_khmer.cc

/*
This file is part of khmer, https://github.com/dib-lab/khmer/, and is
Copyright (C) 2010-2015, Michigan State University.
Copyright (C) 2015, The Regents of the University of California.

Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions are
met:

    * Redistributions of source code must retain the above copyright
      notice, this list of conditions and the following disclaimer.

    * Redistributions in binary form must reproduce the above
      copyright notice, this list of conditions and the following
      disclaimer in the documentation and/or other materials provided
      with the distribution.

    * Neither the name of the Michigan State University nor the names
      of its contributors may be used to endorse or promote products
      derived from this software without specific prior written
      permission.

THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
(INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
LICENSE (END)

Contact: khmer-project@idyll.org
*/

//
// A module for Python that exports khmer C++ library functions.
//

// Must be first.
#include <Python.h>

#include <iostream>

#include "khmer.hh"
#include "kmer_hash.hh"
#include "hashtable.hh"
#include "hashbits.hh"
#include "counting.hh"
#include "read_aligner.hh"
#include "labelhash.hh"
#include "khmer_exception.hh"
#include "hllcounter.hh"

using namespace khmer;
using namespace read_parsers;

//
// Python 2/3 compatibility: PyInt and PyLong
//

#if (PY_MAJOR_VERSION >= 3)
#define PyInt_Check(arg) PyLong_Check(arg)
#define PyInt_AsLong(arg) PyLong_AsLong(arg)
#define PyInt_FromLong(arg) PyLong_FromLong(arg)
#endif

//
// Python 2/3 compatibility: PyBytes and PyString
// https://docs.python.org/2/howto/cporting.html#str-unicode-unification
//

#include "bytesobject.h"

//
// Python 2/3 compatibility: Module initialization
// http://python3porting.com/cextensions.html#module-initialization
//

#if PY_MAJOR_VERSION >= 3
#define MOD_ERROR_VAL NULL
#define MOD_SUCCESS_VAL(val) val
#define MOD_INIT(name) PyMODINIT_FUNC PyInit_##name(void)
#define MOD_DEF(ob, name, doc, methods) \
          static struct PyModuleDef moduledef = { \
            PyModuleDef_HEAD_INIT, name, doc, -1, methods, }; \
          ob = PyModule_Create(&moduledef);
#else
#define MOD_ERROR_VAL
#define MOD_SUCCESS_VAL(val)
#define MOD_INIT(name) void init##name(void)
#define MOD_DEF(ob, name, doc, methods) \
          ob = Py_InitModule3(name, methods, doc);
#endif

using namespace khmer;

//
// Function necessary for Python loading:
//

extern "C" {
    MOD_INIT(_khmer);
}

/***********************************************************************/

//
// Read object -- name, sequence, and FASTQ stuff
//

namespace khmer
{

namespace python
{

typedef struct {
    PyObject_HEAD
    //! Pointer to the low-level genomic read object.
    read_parsers:: Read *   read;
} khmer_Read_Object;


static
void
khmer_Read_dealloc(khmer_Read_Object * obj)
{
    delete obj->read;
    obj->read = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}


static
PyObject *
Read_get_name(khmer_Read_Object * obj, void * closure )
{
    return PyUnicode_FromString(obj->read->name.c_str()) ;
}


static
PyObject *
Read_get_sequence(khmer_Read_Object * obj, void * closure)
{
    return PyUnicode_FromString(obj->read->sequence.c_str()) ;
}


static
PyObject *
Read_get_quality(khmer_Read_Object * obj, void * closure)
{
    return PyUnicode_FromString(obj->read->quality.c_str()) ;
}


static
PyObject *
Read_get_annotations(khmer_Read_Object * obj, void * closure)
{
    return PyUnicode_FromString(obj->read->annotations.c_str()) ;
}


// TODO? Implement setters.


static PyGetSetDef khmer_Read_accessors [ ] = {
    {
        (char *)"name",
        (getter)Read_get_name, (setter)NULL,
        (char *)"Read identifier.", NULL
    },
    {
        (char *)"sequence",
        (getter)Read_get_sequence, (setter)NULL,
        (char *)"Genomic sequence.", NULL
    },
    {
        (char *)"quality",
        (getter)Read_get_quality, (setter)NULL,
        (char *)"Quality scores.", NULL
    },
    {
        (char *)"annotations",
        (getter)Read_get_annotations, (setter)NULL,
        (char *)"Annotations.", NULL
    },

    { NULL, NULL, NULL, NULL, NULL } // sentinel
};


static PyTypeObject khmer_Read_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)        /* init & ob_size */
    "_khmer.Read",                        /* tp_name */
    sizeof(khmer_Read_Object),            /* tp_basicsize */
    0,                                    /* tp_itemsize */
    (destructor)khmer_Read_dealloc,       /* tp_dealloc */
    0,                                    /* tp_print */
    0,                                    /* tp_getattr */
    0,                                    /* tp_setattr */
    0,                                    /* tp_compare */
    0,                                    /* tp_repr */
    0,                                    /* tp_as_number */
    0,                                    /* tp_as_sequence */
    0,                                    /* tp_as_mapping */
    0,                                    /* tp_hash */
    0,                                    /* tp_call */
    0,                                    /* tp_str */
    0,                                    /* tp_getattro */
    0,                                    /* tp_setattro */
    0,                                    /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT,                   /* tp_flags */
    "A FASTQ record plus some metadata.", /* tp_doc */
    0,                                    /* tp_traverse */
    0,                                    /* tp_clear */
    0,                                    /* tp_richcompare */
    0,                                    /* tp_weaklistoffset */
    0,                                    /* tp_iter */
    0,                                    /* tp_iternext */
    0,                                    /* tp_methods */
    0,                                    /* tp_members */
    (PyGetSetDef *)khmer_Read_accessors,  /* tp_getset */
};

/***********************************************************************/

//
// ReadParser object -- parse reads directly from streams
// ReadPairIterator -- return pairs of Read objects
//


typedef struct {
    PyObject_HEAD
    //! Pointer to the low-level parser object.
    read_parsers:: IParser *  parser;
} khmer_ReadParser_Object;


typedef struct {
    PyObject_HEAD
    //! Pointer to Python parser object for reference counting purposes.
    PyObject *  parent;
    //! Persistent value of pair mode across invocations.
    int pair_mode;
} khmer_ReadPairIterator_Object;


static
void
_ReadParser_dealloc(khmer_ReadParser_Object * obj)
{
    Py_DECREF(obj->parser);
    obj->parser = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}


static
void
khmer_ReadPairIterator_dealloc(khmer_ReadPairIterator_Object * obj)
{
    Py_DECREF(obj->parent);
    obj->parent = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}


static
PyObject *
_ReadParser_new( PyTypeObject * subtype, PyObject * args, PyObject * kwds )
{
    const char *      ifile_name_CSTR;

    if (!PyArg_ParseTuple(args, "s", &ifile_name_CSTR )) {
        return NULL;
    }
    std:: string    ifile_name( ifile_name_CSTR );

    PyObject * self     = subtype->tp_alloc( subtype, 1 );
    if (self == NULL) {
        return NULL;
    }
    khmer_ReadParser_Object * myself  = (khmer_ReadParser_Object *)self;

    // Wrap the low-level parser object.
    try {
        myself->parser =
            IParser:: get_parser( ifile_name );
    } catch (khmer_file_exception &exc) {
        PyErr_SetString( PyExc_OSError, exc.what() );
        return NULL;
    }
    return self;
}


static
PyObject *
_ReadParser_iternext( PyObject * self )
{
    khmer_ReadParser_Object * myself  = (khmer_ReadParser_Object *)self;
    IParser *       parser  = myself->parser;

    bool        stop_iteration  = false;
    const char *value_exception = NULL;
    const char *file_exception  = NULL;
    Read       *the_read_PTR    = NULL;
    try {
        the_read_PTR = new Read( );
    } catch (std::bad_alloc &exc) {
        return PyErr_NoMemory();
    }

    Py_BEGIN_ALLOW_THREADS
    stop_iteration = parser->is_complete( );
    if (!stop_iteration) {
        try {
            parser->imprint_next_read( *the_read_PTR );
        } catch (NoMoreReadsAvailable &exc) {
            stop_iteration = true;
        } catch (khmer_file_exception &exc) {
            file_exception = exc.what();
        } catch (khmer_value_exception &exc) {
            value_exception = exc.what();
        }
    }
    Py_END_ALLOW_THREADS

    // Note: Can simply return NULL instead of setting the StopIteration
    //       exception.
    if (stop_iteration) {
        delete the_read_PTR;
        return NULL;
    }

    if (file_exception != NULL) {
        delete the_read_PTR;
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        delete the_read_PTR;
        PyErr_SetString(PyExc_ValueError, value_exception);
        return NULL;
    }

    PyObject * the_read_OBJECT = khmer_Read_Type.tp_alloc( &khmer_Read_Type, 1 );
    ((khmer_Read_Object *)the_read_OBJECT)->read = the_read_PTR;
    return the_read_OBJECT;
}


static
PyObject *
_ReadPairIterator_iternext(khmer_ReadPairIterator_Object * myself)
{
    khmer_ReadParser_Object * parent = (khmer_ReadParser_Object*)myself->parent;
    IParser    *parser    = parent->parser;
    uint8_t     pair_mode = myself->pair_mode;

    ReadPair    the_read_pair;
    bool        stop_iteration  = false;
    const char *value_exception = NULL;
    const char *file_exception  = NULL;

    Py_BEGIN_ALLOW_THREADS
    stop_iteration = parser->is_complete( );
    if (!stop_iteration) {
        try {
            parser->imprint_next_read_pair( the_read_pair, pair_mode );
        } catch (NoMoreReadsAvailable &exc) {
            stop_iteration = true;
        } catch (khmer_file_exception &exc) {
            file_exception = exc.what();
        } catch (khmer_value_exception &exc) {
            value_exception = exc.what();
        }
    }
    Py_END_ALLOW_THREADS

    // Note: Can return NULL instead of setting the StopIteration exception.
    if (stop_iteration) {
        return NULL;
    }
    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
        return NULL;
    }

    // Copy elements of 'ReadPair' object into Python tuple.
    // TODO? Replace dummy reads with 'None' object.
    PyObject * read_1_OBJECT = khmer_Read_Type.tp_alloc( &khmer_Read_Type, 1 );
    try {
        ((khmer_Read_Object *)read_1_OBJECT)->read = new Read( the_read_pair.first );
    } catch (std::bad_alloc &e) {
        return PyErr_NoMemory();
    }
    PyObject * read_2_OBJECT = khmer_Read_Type.tp_alloc( &khmer_Read_Type, 1 );
    try {
        ((khmer_Read_Object *)read_2_OBJECT)->read = new Read( the_read_pair.second );
    } catch (std::bad_alloc &e) {
        delete ((khmer_Read_Object *)read_1_OBJECT)->read;
        return PyErr_NoMemory();
    }
    PyObject * tup = PyTuple_Pack( 2, read_1_OBJECT, read_2_OBJECT );
    Py_XDECREF(read_1_OBJECT);
    Py_XDECREF(read_2_OBJECT);
    return tup;
}

static PyTypeObject khmer_ReadPairIterator_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)              /* init & ob_size */
    "_khmer.ReadPairIterator",                   /* tp_name */
    sizeof(khmer_ReadPairIterator_Object),      /* tp_basicsize */
    0,                                          /* tp_itemsize */
    (destructor)khmer_ReadPairIterator_dealloc, /* tp_dealloc */
    0,                                          /* tp_print */
    0,                                          /* tp_getattr */
    0,                                          /* tp_setattr */
    0,                                          /* tp_compare */
    0,                                          /* tp_repr */
    0,                                          /* tp_as_number */
    0,                                          /* tp_as_sequence */
    0,                                          /* tp_as_mapping */
    0,                                          /* tp_hash */
    0,                                          /* tp_call */
    0,                                          /* tp_str */
    0,                                          /* tp_getattro */
    0,                                          /* tp_setattro */
    0,                                          /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT,                         /* tp_flags */
    "Iterates over 'ReadParser' objects and returns read pairs.", /* tp_doc */
    0,                                          /* tp_traverse */
    0,                                          /* tp_clear */
    0,                                          /* tp_richcompare */
    0,                                          /* tp_weaklistoffset */
    PyObject_SelfIter,                          /* tp_iter */
    (iternextfunc)_ReadPairIterator_iternext,   /* tp_iternext */
};



static
PyObject *
ReadParser_iter_reads(PyObject * self, PyObject * args )
{
    return PyObject_SelfIter( self );
}

static
PyObject *
ReadParser_get_num_reads(khmer_ReadParser_Object * me)
{
    return PyLong_FromLong(me->parser->get_num_reads());
}

static
PyObject *
ReadParser_iter_read_pairs(PyObject * self, PyObject * args )
{
    int  pair_mode  = IParser:: PAIR_MODE_ERROR_ON_UNPAIRED;

    if (!PyArg_ParseTuple( args, "|i", &pair_mode )) {
        return NULL;
    }

    // Capture existing read parser.
    PyObject * obj = khmer_ReadPairIterator_Type.tp_alloc(
                         &khmer_ReadPairIterator_Type, 1
                     );
    if (obj == NULL) {
        return NULL;
    }
    khmer_ReadPairIterator_Object * rpi   = (khmer_ReadPairIterator_Object *)obj;
    rpi->parent             = self;
    rpi->pair_mode          = pair_mode;

    // Increment reference count on existing ReadParser object so that it
    // will not go away until all ReadPairIterator instances have gone away.
    Py_INCREF( self );

    return obj;
}


static PyMethodDef _ReadParser_methods [ ] = {
    {
        "iter_reads",       (PyCFunction)ReadParser_iter_reads,
        METH_NOARGS,        "Iterates over reads."
    },
    {
        "iter_read_pairs",  (PyCFunction)ReadParser_iter_read_pairs,
        METH_VARARGS,       "Iterates over paired reads as pairs."
    },
    { NULL, NULL, 0, NULL } // sentinel
};

static PyGetSetDef khmer_ReadParser_accessors[] = {
    {
        (char *)"num_reads",
        (getter)ReadParser_get_num_reads, NULL,
        (char *)"count of reads processed thus far.",
        NULL
    },
    {NULL, NULL, NULL, NULL, NULL} /* Sentinel */
};

static PyTypeObject khmer_ReadParser_Type
CPYCHECKER_TYPE_OBJECT_FOR_TYPEDEF("khmer_ReadParser_Object")
= {
    PyVarObject_HEAD_INIT(NULL, 0)             /* init & ob_size */
    "_khmer.ReadParser",                        /* tp_name */
    sizeof(khmer_ReadParser_Object),           /* tp_basicsize */
    0,                                         /* tp_itemsize */
    (destructor)_ReadParser_dealloc,           /* tp_dealloc */
    0,                                         /* tp_print */
    0,                                         /* tp_getattr */
    0,                                         /* tp_setattr */
    0,                                         /* tp_compare */
    0,                                         /* tp_repr */
    0,                                         /* tp_as_number */
    0,                                         /* tp_as_sequence */
    0,                                         /* tp_as_mapping */
    0,                                         /* tp_hash */
    0,                                         /* tp_call */
    0,                                         /* tp_str */
    0,                                         /* tp_getattro */
    0,                                         /* tp_setattro */
    0,                                         /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT,                        /* tp_flags */
    "Parses streams from various file formats, " \
    "such as FASTA and FASTQ.",                /* tp_doc */
    0,                                         /* tp_traverse */
    0,                                         /* tp_clear */
    0,                                         /* tp_richcompare */
    0,                                         /* tp_weaklistoffset */
    PyObject_SelfIter,                         /* tp_iter */
    (iternextfunc)_ReadParser_iternext,        /* tp_iternext */
    _ReadParser_methods,                       /* tp_methods */
    0,                                         /* tp_members */
    khmer_ReadParser_accessors,                /* tp_getset */
    0,                                         /* tp_base */
    0,                                         /* tp_dict */
    0,                                         /* tp_descr_get */
    0,                                         /* tp_descr_set */
    0,                                         /* tp_dictoffset */
    0,                                         /* tp_init */
    0,                                         /* tp_alloc */
    _ReadParser_new,                           /* tp_new */
};

void _init_ReadParser_Type_constants()
{
    PyObject * cls_attrs_DICT = PyDict_New( );
    if (cls_attrs_DICT == NULL) {
        return;
    }

    // Place pair mode constants into class dictionary.
    int result;

    PyObject * value = PyLong_FromLong( IParser:: PAIR_MODE_ALLOW_UNPAIRED );
    if (value == NULL) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }
    result = PyDict_SetItemString(cls_attrs_DICT,
                                  "PAIR_MODE_ALLOW_UNPAIRED", value);
    Py_XDECREF(value);
    if (!result) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }

    value = PyLong_FromLong( IParser:: PAIR_MODE_IGNORE_UNPAIRED );
    if (value == NULL) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }
    result = PyDict_SetItemString(cls_attrs_DICT,
                                  "PAIR_MODE_IGNORE_UNPAIRED", value );
    Py_XDECREF(value);
    if (!result) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }

    value = PyLong_FromLong( IParser:: PAIR_MODE_ERROR_ON_UNPAIRED );
    if (value == NULL) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }
    result = PyDict_SetItemString(cls_attrs_DICT,
                                  "PAIR_MODE_ERROR_ON_UNPAIRED", value);
    Py_XDECREF(value);
    if (!result) {
        Py_DECREF(cls_attrs_DICT);
        return;
    }

    khmer_ReadParser_Type.tp_dict     = cls_attrs_DICT;
}

} // namespace python

} // namespace khmer


static
read_parsers:: IParser *
_PyObject_to_khmer_ReadParser( PyObject * py_object )
{
    // TODO: Add type-checking.

    return ((python:: khmer_ReadParser_Object *)py_object)->parser;
}

typedef struct {
    PyObject_HEAD
    pre_partition_info *   PrePartitionInfo;
} khmer_PrePartitionInfo_Object;

static
void
khmer_PrePartitionInfo_dealloc(khmer_PrePartitionInfo_Object * obj)
{
    delete obj->PrePartitionInfo;
    obj->PrePartitionInfo = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}

static PyTypeObject khmer_PrePartitionInfo_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)        /* init & ob_size */
    "_khmer.PrePartitionInfo",            /* tp_name */
    sizeof(khmer_PrePartitionInfo_Object),/* tp_basicsize */
    0,                                    /* tp_itemsize */
    (destructor)khmer_PrePartitionInfo_dealloc,       /* tp_dealloc */
    0,                                    /* tp_print */
    0,                                    /* tp_getattr */
    0,                                    /* tp_setattr */
    0,                                    /* tp_compare */
    0,                                    /* tp_repr */
    0,                                    /* tp_as_number */
    0,                                    /* tp_as_sequence */
    0,                                    /* tp_as_mapping */
    0,                                    /* tp_hash */
    0,                                    /* tp_call */
    0,                                    /* tp_str */
    0,                                    /* tp_getattro */
    0,                                    /* tp_setattro */
    0,                                    /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT,                   /* tp_flags */
    "Stores a k-kmer and a set of tagged seen k-mers.", /* tp_doc */
};


/***********************************************************************/

typedef struct {
    PyObject_HEAD
    Hashtable * hashtable;
} khmer_KHashtable_Object;

typedef struct {
    PyObject_HEAD
    SubsetPartition * subset;
} khmer_KSubsetPartition_Object;

static void khmer_subset_dealloc(khmer_KSubsetPartition_Object * obj);

static PyTypeObject khmer_KSubsetPartition_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)         /* init & ob_size */
    "_khmer.KSubsetPartition",              /* tp_name */
    sizeof(khmer_KSubsetPartition_Object), /* tp_basicsize */
    0,                                     /* tp_itemsize */
    (destructor)khmer_subset_dealloc,      /*tp_dealloc*/
    0,                                     /*tp_print*/
    0,                                     /*tp_getattr*/
    0,                                     /*tp_setattr*/
    0,                                     /*tp_compare*/
    0,                                     /*tp_repr*/
    0,                                     /*tp_as_number*/
    0,                                     /*tp_as_sequence*/
    0,                                     /*tp_as_mapping*/
    0,                                     /*tp_hash */
    0,                                     /*tp_call*/
    0,                                     /*tp_str*/
    0,                                     /*tp_getattro*/
    0,                                     /*tp_setattro*/
    0,                                     /*tp_as_buffer*/
    Py_TPFLAGS_DEFAULT,                    /*tp_flags*/
    "subset object",                       /* tp_doc */
};

typedef struct {
    khmer_KHashtable_Object khashtable;
    Hashbits * hashbits;
} khmer_KHashbits_Object;

static void khmer_hashbits_dealloc(khmer_KHashbits_Object * obj);
static PyObject* khmer_hashbits_new(PyTypeObject * type, PyObject * args,
                                    PyObject * kwds);

static PyTypeObject khmer_KNodegraph_Type
CPYCHECKER_TYPE_OBJECT_FOR_TYPEDEF("khmer_KHashbits_Object")
= {
    PyVarObject_HEAD_INIT(NULL, 0) /* init & ob_size */
    "_khmer.Nodegraph",             /* tp_name */
    sizeof(khmer_KHashbits_Object), /* tp_basicsize */
    0,                             /* tp_itemsize */
    (destructor)khmer_hashbits_dealloc, /*tp_dealloc*/
    0,              /*tp_print*/
    0,              /*tp_getattr*/
    0,              /*tp_setattr*/
    0,              /*tp_compare*/
    0,              /*tp_repr*/
    0,              /*tp_as_number*/
    0,              /*tp_as_sequence*/
    0,              /*tp_as_mapping*/
    0,              /*tp_hash */
    0,              /*tp_call*/
    0,              /*tp_str*/
    0,              /*tp_getattro*/
    0,              /*tp_setattro*/
    0,              /*tp_as_buffer*/
    Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE,       /*tp_flags*/
    "hashbits object",           /* tp_doc */
    0,                       /* tp_traverse */
    0,                       /* tp_clear */
    0,                       /* tp_richcompare */
    0,                       /* tp_weaklistoffset */
    0,                       /* tp_iter */
    0,                       /* tp_iternext */
    0,  /* tp_methods */
    0,                       /* tp_members */
    0,                       /* tp_getset */
    0,                       /* tp_base */
    0,                       /* tp_dict */
    0,                       /* tp_descr_get */
    0,                       /* tp_descr_set */
    0,                       /* tp_dictoffset */
    0,                       /* tp_init */
    0,                       /* tp_alloc */
    khmer_hashbits_new,                  /* tp_new */
};


static
PyObject *
hashtable_get_ksize(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    unsigned int k = hashtable->ksize();

    return PyLong_FromLong(k);
}

static
PyObject *
hashtable_n_occupied(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    HashIntoType n = hashtable->n_occupied();

    return PyLong_FromUnsignedLongLong(n);
}

static
PyObject *
hashtable_n_unique_kmers(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    HashIntoType n = hashtable->n_unique_kmers();

    return PyLong_FromUnsignedLongLong(n);
}

static
PyObject *
hashtable_count(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer;

    if (!PyArg_ParseTuple(args, "s", &kmer)) {
        return NULL;
    }

    if (strlen(kmer) != hashtable->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "k-mer length must be the same as the hashtable k-size");
        return NULL;
    }

    hashtable->count(kmer);

    return PyLong_FromLong(1);
}

static
PyObject *
hashtable_consume_fasta(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable  = me->hashtable;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python
    unsigned long long  n_consumed    = 0;
    unsigned int          total_reads   = 0;
    try {
        hashtable->consume_fasta(filename, total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
hashtable_consume_fasta_with_reads_parser(khmer_KHashtable_Object * me,
        PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    PyObject * rparser_obj = NULL;

    if (!PyArg_ParseTuple(args, "O", &rparser_obj)) {
        return NULL;
    }

    read_parsers:: IParser * rparser =
        _PyObject_to_khmer_ReadParser( rparser_obj );

    // call the C++ function, and trap signals => Python
    unsigned long long  n_consumed      = 0;
    unsigned int        total_reads     = 0;
    const char         *value_exception = NULL;
    const char         *file_exception  = NULL;

    Py_BEGIN_ALLOW_THREADS
    try {
        hashtable->consume_fasta(rparser, total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    } catch (khmer_value_exception &exc) {
        value_exception = exc.what();
    }
    Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
hashtable_consume(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * long_str;

    if (!PyArg_ParseTuple(args, "s", &long_str)) {
        return NULL;
    }

    if (strlen(long_str) < hashtable->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    unsigned int n_consumed;
    n_consumed = hashtable->consume_string(long_str);

    return PyLong_FromLong(n_consumed);
}

static
PyObject *
hashtable_get(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    PyObject * arg;

    if (!PyArg_ParseTuple(args, "O", &arg)) {
        return NULL;
    }

    unsigned long count = 0;

    if (PyInt_Check(arg) || PyLong_Check(arg)) {
        long pos = PyInt_AsLong(arg);
        count = hashtable->get_count((unsigned int) pos);
    } else if (PyUnicode_Check(arg)) {
        std::string s = PyBytes_AsString(PyUnicode_AsEncodedString(
                                             arg, "utf-8", "strict"));
        if (strlen(s.c_str()) != hashtable->ksize()) {
            PyErr_SetString(PyExc_ValueError,
                            "k-mer size must equal the presence table k-mer size");
            return NULL;
        }
        count = hashtable->get_count(s.c_str());
    } else if (PyBytes_Check(arg)) {
        std::string s = PyBytes_AsString(arg);

        if (strlen(s.c_str()) != hashtable->ksize()) {
            PyErr_SetString(PyExc_ValueError,
                            "k-mer size must equal the counting table k-mer size");
            return NULL;
        }

        count = hashtable->get_count(s.c_str());
    } else {
        PyErr_SetString(PyExc_ValueError,
                        "please pass either a hash value or a string");
        return NULL;
    }

    return PyLong_FromLong(count);
}

static
PyObject *
hashtable_load(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->load(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_save(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->save(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_get_hashsizes(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;


    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    std::vector<HashIntoType> ts = hashtable->get_tablesizes();

    PyObject * x = PyList_New(ts.size());
    for (size_t i = 0; i < ts.size(); i++) {
        PyList_SET_ITEM(x, i, PyLong_FromUnsignedLongLong(ts[i]));
    }

    return x;
}

static
PyObject *
hashtable_consume_and_tag(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * seq;

    if (!PyArg_ParseTuple(args, "s", &seq)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    unsigned long long n_consumed = 0;

    // @CTB needs to normalize
    hashtable->consume_sequence_and_tag(seq, n_consumed);

    return Py_BuildValue("K", n_consumed);
}

static
PyObject *
hashtable_get_tags_and_positions(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * seq;

    if (!PyArg_ParseTuple(args, "s", &seq)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    std::vector<unsigned int> posns;
    std::vector<HashIntoType> tags;

    unsigned int pos = 1;
    KmerIterator kmers(seq, hashtable->ksize());

    while (!kmers.done()) {
        HashIntoType kmer = kmers.next();
        if (set_contains(hashtable->all_tags, kmer)) {
            posns.push_back(pos);
            tags.push_back(kmer);
        }
        pos++;
    }

    PyObject * posns_list = PyList_New(posns.size());
    for (size_t i = 0; i < posns.size(); i++) {
        PyObject * tup = Py_BuildValue("IK", posns[i], tags[i]);
        PyList_SET_ITEM(posns_list, i, tup);
    }

    return posns_list;
}

static
PyObject *
hashtable_find_all_tags_list(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer_s = NULL;

    if (!PyArg_ParseTuple(args, "s", &kmer_s)) {
        return NULL;
    }

    if (strlen(kmer_s) != hashtable->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "k-mer length must equal the counting table k-mer size");
        return NULL;
    }

    SeenSet tags;

    Kmer start_kmer = hashtable->build_kmer(kmer_s);

    Py_BEGIN_ALLOW_THREADS

    hashtable->partition->find_all_tags(start_kmer, tags,
                                        hashtable->all_tags);

    Py_END_ALLOW_THREADS

    PyObject * x =  PyList_New(tags.size());
    if (x == NULL) {
        return NULL;
    }
    SeenSet::iterator si;
    unsigned long long i = 0;
    for (si = tags.begin(); si != tags.end(); ++si) {
        // type K for python unsigned long long
        PyList_SET_ITEM(x, i, Py_BuildValue("K", *si));
        i++;
    }

    return x;
}

static
PyObject *
hashtable_consume_fasta_and_tag(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    unsigned long long n_consumed;
    unsigned int total_reads;

    try {
        hashtable->consume_fasta_and_tag(filename, total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
hashtable_get_median_count(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * long_str;

    if (!PyArg_ParseTuple(args, "s", &long_str)) {
        return NULL;
    }

    if (strlen(long_str) < hashtable->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    BoundedCounterType med = 0;
    float average = 0, stddev = 0;

    hashtable->get_median_count(long_str, med, average, stddev);

    return Py_BuildValue("iff", med, average, stddev);
}

static
PyObject *
hashtable_median_at_least(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * long_str;
    unsigned int cutoff;

    if (!PyArg_ParseTuple(args, "sI", &long_str, &cutoff)) {
        return NULL;
    }

    if (strlen(long_str) < hashtable->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    if (hashtable->median_at_least(long_str, cutoff)) {
        Py_RETURN_TRUE;
    }
    Py_RETURN_FALSE;

}

static
PyObject *
hashtable_n_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    return PyLong_FromSize_t(hashtable->n_tags());
}

static
PyObject *
hashtable_print_stop_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    hashtable->print_stop_tags(filename);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_print_tagset(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    hashtable->print_tagset(filename);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_load_stop_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    PyObject * clear_tags_o = NULL;

    if (!PyArg_ParseTuple(args, "s|O", &filename, &clear_tags_o)) {
        return NULL;
    }

    bool clear_tags = true;
    if (clear_tags_o && !PyObject_IsTrue(clear_tags_o)) {
        clear_tags = false;
    }


    try {
        hashtable->load_stop_tags(filename, clear_tags);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}


static
PyObject *
hashtable_save_stop_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->save_stop_tags(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static PyObject * hashtable_traverse_from_tags(khmer_KHashtable_Object * me,
        PyObject * args);

static PyObject * hashtable_repartition_largest_partition(
    khmer_KHashtable_Object * me,
    PyObject * args);

static
PyObject *
hashtable_calc_connected_graph_size(khmer_KHashtable_Object * me,
                                    PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * _kmer;
    unsigned int max_size = 0;
    PyObject * break_on_circum_o = NULL;
    if (!PyArg_ParseTuple(args, "s|IO", &_kmer, &max_size, &break_on_circum_o)) {
        return NULL;
    }

    bool break_on_circum = false;
    if (break_on_circum_o && PyObject_IsTrue(break_on_circum_o)) {
        break_on_circum = true;
    }

    unsigned long long size = 0;
    Kmer start_kmer = hashtable->build_kmer(_kmer);

    Py_BEGIN_ALLOW_THREADS
    KmerSet keeper;
    hashtable->calc_connected_graph_size(start_kmer, size, keeper, max_size,
                                         break_on_circum);
    Py_END_ALLOW_THREADS

    return PyLong_FromUnsignedLongLong(size);
}

static
PyObject *
hashtable_kmer_degree(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer_s = NULL;

    if (!PyArg_ParseTuple(args, "s", &kmer_s)) {
        return NULL;
    }

    return PyLong_FromLong(hashtable->kmer_degree(kmer_s));
}

static
PyObject *
hashtable_trim_on_stoptags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * seq = NULL;

    if (!PyArg_ParseTuple(args, "s", &seq)) {
        return NULL;
    }

    size_t trim_at;
    Py_BEGIN_ALLOW_THREADS

    trim_at = hashtable->trim_on_stoptags(seq);

    Py_END_ALLOW_THREADS;

    PyObject * trim_seq = PyUnicode_FromStringAndSize(seq, trim_at);
    if (trim_seq == NULL) {
        return NULL;
    }
    PyObject * ret = Py_BuildValue("Ok", trim_seq, (unsigned long) trim_at);
    Py_DECREF(trim_seq);

    return ret;
}

static
PyObject *
hashtable_identify_stoptags_by_position(khmer_KHashtable_Object * me,
                                        PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * seq = NULL;

    if (!PyArg_ParseTuple(args, "s", &seq)) {
        return NULL;
    }

    std::vector<unsigned int> posns;
    Py_BEGIN_ALLOW_THREADS

    hashtable->identify_stop_tags_by_position(seq, posns);

    Py_END_ALLOW_THREADS;

    PyObject * x = PyList_New(posns.size());

    for (unsigned int i = 0; i < posns.size(); i++) {
        PyList_SET_ITEM(x, i, Py_BuildValue("I", posns[i]));
    }

    return x;
}

static
PyObject *
hashtable_do_subset_partition(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    HashIntoType start_kmer = 0, end_kmer = 0;
    PyObject * break_on_stop_tags_o = NULL;
    PyObject * stop_big_traversals_o = NULL;

    if (!PyArg_ParseTuple(args, "|KKOO", &start_kmer, &end_kmer,
                          &break_on_stop_tags_o,
                          &stop_big_traversals_o)) {
        return NULL;
    }

    bool break_on_stop_tags = false;
    if (break_on_stop_tags_o && PyObject_IsTrue(break_on_stop_tags_o)) {
        break_on_stop_tags = true;
    }
    bool stop_big_traversals = false;
    if (stop_big_traversals_o && PyObject_IsTrue(stop_big_traversals_o)) {
        stop_big_traversals = true;
    }

    SubsetPartition * subset_p = NULL;
    try {
        Py_BEGIN_ALLOW_THREADS
        subset_p = new SubsetPartition(hashtable);
        subset_p->do_partition(start_kmer, end_kmer, break_on_stop_tags,
                               stop_big_traversals);
        Py_END_ALLOW_THREADS
    } catch (std::bad_alloc &e) {
        return PyErr_NoMemory();
    }

    khmer_KSubsetPartition_Object * subset_obj = (khmer_KSubsetPartition_Object *)\
            PyObject_New(khmer_KSubsetPartition_Object, &khmer_KSubsetPartition_Type);

    if (subset_obj == NULL) {
        delete subset_p;
        return NULL;
    }

    subset_obj->subset = subset_p;

    return (PyObject *) subset_obj;
}


static
PyObject *
hashtable_join_partitions_by_path(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * sequence = NULL;
    if (!PyArg_ParseTuple(args, "s", &sequence)) {
        return NULL;
    }

    hashtable->partition->join_partitions_by_path(sequence);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_merge_subset(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    khmer_KSubsetPartition_Object * subset_obj = NULL;
    if (!PyArg_ParseTuple(args, "O!", &khmer_KSubsetPartition_Type,
                          &subset_obj)) {
        return NULL;
    }

    SubsetPartition * subset_p = subset_obj->subset;

    hashtable->partition->merge(subset_p);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_merge_from_disk(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->partition->merge_from_disk(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_consume_fasta_and_tag_with_reads_parser(khmer_KHashtable_Object * me,
        PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    python::khmer_ReadParser_Object * rparser_obj = NULL;

    if (!PyArg_ParseTuple( args, "O!", &python::khmer_ReadParser_Type,
                           &rparser_obj)) {
        return NULL;
    }

    read_parsers:: IParser * rparser = rparser_obj-> parser;

    // call the C++ function, and trap signals => Python
    const char         *value_exception = NULL;
    const char         *file_exception  = NULL;
    unsigned long long  n_consumed      = 0;
    unsigned int        total_reads     = 0;

    Py_BEGIN_ALLOW_THREADS
    try {
        hashtable->consume_fasta_and_tag(rparser, total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    } catch (khmer_value_exception &exc) {
        value_exception = exc.what();
    }
    Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static PyObject * hashtable_consume_fasta_and_traverse(
    khmer_KHashtable_Object * me,
    PyObject * args);


static
PyObject *
hashtable_consume_fasta_and_tag_with_stoptags(khmer_KHashtable_Object * me,
        PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    unsigned long long  n_consumed;
    unsigned int        total_reads;
    try {
        hashtable->consume_fasta_and_tag_with_stoptags(filename,
                total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
hashtable_consume_partitioned_fasta(khmer_KHashtable_Object * me,
                                    PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    unsigned long long n_consumed;
    unsigned int total_reads;

    try {
        hashtable->consume_partitioned_fasta(filename, total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
hashtable_find_all_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer_s = NULL;

    if (!PyArg_ParseTuple(args, "s", &kmer_s)) {
        return NULL;
    }

    if (strlen(kmer_s) != hashtable->ksize()) {
        PyErr_SetString( PyExc_ValueError,
                         "k-mer size must equal the k-mer size of the presence table");
        return NULL;
    }

    pre_partition_info * ppi = NULL;

    Kmer kmer = hashtable->build_kmer(kmer_s);

    Py_BEGIN_ALLOW_THREADS

    try {
        ppi = new pre_partition_info(kmer);
    } catch (std::bad_alloc &e) {
        return PyErr_NoMemory();
    }
    hashtable->partition->find_all_tags(kmer, ppi->tagged_kmers,
                                        hashtable->all_tags);
    hashtable->add_kmer_to_tags(kmer);

    Py_END_ALLOW_THREADS

    khmer_PrePartitionInfo_Object * ppi_obj = (khmer_PrePartitionInfo_Object *) \
            PyObject_New(khmer_PrePartitionInfo_Object, &khmer_PrePartitionInfo_Type);

    ppi_obj->PrePartitionInfo = ppi;

    return (PyObject*)ppi_obj;
}

static
PyObject *
hashtable_assign_partition_id(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    khmer_PrePartitionInfo_Object * ppi_obj;
    if (!PyArg_ParseTuple(args, "O!", &khmer_PrePartitionInfo_Type, &ppi_obj)) {
        return NULL;
    }

    pre_partition_info * ppi;
    ppi = ppi_obj->PrePartitionInfo;

    PartitionID p;
    p = hashtable->partition->assign_partition_id(ppi->kmer,
            ppi->tagged_kmers);

    return PyLong_FromLong(p);
}

static
PyObject *
hashtable_add_tag(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer_s = NULL;
    if (!PyArg_ParseTuple(args, "s", &kmer_s)) {
        return NULL;
    }

    HashIntoType kmer = _hash(kmer_s, hashtable->ksize());
    hashtable->add_tag(kmer);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_add_stop_tag(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer_s = NULL;
    if (!PyArg_ParseTuple(args, "s", &kmer_s)) {
        return NULL;
    }

    HashIntoType kmer = _hash(kmer_s, hashtable->ksize());
    hashtable->add_stop_tag(kmer);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_get_stop_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    WordLength k = hashtable->ksize();
    SeenSet::const_iterator si;

    PyObject * x = PyList_New(hashtable->stop_tags.size());
    unsigned long long i = 0;
    for (si = hashtable->stop_tags.begin(); si != hashtable->stop_tags.end();
            ++si) {
        std::string s = _revhash(*si, k);
        PyList_SET_ITEM(x, i, Py_BuildValue("s", s.c_str()));
        i++;
    }

    return x;
}

static
PyObject *
hashtable_get_tagset(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    WordLength k = hashtable->ksize();
    SeenSet::const_iterator si;

    PyObject * x = PyList_New(hashtable->all_tags.size());
    unsigned long long i = 0;
    for (si = hashtable->all_tags.begin(); si != hashtable->all_tags.end();
            ++si) {
        std::string s = _revhash(*si, k);
        PyList_SET_ITEM(x, i, Py_BuildValue("s", s.c_str()));
        i++;
    }

    return x;
}

static
PyObject *
hashtable_output_partitions(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    const char * output = NULL;
    PyObject * output_unassigned_o = NULL;

    if (!PyArg_ParseTuple(args, "ss|O", &filename, &output,
                          &output_unassigned_o)) {
        return NULL;
    }

    bool output_unassigned = false;
    if (output_unassigned_o != NULL && PyObject_IsTrue(output_unassigned_o)) {
        output_unassigned = true;
    }

    size_t n_partitions = 0;

    try {
        SubsetPartition * subset_p = hashtable->partition;
        n_partitions = subset_p->output_partitioned_file(filename,
                       output,
                       output_unassigned);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return PyLong_FromLong(n_partitions);
}

static
PyObject *
hashtable_find_unpart(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    PyObject * traverse_o = NULL;
    PyObject * stop_big_traversals_o = NULL;

    if (!PyArg_ParseTuple(args, "sOO", &filename, &traverse_o,
                          &stop_big_traversals_o)) {
        return NULL;
    }

    bool traverse = PyObject_IsTrue(traverse_o);
    bool stop_big_traversals = PyObject_IsTrue(stop_big_traversals_o);
    unsigned int n_singletons = 0;

    try {
        SubsetPartition * subset_p = hashtable->partition;
        n_singletons = subset_p->find_unpart(filename, traverse,
                                             stop_big_traversals);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return PyLong_FromLong(n_singletons);
}

static
PyObject *
hashtable_filter_if_present(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    const char * output = NULL;

    if (!PyArg_ParseTuple(args, "ss", &filename, &output)) {
        return NULL;
    }

    try {
        hashtable->filter_if_present(filename, output);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_save_partitionmap(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->partition->save_partitionmap(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_load_partitionmap(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->partition->load_partitionmap(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable__validate_partitionmap(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    hashtable->partition->_validate_pmap();

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_count_partitions(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    size_t n_partitions = 0, n_unassigned = 0;
    hashtable->partition->count_partitions(n_partitions, n_unassigned);

    return Py_BuildValue("nn", (Py_ssize_t) n_partitions,
                         (Py_ssize_t) n_unassigned);
}

static
PyObject *
hashtable_subset_count_partitions(khmer_KHashtable_Object * me, PyObject * args)
{
    khmer_KSubsetPartition_Object * subset_obj = NULL;

    if (!PyArg_ParseTuple(args, "O!", &khmer_KSubsetPartition_Type,
                          &subset_obj)) {
        return NULL;
    }


    size_t n_partitions = 0, n_unassigned = 0;
    subset_obj->subset->count_partitions(n_partitions, n_unassigned);

    return Py_BuildValue("nn", (Py_ssize_t) n_partitions,
                         (Py_ssize_t) n_unassigned);
}

static
PyObject *
hashtable_subset_partition_size_distribution(khmer_KHashtable_Object * me,
        PyObject * args)
{
    khmer_KSubsetPartition_Object * subset_obj = NULL;
    if (!PyArg_ParseTuple(args, "O!", &khmer_KSubsetPartition_Type,
                          &subset_obj)) {
        return NULL;
    }

    SubsetPartition * subset_p = subset_obj->subset;

    PartitionCountDistribution d;

    unsigned int n_unassigned = 0;
    subset_p->partition_size_distribution(d, n_unassigned);

    PyObject * x = PyList_New(d.size());
    if (x == NULL) {
        return NULL;
    }
    PartitionCountDistribution::iterator di;

    unsigned int i;
    for (i = 0, di = d.begin(); di != d.end(); ++di, i++) {
        PyObject * value =  Py_BuildValue("KK", di->first, di->second);
        if (value == NULL) {
            Py_DECREF(x);
            return NULL;
        }
        PyList_SET_ITEM(x, i, value);
    }
    if (!(i == d.size())) {
        throw khmer_exception();
    }

    PyObject * returnValue = Py_BuildValue("NI", x, n_unassigned);
    if (returnValue == NULL) {
        Py_DECREF(x);
        return NULL;
    }
    return returnValue;
}

static
PyObject *
hashtable_load_tagset(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;
    PyObject * clear_tags_o = NULL;

    if (!PyArg_ParseTuple(args, "s|O", &filename, &clear_tags_o)) {
        return NULL;
    }

    bool clear_tags = true;
    if (clear_tags_o && !PyObject_IsTrue(clear_tags_o)) {
        clear_tags = false;
    }

    try {
        hashtable->load_tagset(filename, clear_tags);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_save_tagset(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        hashtable->save_tagset(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_save_subset_partitionmap(khmer_KHashtable_Object * me,
                                   PyObject * args)
{
    const char * filename = NULL;
    khmer_KSubsetPartition_Object * subset_obj = NULL;

    if (!PyArg_ParseTuple(args, "O!s", &khmer_KSubsetPartition_Type,
                          &subset_obj, &filename)) {
        return NULL;
    }

    SubsetPartition * subset_p = subset_obj->subset;

    Py_BEGIN_ALLOW_THREADS

    try {
        subset_p->save_partitionmap(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_END_ALLOW_THREADS

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_load_subset_partitionmap(khmer_KHashtable_Object * me,
                                   PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename = NULL;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    SubsetPartition * subset_p;
    try {
        subset_p = new SubsetPartition(hashtable);
    } catch (std::bad_alloc &e) {
        return PyErr_NoMemory();
    }

    const char         *file_exception  = NULL;

    Py_BEGIN_ALLOW_THREADS
    try {
        subset_p->load_partitionmap(filename);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    }
    Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        delete subset_p;
        return NULL;
    }

    khmer_KSubsetPartition_Object * subset_obj = (khmer_KSubsetPartition_Object *)\
            PyObject_New(khmer_KSubsetPartition_Object, &khmer_KSubsetPartition_Type);

    if (subset_obj == NULL) {
        delete subset_p;
        return NULL;
    }

    subset_obj->subset = subset_p;

    return (PyObject *) subset_obj;
}

static
PyObject *
hashtable__set_tag_density(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    unsigned int d;
    if (!PyArg_ParseTuple(args, "I", &d)) {
        return NULL;
    }

    hashtable->_set_tag_density(d);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable__get_tag_density(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    unsigned int d = hashtable->_get_tag_density();

    return PyLong_FromLong(d);
}

static
PyObject *
hashtable__validate_subset_partitionmap(khmer_KHashtable_Object * me,
                                        PyObject * args)
{
    khmer_KSubsetPartition_Object * subset_obj = NULL;

    if (!PyArg_ParseTuple(args, "O!", &khmer_KSubsetPartition_Type,
                          &subset_obj)) {
        return NULL;
    }

    SubsetPartition * subset_p = subset_obj->subset;

    subset_p->_validate_pmap();

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_set_partition_id(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer = NULL;
    PartitionID p = 0;

    if (!PyArg_ParseTuple(args, "sI", &kmer, &p)) {
        return NULL;
    }

    hashtable->partition->set_partition_id(kmer, p);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_join_partitions(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    PartitionID p1 = 0, p2 = 0;

    if (!PyArg_ParseTuple(args, "II", &p1, &p2)) {
        return NULL;
    }

    p1 = hashtable->partition->join_partitions(p1, p2);

    return PyLong_FromLong(p1);
}

static
PyObject *
hashtable_get_partition_id(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer = NULL;

    if (!PyArg_ParseTuple(args, "s", &kmer)) {
        return NULL;
    }

    PartitionID partition_id;
    partition_id = hashtable->partition->get_partition_id(kmer);

    return PyLong_FromLong(partition_id);
}

static
PyObject *
hashtable_is_single_partition(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * seq = NULL;

    if (!PyArg_ParseTuple(args, "s", &seq)) {
        return NULL;
    }

    bool v = hashtable->partition->is_single_partition(seq);

    PyObject * val;
    if (v) {
        val = Py_True;
    } else {
        val = Py_False;
    }
    Py_INCREF(val);

    return val;
}

static
PyObject *
hashtable_divide_tags_into_subsets(khmer_KHashtable_Object * me,
                                   PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    unsigned int subset_size = 0;

    if (!PyArg_ParseTuple(args, "I", &subset_size)) {
        return NULL;
    }

    SeenSet divvy;
    hashtable->divide_tags_into_subsets(subset_size, divvy);

    PyObject * x = PyList_New(divvy.size());
    unsigned int i = 0;
    for (SeenSet::const_iterator si = divvy.begin(); si != divvy.end();
            ++si, i++) {
        PyList_SET_ITEM(x, i, PyLong_FromUnsignedLongLong(*si));
    }

    return x;
}

static
PyObject *
hashtable_count_kmers_within_radius(khmer_KHashtable_Object * me,
                                    PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * kmer = NULL;
    unsigned int radius = 0;
    unsigned int max_count = 0;

    if (!PyArg_ParseTuple(args, "sI|I", &kmer, &radius, &max_count)) {
        return NULL;
    }

    unsigned int n;

    Py_BEGIN_ALLOW_THREADS
    Kmer start_kmer = hashtable->build_kmer(kmer);
    KmerSet seen;
    n = hashtable->traverse_from_kmer(start_kmer, radius,
                                      seen, max_count);

    Py_END_ALLOW_THREADS

    return PyLong_FromUnsignedLong(n);
}

static
PyObject *
hashtable_extract_unique_paths(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * sequence = NULL;
    unsigned int min_length = 0;
    float min_unique_f = 0;
    if (!PyArg_ParseTuple(args, "sIf", &sequence, &min_length, &min_unique_f)) {
        return NULL;
    }

    std::vector<std::string> results;
    hashtable->extract_unique_paths(sequence, min_length, min_unique_f, results);

    PyObject * x = PyList_New(results.size());
    if (x == NULL) {
        return NULL;
    }

    for (unsigned int i = 0; i < results.size(); i++) {
        PyList_SET_ITEM(x, i, PyUnicode_FromString(results[i].c_str()));
    }

    return x;
}


static
PyObject *
hashtable_get_kmers(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;
    const char * sequence;

    if (!PyArg_ParseTuple(args, "s", &sequence)) {
        return NULL;
    }

    std::vector<std::string> kmers;

    hashtable->get_kmers(sequence, kmers);

    PyObject * x = PyList_New(kmers.size());
    for (unsigned int i = 0; i < kmers.size(); i++) {
        PyObject * obj = PyUnicode_FromString(kmers[i].c_str());
        PyList_SET_ITEM(x, i, obj);
    }

    return x;
}

static
PyObject *
hashtable_get_kmer_counts(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;
    const char * sequence;

    if (!PyArg_ParseTuple(args, "s", &sequence)) {
        return NULL;
    }

    std::vector<BoundedCounterType> counts;
    hashtable->get_kmer_counts(sequence, counts);

    PyObject * x = PyList_New(counts.size());
    for (unsigned int i = 0; i <counts.size(); i++) {
        PyObject * obj = PyInt_FromLong(counts[i]);
        PyList_SET_ITEM(x, i, obj);
    }

    return x;
}


static
PyObject *
hashtable_get_kmer_hashes(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;
    const char * sequence;

    if (!PyArg_ParseTuple(args, "s", &sequence)) {
        return NULL;
    }

    std::vector<HashIntoType> hashes;
    hashtable->get_kmer_hashes(sequence, hashes);

    PyObject * x = PyList_New(hashes.size());
    for (unsigned int i = 0; i < hashes.size(); i++) {
        PyObject * obj = PyLong_FromUnsignedLongLong(hashes[i]);
        PyList_SET_ITEM(x, i, obj);
    }

    return x;
}


static PyMethodDef khmer_hashtable_methods[] = {
    //
    // Basic methods
    //

    {
        "ksize",
        (PyCFunction)hashtable_get_ksize, METH_VARARGS,
        "Returns the k-mer size of this graph."
    },
    { "hashsizes", (PyCFunction)hashtable_get_hashsizes, METH_VARARGS, "" },
    {
        "n_unique_kmers",
        (PyCFunction)hashtable_n_unique_kmers, METH_VARARGS,
        "Count the number of unique kmers in this graph."
    },
    {
        "n_occupied", (PyCFunction)hashtable_n_occupied, METH_VARARGS,
        "Count the number of occupied bins."
    },
    {
        "count",
        (PyCFunction)hashtable_count, METH_VARARGS,
        "Increment the count of this k-mer."
    },
    {
        "consume",
        (PyCFunction)hashtable_consume, METH_VARARGS,
        "Increment the counts of all of the k-mers in the string."
    },
    {
        "consume_fasta",
        (PyCFunction)hashtable_consume_fasta, METH_VARARGS,
        "Incrment the counts of all the k-mers in the sequences in the "
        "given file"
    },
    {
        "consume_fasta_with_reads_parser",
        (PyCFunction)hashtable_consume_fasta_with_reads_parser, METH_VARARGS,
        "Count all k-mers retrieved with this reads parser object."
    },
    {
        "get",
        (PyCFunction)hashtable_get, METH_VARARGS,
        "Retrieve the count for the given k-mer."
    },
    {
        "load",
        (PyCFunction)hashtable_load, METH_VARARGS,
        "Load the graph from the specified file."
    },
    {
        "save",
        (PyCFunction)hashtable_save, METH_VARARGS,
        "Save the graph to the specified file."
    },
    {
        "get_median_count",
        (PyCFunction)hashtable_get_median_count, METH_VARARGS,
        "Get the median, average, and stddev of the k-mer counts "
        " in the string"
    },
    {
        "get_kmers",
        (PyCFunction)hashtable_get_kmers, METH_VARARGS,
        "Generate an ordered list of all substrings of length k in the string."
    },
    {
        "get_kmer_hashes",
        (PyCFunction)hashtable_get_kmer_hashes, METH_VARARGS,
        "Retrieve an ordered list of all hashes of all k-mers in the string."
    },
    {
        "get_kmer_counts",
        (PyCFunction)hashtable_get_kmer_counts, METH_VARARGS,
        "Retrieve an ordered list of the counts of all k-mers in the string."
    },

    //
    // graph/traversal functionality
    //

    {
        "calc_connected_graph_size",
        (PyCFunction)hashtable_calc_connected_graph_size, METH_VARARGS, ""
    },
    {
        "kmer_degree",
        (PyCFunction)hashtable_kmer_degree, METH_VARARGS,
        "Calculate the number of immediate neighbors this k-mer has in "
        "the graph."
    },
    {
        "count_kmers_within_radius",
        (PyCFunction)hashtable_count_kmers_within_radius, METH_VARARGS,
        "Calculate the number of neighbors with given radius in the graph."
    },

    //
    // tagging / sparse graph functionality
    //

    { "consume_and_tag", (PyCFunction)hashtable_consume_and_tag, METH_VARARGS, "Consume a sequence and tag it" },
    { "get_tags_and_positions", (PyCFunction)hashtable_get_tags_and_positions, METH_VARARGS, "Retrieve tags and their positions in a sequence." },
    { "find_all_tags_list", (PyCFunction)hashtable_find_all_tags_list, METH_VARARGS, "Find all tags within range of the given k-mer, return as list" },
    { "consume_fasta_and_tag", (PyCFunction)hashtable_consume_fasta_and_tag, METH_VARARGS, "Count all k-mers in a given file" },
    { "get_median_count", (PyCFunction)hashtable_get_median_count, METH_VARARGS, "Get the median, average, and stddev of the k-mer counts in the string" },
    { "median_at_least", (PyCFunction)hashtable_median_at_least, METH_VARARGS, "Return true if the median is at least the given cutoff" },
    { "extract_unique_paths", (PyCFunction)hashtable_extract_unique_paths, METH_VARARGS, "" },
    { "print_tagset", (PyCFunction)hashtable_print_tagset, METH_VARARGS, "" },
    { "add_tag", (PyCFunction)hashtable_add_tag, METH_VARARGS, "" },
    { "get_tagset", (PyCFunction)hashtable_get_tagset, METH_VARARGS, "" },
    { "load_tagset", (PyCFunction)hashtable_load_tagset, METH_VARARGS, "" },
    { "save_tagset", (PyCFunction)hashtable_save_tagset, METH_VARARGS, "" },
    { "n_tags", (PyCFunction)hashtable_n_tags, METH_VARARGS, "" },
    { "divide_tags_into_subsets", (PyCFunction)hashtable_divide_tags_into_subsets, METH_VARARGS, "" },
    { "_get_tag_density", (PyCFunction)hashtable__get_tag_density, METH_VARARGS, "" },
    { "_set_tag_density", (PyCFunction)hashtable__set_tag_density, METH_VARARGS, "" },

    // partitioning
    { "do_subset_partition", (PyCFunction)hashtable_do_subset_partition, METH_VARARGS, "" },
    { "find_all_tags", (PyCFunction)hashtable_find_all_tags, METH_VARARGS, "" },
    { "assign_partition_id", (PyCFunction)hashtable_assign_partition_id, METH_VARARGS, "" },
    { "output_partitions", (PyCFunction)hashtable_output_partitions, METH_VARARGS, "" },
    { "find_unpart", (PyCFunction)hashtable_find_unpart, METH_VARARGS, "" },
    { "load_partitionmap", (PyCFunction)hashtable_load_partitionmap, METH_VARARGS, "" },
    { "save_partitionmap", (PyCFunction)hashtable_save_partitionmap, METH_VARARGS, "" },
    { "_validate_partitionmap", (PyCFunction)hashtable__validate_partitionmap, METH_VARARGS, "" },
    { "consume_fasta_and_traverse", (PyCFunction)hashtable_consume_fasta_and_traverse, METH_VARARGS, "" },
    {
        "consume_fasta_and_tag_with_reads_parser", (PyCFunction)hashtable_consume_fasta_and_tag_with_reads_parser,
        METH_VARARGS, "Count all k-mers using a given reads parser"
    },
    { "consume_partitioned_fasta", (PyCFunction)hashtable_consume_partitioned_fasta, METH_VARARGS, "Count all k-mers in a given file" },
    { "join_partitions_by_path", (PyCFunction)hashtable_join_partitions_by_path, METH_VARARGS, "" },
    { "merge_subset", (PyCFunction)hashtable_merge_subset, METH_VARARGS, "" },
    { "merge_subset_from_disk", (PyCFunction)hashtable_merge_from_disk, METH_VARARGS, "" },
    { "count_partitions", (PyCFunction)hashtable_count_partitions, METH_VARARGS, "" },
    { "subset_count_partitions", (PyCFunction)hashtable_subset_count_partitions, METH_VARARGS, "" },
    { "subset_partition_size_distribution", (PyCFunction)hashtable_subset_partition_size_distribution, METH_VARARGS, "" },
    { "save_subset_partitionmap", (PyCFunction)hashtable_save_subset_partitionmap, METH_VARARGS },
    { "load_subset_partitionmap", (PyCFunction)hashtable_load_subset_partitionmap, METH_VARARGS },
    { "_validate_subset_partitionmap", (PyCFunction)hashtable__validate_subset_partitionmap, METH_VARARGS, "" },
    { "set_partition_id", (PyCFunction)hashtable_set_partition_id, METH_VARARGS, "" },
    { "join_partitions", (PyCFunction)hashtable_join_partitions, METH_VARARGS, "" },
    { "get_partition_id", (PyCFunction)hashtable_get_partition_id, METH_VARARGS, "" },
    { "is_single_partition", (PyCFunction)hashtable_is_single_partition, METH_VARARGS, "" },
    { "traverse_from_tags", (PyCFunction)hashtable_traverse_from_tags, METH_VARARGS, "" },
    { "repartition_largest_partition", (PyCFunction)hashtable_repartition_largest_partition, METH_VARARGS, "" },

    // stop tags
    { "load_stop_tags", (PyCFunction)hashtable_load_stop_tags, METH_VARARGS, "" },
    { "save_stop_tags", (PyCFunction)hashtable_save_stop_tags, METH_VARARGS, "" },
    { "print_stop_tags", (PyCFunction)hashtable_print_stop_tags, METH_VARARGS, "" },
    { "trim_on_stoptags", (PyCFunction)hashtable_trim_on_stoptags, METH_VARARGS, "" },
    { "identify_stoptags_by_position", (PyCFunction)hashtable_identify_stoptags_by_position, METH_VARARGS, "" },
    { "filter_if_present", (PyCFunction)hashtable_filter_if_present, METH_VARARGS, "" },
    { "add_stop_tag", (PyCFunction)hashtable_add_stop_tag, METH_VARARGS, "" },
    { "get_stop_tags", (PyCFunction)hashtable_get_stop_tags, METH_VARARGS, "" },
    { "consume_fasta_and_tag_with_stoptags", (PyCFunction)hashtable_consume_fasta_and_tag_with_stoptags, METH_VARARGS, "Count all k-mers in a given file" },

    {NULL, NULL, 0, NULL}           /* sentinel */
};

static PyTypeObject khmer_KHashtable_Type
CPYCHECKER_TYPE_OBJECT_FOR_TYPEDEF("khmer_KHashtable_Object")
= {
    PyVarObject_HEAD_INIT(NULL, 0)       /* init & ob_size */
    "_khmer.KHashtable   ",              /*tp_name*/
    sizeof(khmer_KHashtable_Object)   ,  /*tp_basicsize*/
    0,                                   /*tp_itemsize*/
    0,                                   /*tp_dealloc*/
    0,                                   /*tp_print*/
    0,                                   /*tp_getattr*/
    0,                                   /*tp_setattr*/
    0,                                   /*tp_compare*/
    0,                                   /*tp_repr*/
    0,                                   /*tp_as_number*/
    0,                                   /*tp_as_sequence*/
    0,                                   /*tp_as_mapping*/
    0,                                   /*tp_hash */
    0,                                   /*tp_call*/
    0,                                   /*tp_str*/
    0,                                   /*tp_getattro*/
    0,                                   /*tp_setattro*/
    0,                                   /*tp_as_buffer*/
    Py_TPFLAGS_DEFAULT,                  /*tp_flags*/
    "base hashtable object",             /* tp_doc */
    0,                                   /* tp_traverse */
    0,                                   /* tp_clear */
    0,                                   /* tp_richcompare */
    0,                                   /* tp_weaklistoffset */
    0,                                   /* tp_iter */
    0,                                   /* tp_iternext */
    khmer_hashtable_methods,             /* tp_methods */
    0,                                   /* tp_members */
    0,                                   /* tp_getset */
    0,                                   /* tp_base */
    0,                                   /* tp_dict */
    0,                                   /* tp_descr_get */
    0,                                   /* tp_descr_set */
    0,                                   /* tp_dictoffset */
    0,                                   /* tp_init */
    0,                                   /* tp_alloc */
    0,                                   /* tp_new */
};

#define is_hashtable_obj(v)  (Py_TYPE(v) == &khmer_KHashtable_Type)

//
// KCountingHash object
//

typedef struct {
    khmer_KHashtable_Object khashtable;
    CountingHash * counting;
} khmer_KCountingHash_Object;

typedef struct {
    PyObject_HEAD
    ReadAligner * aligner;
} khmer_ReadAligner_Object;

static void khmer_counting_dealloc(khmer_KCountingHash_Object * obj);

static
PyObject *
count_trim_on_abundance(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * seq = NULL;
    unsigned int min_count_i = 0;

    if (!PyArg_ParseTuple(args, "sI", &seq, &min_count_i)) {
        return NULL;
    }

    unsigned long trim_at;
    Py_BEGIN_ALLOW_THREADS

    BoundedCounterType min_count = min_count_i;

    trim_at = counting->trim_on_abundance(seq, min_count);

    Py_END_ALLOW_THREADS;

    PyObject * trim_seq = PyUnicode_FromStringAndSize(seq, trim_at);
    if (trim_seq == NULL) {
        return NULL;
    }
    PyObject * ret = Py_BuildValue("Ok", trim_seq, trim_at);
    Py_DECREF(trim_seq);

    return ret;
}

static
PyObject *
count_trim_below_abundance(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * seq = NULL;
    BoundedCounterType max_count_i = 0;

    if (!PyArg_ParseTuple(args, "sH", &seq, &max_count_i)) {
        return NULL;
    }

    unsigned long trim_at;
    Py_BEGIN_ALLOW_THREADS

    BoundedCounterType max_count = max_count_i;

    trim_at = counting->trim_below_abundance(seq, max_count);

    Py_END_ALLOW_THREADS;

    PyObject * trim_seq = PyUnicode_FromStringAndSize(seq, trim_at);
    if (trim_seq == NULL) {
        return NULL;
    }
    PyObject * ret = Py_BuildValue("Ok", trim_seq, trim_at);
    Py_DECREF(trim_seq);

    return ret;
}

static
PyObject *
count_find_spectral_error_positions(khmer_KCountingHash_Object * me,
                                    PyObject * args)
{
    khmer::CountingHash * counting = me->counting;

    const char * seq = NULL;
    khmer::BoundedCounterType max_count = 0; // unsigned short int

    if (!PyArg_ParseTuple(args, "sH", &seq, &max_count)) {
        return NULL;
    }

    std::vector<unsigned int> posns;

    try {
        posns = counting->find_spectral_error_positions(seq, max_count);
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return NULL;
    }

    Py_ssize_t posns_size = posns.size();

    PyObject * x = PyList_New(posns_size);
    if (x == NULL) {
        return NULL;
    }
    for (Py_ssize_t i = 0; i < posns_size; i++) {
        PyList_SET_ITEM(x, i, PyLong_FromLong(posns[i]));
    }

    return x;
}

static
PyObject *
count_fasta_dump_kmers_by_abundance(khmer_KCountingHash_Object * me,
                                    PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * inputfile;
    int limit_by = 0;

    if (!PyArg_ParseTuple(args, "si", &inputfile, &limit_by)) {
        return NULL;
    }

    try {
        counting->fasta_dump_kmers_by_abundance(inputfile,
                                                limit_by);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
count_get_raw_tables(khmer_KCountingHash_Object * self, PyObject * args)
{
    CountingHash * counting = self->counting;

    khmer::Byte ** table_ptrs = counting->get_raw_tables();
    std::vector<HashIntoType> sizes = counting->get_tablesizes();

    PyObject * raw_tables = PyList_New(sizes.size());
    for (unsigned int i=0; i<sizes.size(); ++i) {
        Py_buffer buffer;
        int res = PyBuffer_FillInfo(&buffer, NULL, table_ptrs[i], sizes[i], 0,
                                    PyBUF_FULL_RO);
        if (res == -1) {
            return NULL;
        }
        PyObject * buf = PyMemoryView_FromBuffer(&buffer);
        if(!PyMemoryView_Check(buf)) {
            return NULL;
        }
        PyList_SET_ITEM(raw_tables, i, buf);
    }

    return raw_tables;
}

static
PyObject *
count_set_use_bigcount(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    PyObject * x;
    if (!PyArg_ParseTuple(args, "O", &x)) {
        return NULL;
    }
    int setme = PyObject_IsTrue(x);
    if (setme < 0) {
        return NULL;
    }
    counting->set_use_bigcount((bool)setme);

    Py_RETURN_NONE;
}

static
PyObject *
count_get_use_bigcount(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    bool val = counting->get_use_bigcount();

    return PyBool_FromLong((int)val);
}

static
PyObject *
count_get_min_count(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * long_str;

    if (!PyArg_ParseTuple(args, "s", &long_str)) {
        return NULL;
    }

    if (strlen(long_str) < counting->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    BoundedCounterType c = counting->get_min_count(long_str);
    unsigned int N = c;

    return PyLong_FromLong(N);
}

static
PyObject *
count_get_max_count(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * long_str;

    if (!PyArg_ParseTuple(args, "s", &long_str)) {
        return NULL;
    }

    if (strlen(long_str) < counting->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    BoundedCounterType c = counting->get_max_count(long_str);
    unsigned int N = c;

    return PyLong_FromLong(N);
}

static
PyObject *
count_output_fasta_kmer_pos_freq(khmer_KCountingHash_Object * me,
                                 PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * infile;
    const char * outfile;

    if (!PyArg_ParseTuple(args, "ss", &infile, &outfile)) {
        return NULL;
    }

    counting->output_fasta_kmer_pos_freq(infile, outfile);

    return PyLong_FromLong(0);
}

static
PyObject *
count_fasta_count_kmers_by_position(khmer_KCountingHash_Object * me,
                                    PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * inputfile;
    unsigned int max_read_len = 0;
    long max_read_len_long;
    int limit_by_count_int;

    if (!PyArg_ParseTuple(args, "sli", &inputfile, &max_read_len_long,
                          &limit_by_count_int)) {
        return NULL;
    }
    if (max_read_len_long < 0 || max_read_len_long >= pow(2, 32)) {
        PyErr_SetString(
            PyExc_ValueError,
            "The 2nd argument must be positive and less than 2^32");
        return NULL;
    }
    if (limit_by_count_int < 0 || limit_by_count_int >= pow(2, 16)) {
        PyErr_SetString(
            PyExc_ValueError,
            "The 3rd argument must be positive and less than 2^16");
        return NULL;
    }
    max_read_len = (unsigned int) max_read_len_long;

    unsigned long long * counts;
    try {
        counts = counting->fasta_count_kmers_by_position(inputfile,
                 max_read_len,
                 (unsigned short) limit_by_count_int);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    PyObject * x = PyList_New(max_read_len);
    if (x == NULL) {
        delete[] counts;
        return NULL;
    }

    for (unsigned int i = 0; i < max_read_len; i++) {
        int ret = PyList_SetItem(x, i, PyLong_FromUnsignedLongLong(counts[i]));
        if (ret < 0) {
            delete[] counts;
            return NULL;
        }
    }

    delete[] counts;

    return x;
}

static
PyObject *
count_abundance_distribution_with_reads_parser(khmer_KCountingHash_Object * me,
        PyObject * args)
{
    CountingHash * counting = me->counting;

    khmer :: python :: khmer_ReadParser_Object * rparser_obj = NULL;
    khmer_KHashbits_Object *tracking_obj = NULL;

    if (!PyArg_ParseTuple(args, "O!O!", &python::khmer_ReadParser_Type,
                          &rparser_obj, &khmer_KNodegraph_Type, &tracking_obj)) {
        return NULL;
    }

    read_parsers::IParser *rparser      = rparser_obj->parser;
    Hashbits           *hashbits        = tracking_obj->hashbits;
    HashIntoType       *dist            = NULL;
    const char         *value_exception = NULL;
    const char         *file_exception  = NULL;

    Py_BEGIN_ALLOW_THREADS
    try {
        dist = counting->abundance_distribution(rparser, hashbits);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    } catch (khmer_value_exception &exc) {
        value_exception = exc.what();
    }
    Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
        return NULL;
    }

    PyObject * x = PyList_New(MAX_BIGCOUNT + 1);
    if (x == NULL) {
        delete[] dist;
        return NULL;
    }
    for (int i = 0; i < MAX_BIGCOUNT + 1; i++) {
        PyList_SET_ITEM(x, i, PyLong_FromUnsignedLongLong(dist[i]));
    }

    delete[] dist;
    return x;
}

static
PyObject *
count_abundance_distribution(khmer_KCountingHash_Object * me, PyObject * args)
{
    CountingHash * counting = me->counting;

    const char * filename = NULL;
    khmer_KHashbits_Object * tracking_obj = NULL;
    if (!PyArg_ParseTuple(args, "sO!", &filename, &khmer_KNodegraph_Type,
                          &tracking_obj)) {
        return NULL;
    }

    Hashbits           *hashbits        = tracking_obj->hashbits;
    HashIntoType       *dist            = NULL;
    const char         *value_exception = NULL;
    const char         *file_exception  = NULL;
    Py_BEGIN_ALLOW_THREADS
    try {
        dist = counting->abundance_distribution(filename, hashbits);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    } catch (khmer_value_exception &exc) {
        value_exception = exc.what();
    }
    Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        if (dist != NULL) {
            delete []dist;
        }
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
        if (dist != NULL) {
            delete []dist;
        }
        return NULL;
    }

    PyObject * x = PyList_New(MAX_BIGCOUNT + 1);
    if (x == NULL) {
        if (dist != NULL) {
            delete []dist;
        }
        return NULL;
    }
    for (int i = 0; i < MAX_BIGCOUNT + 1; i++) {
        PyList_SET_ITEM(x, i, PyLong_FromUnsignedLongLong(dist[i]));
    }

    if (dist != NULL) {
        delete []dist;
    }

    return x;
}

static
PyObject *
count_do_subset_partition_with_abundance(khmer_KCountingHash_Object * me,
        PyObject * args)
{
    CountingHash * counting = me->counting;

    HashIntoType start_kmer = 0, end_kmer = 0;
    PyObject * break_on_stop_tags_o = NULL;
    PyObject * stop_big_traversals_o = NULL;
    BoundedCounterType min_count, max_count;

    if (!PyArg_ParseTuple(args, "HH|KKOO",
                          &min_count, &max_count,
                          &start_kmer, &end_kmer,
                          &break_on_stop_tags_o,
                          &stop_big_traversals_o)) {
        return NULL;
    }

    bool break_on_stop_tags = false;
    if (break_on_stop_tags_o && PyObject_IsTrue(break_on_stop_tags_o)) {
        break_on_stop_tags = true;
    }
    bool stop_big_traversals = false;
    if (stop_big_traversals_o && PyObject_IsTrue(stop_big_traversals_o)) {
        stop_big_traversals = true;
    }

    SubsetPartition * subset_p = NULL;
    try {
        Py_BEGIN_ALLOW_THREADS
        subset_p = new SubsetPartition(counting);
        subset_p->do_partition_with_abundance(start_kmer, end_kmer,
                                              min_count, max_count,
                                              break_on_stop_tags,
                                              stop_big_traversals);
        Py_END_ALLOW_THREADS
    } catch (std::bad_alloc &e) {
        return PyErr_NoMemory();
    }

    khmer_KSubsetPartition_Object * subset_obj = (khmer_KSubsetPartition_Object *)\
            PyObject_New(khmer_KSubsetPartition_Object, &khmer_KSubsetPartition_Type);

    if (subset_obj == NULL) {
        delete subset_p;
        return NULL;
    }

    subset_obj->subset = subset_p;

    return (PyObject *) subset_obj;
}

static PyMethodDef khmer_counting_methods[] = {
    { "set_use_bigcount", (PyCFunction)count_set_use_bigcount, METH_VARARGS, "" },
    { "get_use_bigcount", (PyCFunction)count_get_use_bigcount, METH_VARARGS, "" },
    { "output_fasta_kmer_pos_freq", (PyCFunction)count_output_fasta_kmer_pos_freq, METH_VARARGS, "" },
    { "get_min_count", (PyCFunction)count_get_min_count, METH_VARARGS, "Get the smallest count of all the k-mers in the string" },
    { "get_max_count", (PyCFunction)count_get_max_count, METH_VARARGS, "Get the largest count of all the k-mers in the string" },
    { "trim_on_abundance", (PyCFunction)count_trim_on_abundance, METH_VARARGS, "Trim on >= abundance" },
    { "trim_below_abundance", (PyCFunction)count_trim_below_abundance, METH_VARARGS, "Trim on >= abundance" },
    { "find_spectral_error_positions", (PyCFunction)count_find_spectral_error_positions, METH_VARARGS, "Identify positions of low-abundance k-mers" },
    { "abundance_distribution", (PyCFunction)count_abundance_distribution, METH_VARARGS, "" },
    { "abundance_distribution_with_reads_parser", (PyCFunction)count_abundance_distribution_with_reads_parser, METH_VARARGS, "" },
    { "fasta_count_kmers_by_position", (PyCFunction)count_fasta_count_kmers_by_position, METH_VARARGS, "" },
    { "fasta_dump_kmers_by_abundance", (PyCFunction)count_fasta_dump_kmers_by_abundance, METH_VARARGS, "" },
    {
        "get_raw_tables", (PyCFunction)count_get_raw_tables,
        METH_VARARGS, "Get a list of the raw tables as memoryview objects"
    },
    { "do_subset_partition_with_abundance", (PyCFunction)count_do_subset_partition_with_abundance, METH_VARARGS, "" },
    {NULL, NULL, 0, NULL}           /* sentinel */
};

static PyObject* _new_counting_hash(PyTypeObject * type, PyObject * args,
                                    PyObject * kwds);

static PyTypeObject khmer_KCountgraph_Type
CPYCHECKER_TYPE_OBJECT_FOR_TYPEDEF("khmer_KCountingHash_Object")
= {
    PyVarObject_HEAD_INIT(NULL, 0)       /* init & ob_size */
    "_khmer.Countgraph",                 /*tp_name*/
    sizeof(khmer_KCountingHash_Object),  /*tp_basicsize*/
    0,                                   /*tp_itemsize*/
    (destructor)khmer_counting_dealloc,  /*tp_dealloc*/
    0,                                   /*tp_print*/
    0,                                   /*tp_getattr*/
    0,                                   /*tp_setattr*/
    0,                                   /*tp_compare*/
    0,                                   /*tp_repr*/
    0,                                   /*tp_as_number*/
    0,                                   /*tp_as_sequence*/
    0,                                   /*tp_as_mapping*/
    0,                                   /*tp_hash */
    0,                                   /*tp_call*/
    0,                                   /*tp_str*/
    0,                                   /*tp_getattro*/
    0,                                   /*tp_setattro*/
    0,                                   /*tp_as_buffer*/
    Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE,                  /*tp_flags*/
    "counting hash object",              /* tp_doc */
    0,                                   /* tp_traverse */
    0,                                   /* tp_clear */
    0,                                   /* tp_richcompare */
    0,                                   /* tp_weaklistoffset */
    0,                                   /* tp_iter */
    0,                                   /* tp_iternext */
    khmer_counting_methods,              /* tp_methods */
    0,                                   /* tp_members */
    0,                                   /* tp_getset */
    0,                                   /* tp_base */
    0,                                   /* tp_dict */
    0,                                   /* tp_descr_get */
    0,                                   /* tp_descr_set */
    0,                                   /* tp_dictoffset */
    0,                                   /* tp_init */
    0,                                   /* tp_alloc */
    _new_counting_hash,                  /* tp_new */
};

#define is_counting_obj(v)  (Py_TYPE(v) == &khmer_KCountgraph_Type)

//
// _new_counting_hash
//

static PyObject* _new_counting_hash(PyTypeObject * type, PyObject * args,
                                    PyObject * kwds)
{
    khmer_KCountingHash_Object * self;

    self = (khmer_KCountingHash_Object *)type->tp_alloc(type, 0);

    if (self != NULL) {
        WordLength k = 0;
        PyListObject * sizes_list_o = NULL;

        if (!PyArg_ParseTuple(args, "bO!", &k, &PyList_Type, &sizes_list_o)) {
            Py_DECREF(self);
            return NULL;
        }

        std::vector<HashIntoType> sizes;
        Py_ssize_t sizes_list_o_length = PyList_GET_SIZE(sizes_list_o);
        if (sizes_list_o_length == -1) {
            Py_DECREF(self);
            PyErr_SetString(PyExc_ValueError, "error with hashtable primes!");
            return NULL;
        }
        for (Py_ssize_t i = 0; i < sizes_list_o_length; i++) {
            PyObject * size_o = PyList_GET_ITEM(sizes_list_o, i);
            if (PyLong_Check(size_o)) {
                sizes.push_back((HashIntoType) PyLong_AsUnsignedLongLong(size_o));
            } else if (PyInt_Check(size_o)) {
                sizes.push_back((HashIntoType) PyInt_AsLong(size_o));
            } else if (PyFloat_Check(size_o)) {
                sizes.push_back((HashIntoType) PyFloat_AS_DOUBLE(size_o));
            } else {
                Py_DECREF(self);
                PyErr_SetString(PyExc_TypeError,
                                "2nd argument must be a list of ints, longs, or floats");
                return NULL;
            }
        }

        try {
            self->counting = new CountingHash(k, sizes);
        } catch (std::bad_alloc &e) {
            Py_DECREF(self);
            return PyErr_NoMemory();
        }
        self->khashtable.hashtable = dynamic_cast<Hashtable*>(self->counting);
    }

    return (PyObject *) self;
}

static
PyObject *
hashbits_update(khmer_KHashbits_Object * me, PyObject * args)
{
    Hashbits * hashbits = me->hashbits;
    Hashbits * other;
    khmer_KHashbits_Object * other_o;

    if (!PyArg_ParseTuple(args, "O!", &khmer_KNodegraph_Type, &other_o)) {
        return NULL;
    }

    other = other_o->hashbits;

    try {
        hashbits->update_from(*other);
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static PyMethodDef khmer_hashbits_methods[] = {
    {
        "update",
        (PyCFunction) hashbits_update, METH_VARARGS,
        "a set update: update this nodegraph with all the entries from the other"
    },
    {NULL, NULL, 0, NULL}           /* sentinel */
};

// __new__ for hashbits; necessary for proper subclassing
// This will essentially do what the old factory function did. Unlike many __new__
// methods, we take our arguments here, because there's no "uninitialized" hashbits
// object; we have to have k and the table sizes before creating the new objects
static PyObject* khmer_hashbits_new(PyTypeObject * type, PyObject * args,
                                    PyObject * kwds)
{
    khmer_KHashbits_Object * self;
    self = (khmer_KHashbits_Object *)type->tp_alloc(type, 0);

    if (self != NULL) {
        WordLength k = 0;
        PyListObject* sizes_list_o = NULL;

        if (!PyArg_ParseTuple(args, "bO!", &k, &PyList_Type, &sizes_list_o)) {
            Py_DECREF(self);
            return NULL;
        }

        std::vector<HashIntoType> sizes;
        Py_ssize_t sizes_list_o_length = PyList_GET_SIZE(sizes_list_o);
        for (Py_ssize_t i = 0; i < sizes_list_o_length; i++) {
            PyObject * size_o = PyList_GET_ITEM(sizes_list_o, i);
            if (PyLong_Check(size_o)) {
                sizes.push_back((HashIntoType) PyLong_AsUnsignedLongLong(size_o));
            } else if (PyInt_Check(size_o)) {
                sizes.push_back((HashIntoType) PyInt_AsLong(size_o));
            } else if (PyFloat_Check(size_o)) {
                sizes.push_back((HashIntoType) PyFloat_AS_DOUBLE(size_o));
            } else {
                Py_DECREF(self);
                PyErr_SetString(PyExc_TypeError,
                                "2nd argument must be a list of ints, longs, or floats");
                return NULL;
            }
        }

        try {
            self->hashbits = new Hashbits(k, sizes);
        } catch (std::bad_alloc &e) {
            Py_DECREF(self);
            return PyErr_NoMemory();
        }
        self->khashtable.hashtable = self->hashbits;
    }
    return (PyObject *) self;
}

#define is_hashbits_obj(v)  (Py_TYPE(v) == &khmer_KNodegraph_Type)

////////////////////////////////////////////////////////////////////////////

static
PyObject *
subset_count_partitions(khmer_KSubsetPartition_Object * me, PyObject * args)
{
    SubsetPartition * subset_p = me->subset;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    size_t n_partitions = 0, n_unassigned = 0;
    subset_p->count_partitions(n_partitions, n_unassigned);

    return Py_BuildValue("nn", (Py_ssize_t) n_partitions,
                         (Py_ssize_t) n_unassigned);
}

static
PyObject *
subset_report_on_partitions(khmer_KSubsetPartition_Object * me, PyObject * args)
{
    SubsetPartition * subset_p = me->subset;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    subset_p->report_on_partitions();

    Py_RETURN_NONE;
}

static
PyObject *
subset_compare_partitions(khmer_KSubsetPartition_Object * me, PyObject * args)
{
    SubsetPartition * subset1_p = me->subset;

    PyObject * subset2_obj = NULL;
    PartitionID pid1, pid2; // @CTB ensure that these are unsigned?

    if (!PyArg_ParseTuple(args, "IOI",
                          &pid1, &subset2_obj, &pid2)) {
        return NULL;
    }

    khmer_KSubsetPartition_Object *other = (khmer_KSubsetPartition_Object *)
                                           subset2_obj;
    SubsetPartition * subset2_p = other->subset;

    unsigned int n_only1 = 0, n_only2 = 0, n_shared = 0;
    subset1_p->compare_to_partition(pid1, subset2_p, pid2,
                                    n_only1, n_only2, n_shared);

    return Py_BuildValue("III", n_only1, n_only2, n_shared);
}

static
PyObject *
subset_partition_size_distribution(khmer_KSubsetPartition_Object * me,
                                   PyObject * args)
{
    SubsetPartition * subset_p = me->subset;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    PartitionCountDistribution d;

    unsigned int n_unassigned = 0;
    subset_p->partition_size_distribution(d, n_unassigned);

    PyObject * x = PyList_New(d.size());
    if (x == NULL) {
        return NULL;
    }
    PartitionCountDistribution::iterator di;

    unsigned int i;
    for (i = 0, di = d.begin(); di != d.end(); ++di, i++) {
        PyObject * tup = Py_BuildValue("KK", di->first, di->second);
        if (tup != NULL) {
            PyList_SET_ITEM(x, i, tup);
        }
        Py_XDECREF(tup);
    }
    if (!(i == d.size())) {
        throw khmer_exception();
    }

    PyObject * ret = Py_BuildValue("OI", x, n_unassigned);
    Py_DECREF(x);
    return ret;
}

static
PyObject *
subset_partition_sizes(khmer_KSubsetPartition_Object * me, PyObject * args)
{
    SubsetPartition * subset_p = me->subset;

    unsigned int min_size = 0;

    if (!PyArg_ParseTuple(args, "|I", &min_size)) {
        return NULL;
    }

    PartitionCountMap cm;
    unsigned int n_unassigned = 0;
    subset_p->partition_sizes(cm, n_unassigned);

    unsigned int i = 0;
    PartitionCountMap::const_iterator mi;
    for (mi = cm.begin(); mi != cm.end(); ++mi) {
        if (mi->second >= min_size) {
            i++;
        }
    }

    PyObject * x = PyList_New(i);
    if (x == NULL) {
        return NULL;
    }

    // this should probably be a dict. @CTB
    for (i = 0, mi = cm.begin(); mi != cm.end(); ++mi) {
        if (mi->second >= min_size) {
            PyObject * tup = Py_BuildValue("II", mi->first, mi->second);
            if (tup != NULL) {
                PyList_SET_ITEM(x, i, tup);
            }
            i++;
        }
    }

    PyObject * ret = Py_BuildValue("OI", x, n_unassigned);
    Py_DECREF(x);

    return ret;
}

static
PyObject *
subset_partition_average_coverages(khmer_KSubsetPartition_Object * me,
                                   PyObject * args)
{
    SubsetPartition * subset_p = me->subset;

    khmer_KCountingHash_Object * counting_o;

    if (!PyArg_ParseTuple(args, "O!", &khmer_KCountgraph_Type, &counting_o)) {
        return NULL;
    }

    PartitionCountMap cm;
    subset_p->partition_average_coverages(cm, counting_o -> counting);

    unsigned int i;
    PartitionCountMap::iterator mi;

    PyObject * x = PyList_New(cm.size());
    if (x == NULL) {
        return NULL;
    }

    // this should probably be a dict. @CTB
    for (i = 0, mi = cm.begin(); mi != cm.end(); ++mi, i++) {
        PyObject * tup = Py_BuildValue("II", mi->first, mi->second);
        if (tup != NULL) {
            PyList_SET_ITEM(x, i, tup);
        }
    }

    return x;
}

static PyMethodDef khmer_subset_methods[] = {
    {
        "count_partitions",
        (PyCFunction)subset_count_partitions,
        METH_VARARGS,
        ""
    },
    {
        "report_on_partitions",
        (PyCFunction)subset_report_on_partitions,
        METH_VARARGS,
        ""
    },
    {
        "compare_partitions",
        (PyCFunction)subset_compare_partitions,
        METH_VARARGS,
        ""
    },
    {
        "partition_size_distribution",
        (PyCFunction)subset_partition_size_distribution,
        METH_VARARGS,
        ""
    },
    {
        "partition_sizes",
        (PyCFunction)subset_partition_sizes,
        METH_VARARGS,
        ""
    },
    {
        "partition_average_coverages",
        (PyCFunction)subset_partition_average_coverages,
        METH_VARARGS,
        ""
    },
    {NULL, NULL, 0, NULL}           /* sentinel */
};

typedef struct {
    PyObject_HEAD
    LabelHash * labelhash;
} khmer_KGraphLabels_Object;

static PyObject * khmer_graphlabels_new(PyTypeObject * type, PyObject *args,
                                        PyObject *kwds);

#define is_graphlabels_obj(v)  (Py_TYPE(v) == &khmer_KGraphLabels_Type)

static void khmer_graphlabels_dealloc(khmer_KGraphLabels_Object * obj)
{
    delete obj->labelhash;
    obj->labelhash = NULL;

    Py_TYPE(obj)->tp_free((PyObject*)obj);
}

static PyObject * khmer_graphlabels_new(PyTypeObject *type, PyObject *args,
                                        PyObject *kwds)
{
    khmer_KGraphLabels_Object *self;
    self = (khmer_KGraphLabels_Object*)type->tp_alloc(type, 0);

    if (self != NULL) {
        PyObject * hashtable_o;
        khmer::Hashtable * hashtable = NULL;

        if (!PyArg_ParseTuple(args, "O", &hashtable_o)) {
            Py_DECREF(self);
            return NULL;
        }

        if (PyObject_TypeCheck(hashtable_o, &khmer_KNodegraph_Type)) {
            khmer_KHashbits_Object * kho = (khmer_KHashbits_Object *) hashtable_o;
            hashtable = kho->hashbits;
        } else if (PyObject_TypeCheck(hashtable_o, &khmer_KCountgraph_Type)) {
            khmer_KCountingHash_Object * cho = (khmer_KCountingHash_Object *) hashtable_o;
            hashtable = cho->counting;
        } else {
            PyErr_SetString(PyExc_ValueError,
                            "graph object must be a NodeGraph or CountGraph");
            Py_DECREF(self);
            return NULL;
        }

        try {
            self->labelhash = new LabelHash(hashtable);
        } catch (std::bad_alloc &e) {
            Py_DECREF(self);
            return PyErr_NoMemory();
        }
    }

    return (PyObject *) self;
}

static
PyObject *
labelhash_get_label_dict(khmer_KGraphLabels_Object * me, PyObject * args)
{
    LabelHash * hb = me->labelhash;

    PyObject * d = PyDict_New();
    if (d == NULL) {
        return NULL;
    }
    LabelPtrMap::iterator it;

    for (it = hb->label_ptrs.begin(); it != hb->label_ptrs.end(); ++it) {
        PyObject * key = Py_BuildValue("K", it->first);
        PyObject * val = Py_BuildValue("K", it->second);
        if (key != NULL && val != NULL) {
            PyDict_SetItem(d, key, val);
        }
        Py_XDECREF(key);
        Py_XDECREF(val);
    }

    return d;
}

static
PyObject *
labelhash_consume_fasta_and_tag_with_labels(khmer_KGraphLabels_Object * me,
        PyObject * args)
{
    LabelHash * hb = me->labelhash;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    const char         *value_exception = NULL;
    const char         *file_exception  = NULL;
    unsigned long long  n_consumed      = 0;
    unsigned int        total_reads     = 0;
    //Py_BEGIN_ALLOW_THREADS
    try {
        hb->consume_fasta_and_tag_with_labels(filename, total_reads,
                                              n_consumed);
    } catch (khmer_file_exception &exc) {
        file_exception = exc.what();
    } catch (khmer_value_exception &exc) {
        value_exception = exc.what();
    }
    //Py_END_ALLOW_THREADS

    if (file_exception != NULL) {
        PyErr_SetString(PyExc_OSError, file_exception);
        return NULL;
    }
    if (value_exception != NULL) {
        PyErr_SetString(PyExc_ValueError, value_exception);
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
labelhash_consume_partitioned_fasta_and_tag_with_labels(
    khmer_KGraphLabels_Object * me, PyObject * args)
{
    LabelHash * labelhash = me->labelhash;

    const char * filename;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    // call the C++ function, and trap signals => Python

    unsigned long long  n_consumed  = 0;
    unsigned int        total_reads = 0;

    try {
        labelhash->consume_partitioned_fasta_and_tag_with_labels(filename,
                total_reads, n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static
PyObject *
labelhash_consume_sequence_and_tag_with_labels(khmer_KGraphLabels_Object * me,
        PyObject * args)
{
    LabelHash * hb = me->labelhash;
    const char * seq = NULL;
    unsigned long long c = 0;
    if (!PyArg_ParseTuple(args, "sK", &seq, &c)) {
        return NULL;
    }
    unsigned long long n_consumed = 0;
    Label * the_label = hb->check_and_allocate_label(c);

    hb->consume_sequence_and_tag_with_labels(seq, n_consumed, *the_label);
    return Py_BuildValue("K", n_consumed);
}

static
PyObject *
labelhash_sweep_label_neighborhood(khmer_KGraphLabels_Object * me,
                                   PyObject * args)
{
    LabelHash * hb = me->labelhash;

    const char * seq = NULL;
    int r = 0;
    PyObject * break_on_stop_tags_o = NULL;
    PyObject * stop_big_traversals_o = NULL;

    if (!PyArg_ParseTuple(args, "s|iOO", &seq, &r,
                          &break_on_stop_tags_o,
                          &stop_big_traversals_o)) {
        return NULL;
    }

    unsigned int range = (2 * hb->graph->_get_tag_density()) + 1;
    if (r >= 0) {
        range = r;
    }

    bool break_on_stop_tags = false;
    if (break_on_stop_tags_o && PyObject_IsTrue(break_on_stop_tags_o)) {
        break_on_stop_tags = true;
    }
    bool stop_big_traversals = false;
    if (stop_big_traversals_o && PyObject_IsTrue(stop_big_traversals_o)) {
        stop_big_traversals = true;
    }

    if (strlen(seq) < hb->graph->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    //std::pair<TagLabelPtrPair::iterator, TagLabelPtrPair::iterator> ret;
    LabelPtrSet found_labels;

    //unsigned int num_traversed = 0;
    //Py_BEGIN_ALLOW_THREADS
    hb->sweep_label_neighborhood(seq, found_labels, range, break_on_stop_tags,
                                 stop_big_traversals);
    //Py_END_ALLOW_THREADS

    //printf("...%u kmers traversed\n", num_traversed);

    PyObject * x =  PyList_New(found_labels.size());
    LabelPtrSet::const_iterator si;
    unsigned long long i = 0;
    for (si = found_labels.begin(); si != found_labels.end(); ++si) {
        PyList_SET_ITEM(x, i, Py_BuildValue("K", *(*si)));
        i++;
    }

    return x;
}

// Similar to find_all_tags, but returns tags in a way actually usable by python
// need a tags_in_sequence iterator or function in c++ land for reuse in all
// these functions

static
PyObject *
labelhash_sweep_tag_neighborhood(khmer_KGraphLabels_Object * me,
                                 PyObject * args)
{
    LabelHash * labelhash = me->labelhash;

    const char * seq = NULL;
    int r = 0;
    PyObject * break_on_stop_tags_o = NULL;
    PyObject * stop_big_traversals_o = NULL;

    if (!PyArg_ParseTuple(args, "s|iOO", &seq, &r,
                          &break_on_stop_tags_o,
                          &stop_big_traversals_o)) {
        return NULL;
    }

    unsigned int range = (2 * labelhash->graph->_get_tag_density()) + 1;
    if (r >= 0) {
        range = r;
    }

    bool break_on_stop_tags = false;
    if (break_on_stop_tags_o && PyObject_IsTrue(break_on_stop_tags_o)) {
        break_on_stop_tags = true;
    }
    bool stop_big_traversals = false;
    if (stop_big_traversals_o && PyObject_IsTrue(stop_big_traversals_o)) {
        stop_big_traversals = true;
    }

    if (strlen(seq) < labelhash->graph->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }

    SeenSet tagged_kmers;

    //Py_BEGIN_ALLOW_THREADS

    labelhash->graph->partition->sweep_for_tags(seq, tagged_kmers,
            labelhash->graph->all_tags,
            range, break_on_stop_tags,
            stop_big_traversals);

    //Py_END_ALLOW_THREADS

    PyObject * x =  PyList_New(tagged_kmers.size());
    if (x == NULL) {
        return NULL;
    }
    SeenSet::iterator si;
    unsigned long long i = 0;
    for (si = tagged_kmers.begin(); si != tagged_kmers.end(); ++si) {
        //std::string kmer_s = _revhash(*si, labelhash->ksize());
        // type K for python unsigned long long
        PyList_SET_ITEM(x, i, Py_BuildValue("K", *si));
        i++;
    }

    return x;
}

static
PyObject *
labelhash_get_tag_labels(khmer_KGraphLabels_Object * me, PyObject * args)
{
    LabelHash * labelhash = me->labelhash;

    HashIntoType tag;

    if (!PyArg_ParseTuple(args, "K", &tag)) {
        return NULL;
    }

    LabelPtrSet labels;

    labels = labelhash->get_tag_labels(tag);

    PyObject * x =  PyList_New(labels.size());
    LabelPtrSet::const_iterator si;
    unsigned long long i = 0;
    for (si = labels.begin(); si != labels.end(); ++si) {
        //std::string kmer_s = _revhash(*si, labelhash->ksize());
        PyList_SET_ITEM(x, i, Py_BuildValue("K", *(*si)));
        i++;
    }

    return x;
}

static
PyObject *
labelhash_n_labels(khmer_KGraphLabels_Object * me, PyObject * args)
{
    LabelHash * labelhash = me->labelhash;

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }

    return PyLong_FromSize_t(labelhash->n_labels());
}

static
PyObject *
labelhash_save_labels_and_tags(khmer_KGraphLabels_Object * me, PyObject * args)
{
    const char * filename = NULL;
    LabelHash * labelhash = me->labelhash;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        labelhash->save_labels_and_tags(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
labelhash_load_labels_and_tags(khmer_KGraphLabels_Object * me, PyObject * args)
{
    const char * filename = NULL;
    LabelHash * labelhash = me->labelhash;

    if (!PyArg_ParseTuple(args, "s", &filename)) {
        return NULL;
    }

    try {
        labelhash->load_labels_and_tags(filename);
    } catch (khmer_file_exception &e) {
        PyErr_SetString(PyExc_OSError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static PyMethodDef khmer_graphlabels_methods[] = {
    { "consume_fasta_and_tag_with_labels", (PyCFunction)labelhash_consume_fasta_and_tag_with_labels, METH_VARARGS, "" },
    { "sweep_label_neighborhood", (PyCFunction)labelhash_sweep_label_neighborhood, METH_VARARGS, "" },
    {"consume_partitioned_fasta_and_tag_with_labels", (PyCFunction)labelhash_consume_partitioned_fasta_and_tag_with_labels, METH_VARARGS, "" },
    {"sweep_tag_neighborhood", (PyCFunction)labelhash_sweep_tag_neighborhood, METH_VARARGS, "" },
    {"get_tag_labels", (PyCFunction)labelhash_get_tag_labels, METH_VARARGS, ""},
    {"consume_sequence_and_tag_with_labels", (PyCFunction)labelhash_consume_sequence_and_tag_with_labels, METH_VARARGS, "" },
    {"n_labels", (PyCFunction)labelhash_n_labels, METH_VARARGS, ""},
    {"get_label_dict", (PyCFunction)labelhash_get_label_dict, METH_VARARGS, "" },
    { "save_labels_and_tags", (PyCFunction)labelhash_save_labels_and_tags, METH_VARARGS, "" },
    { "load_labels_and_tags", (PyCFunction)labelhash_load_labels_and_tags, METH_VARARGS, "" },    {NULL, NULL, 0, NULL}           /* sentinel */
};

static PyTypeObject khmer_KGraphLabels_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)  /* init & ob_size */
    "_khmer.LabelHash",            /* tp_name */
    sizeof(khmer_KGraphLabels_Object), /* tp_basicsize */
    0,                       /* tp_itemsize */
    (destructor)khmer_graphlabels_dealloc, /* tp_dealloc */
    0,                       /* tp_print */
    0,                       /* tp_getattr */
    0,                       /* tp_setattr */
    0,                       /* tp_compare */
    0,                       /* tp_repr */
    0,                       /* tp_as_number */
    0,                       /* tp_as_sequence */
    0,                       /* tp_as_mapping */
    0,                       /* tp_hash */
    0,                       /* tp_call */
    0,                       /* tp_str */
    0,                       /* tp_getattro */
    0,                       /* tp_setattro */
    0,                       /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE,   /* tp_flags */
    0,                       /* tp_doc */
    0,                       /* tp_traverse */
    0,                       /* tp_clear */
    0,                       /* tp_richcompare */
    0,                       /* tp_weaklistoffset */
    0,                       /* tp_iter */
    0,                       /* tp_iternext */
    khmer_graphlabels_methods, /* tp_methods */
    0,                       /* tp_members */
    0,                       /* tp_getset */
    0,                       /* tp_base */
    0,                       /* tp_dict */
    0,                       /* tp_descr_get */
    0,                       /* tp_descr_set */
    0,                       /* tp_dictoffset */
    0,                       /* tp_init */
    0,                       /* tp_alloc */
    khmer_graphlabels_new,      /* tp_new */
};

static
PyObject *
hashtable_traverse_from_tags(khmer_KHashtable_Object * me, PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    khmer_KCountingHash_Object * counting_o = NULL;
    unsigned int distance, threshold, frequency;

    if (!PyArg_ParseTuple(args, "O!III", &khmer_KCountgraph_Type, &counting_o,
                          &distance, &threshold, &frequency)) {
        return NULL;
    }

    hashtable->traverse_from_tags(distance, threshold, frequency,
                                  * counting_o->counting);

    Py_RETURN_NONE;
}

static
PyObject *
hashtable_repartition_largest_partition(khmer_KHashtable_Object * me,
                                        PyObject * args)
{
    Hashtable * hashtable = me->hashtable;
    khmer_KCountingHash_Object * counting_o = NULL;
    PyObject * subset_o = NULL;
    SubsetPartition * subset_p;
    unsigned int distance, threshold, frequency;

    if (!PyArg_ParseTuple(args, "OO!III",
                          &subset_o,
                          &khmer_KCountgraph_Type, &counting_o,
                          &distance, &threshold, &frequency)) {
        return NULL;
    }

    if (PyObject_TypeCheck(subset_o, &khmer_KSubsetPartition_Type)) {
        subset_p = ((khmer_KSubsetPartition_Object *) subset_o)->subset;
    } else {
        subset_p = hashtable->partition;
    }

    CountingHash * counting = counting_o->counting;

    unsigned long next_largest;
    try {
        next_largest = subset_p->repartition_largest_partition(distance,
                       threshold, frequency, *counting);
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_RuntimeError, e.what());
        return NULL;
    }

    return PyLong_FromLong(next_largest);
}

static PyObject * readaligner_align(khmer_ReadAligner_Object * me,
                                    PyObject * args)
{
    const char * read;

    if (!PyArg_ParseTuple(args, "s", &read)) {
        return NULL;
    }

    /*if (strlen(read) < (unsigned int)aligner->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }*/

    Alignment * aln = me->aligner->Align(read);

    const char* alignment = aln->graph_alignment.c_str();
    const char* readAlignment = aln->read_alignment.c_str();
    PyObject * ret = Py_BuildValue("dssO", aln->score, alignment,
                                   readAlignment, (aln->truncated)? Py_True : Py_False);
    delete aln;

    return ret;
}

static PyObject * readaligner_align_forward(khmer_ReadAligner_Object * me,
        PyObject * args)
{
    ReadAligner * aligner = me->aligner;

    const char * read;

    if (!PyArg_ParseTuple(args, "s", &read)) {
        return NULL;
    }

    /*if (strlen(read) < (unsigned int)aligner->ksize()) {
        PyErr_SetString(PyExc_ValueError,
                        "string length must >= the hashtable k-mer size");
        return NULL;
    }*/

    Alignment * aln;
    aln = aligner->AlignForward(read);

    const char* alignment = aln->graph_alignment.c_str();
    const char* readAlignment = aln->read_alignment.c_str();
    PyObject * x = PyList_New(aln->covs.size());
    for (size_t i = 0; i < aln->covs.size(); i++ ) {
        PyList_SET_ITEM(x, i, PyLong_FromLong(aln->covs[i]));
    }

    PyObject * ret = Py_BuildValue("dssOO", aln->score, alignment,
                                   readAlignment,
                                   (aln->truncated)? Py_True : Py_False,
                                   x);
    delete aln;
    Py_DECREF(x);

    return ret;
}

static PyObject* khmer_ReadAligner_get_scoring_matrix(
    khmer_ReadAligner_Object * me, PyObject * args)
{

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }
    ScoringMatrix matrix = me->aligner->getScoringMatrix();

    return Py_BuildValue( "dddd", matrix.trusted_match, matrix.trusted_mismatch,
                          matrix.untrusted_match, matrix.untrusted_mismatch);
}

static PyObject* khmer_ReadAligner_get_transition_probabilities(
    khmer_ReadAligner_Object * me, PyObject * args)
{

    if (!PyArg_ParseTuple(args, "")) {
        return NULL;
    }
    ScoringMatrix matrix = me->aligner->getScoringMatrix();

    return Py_BuildValue( "(dddddd)(dddd)(dddd)(dddddd)(dddd)(dddd)",
                          matrix.tsc[0], matrix.tsc[1], matrix.tsc[2],
                          matrix.tsc[3], matrix.tsc[4], matrix.tsc[5],
                          matrix.tsc[6], matrix.tsc[7], matrix.tsc[8],
                          matrix.tsc[9], matrix.tsc[10], matrix.tsc[11],
                          matrix.tsc[12], matrix.tsc[13], matrix.tsc[14],
                          matrix.tsc[15], matrix.tsc[16], matrix.tsc[17],
                          matrix.tsc[18], matrix.tsc[19], matrix.tsc[20],
                          matrix.tsc[21], matrix.tsc[22], matrix.tsc[23],
                          matrix.tsc[24], matrix.tsc[25], matrix.tsc[26],
                          matrix.tsc[27]);
}

static PyMethodDef khmer_ReadAligner_methods[] = {
    {"align", (PyCFunction)readaligner_align, METH_VARARGS, ""},
    {"align_forward", (PyCFunction)readaligner_align_forward, METH_VARARGS, ""},
    {
        "get_scoring_matrix", (PyCFunction)khmer_ReadAligner_get_scoring_matrix,
        METH_VARARGS,
        "Get the scoring matrix in use.\n\n\
Returns a tuple of floats: (trusted_match, trusted_mismatch, untrusted_match, \
untrusted_mismatch)"
    },
    {
        "get_transition_probabilities",
        (PyCFunction)khmer_ReadAligner_get_transition_probabilities,
        METH_VARARGS,
        "Get the transition probabilties in use.\n\n\
HMM state notation abbreviations:\n\
    M_t - trusted match; M_u - untrusted match\n\
    Ir_t - trusted read insert; Ir_u - untrusted read insert\n\
    Ig_t - trusted graph insert; Ig_u - untrusted graph insert\n\
\
Returns a sparse matrix as a tuple of six tuples.\n\
The inner tuples contain 6, 4, 4, 6, 4, and 4 floats respectively.\n\
Transition are notated as 'StartState-NextState':\n\
(\n\
  ( M_t-M_t,  M_t-Ir_t,  M_t-Ig_t,  M_t-M_u,  M_t-Ir_u,  M_t-Ig_u),\n\
  (Ir_t-M_t, Ir_t-Ir_t,            Ir_t-M_u, Ir_t-Ir_u           ),\n\
  (Ig_t-M_t,          , Ig_t-Ig_t, Ig_t-M_u,            Ig_t-Ig_u),\n\
  ( M_u-M_t,  M_u-Ir_t,  M_u-Ig_t,  M_u-M_u,  M_u-Ir_u,  M_u-Ig_u),\n\
  (Ir_u-M_t, Ir_u-Ir_t,            Ir_u-M_u, Ir_u-Ir_u           ),\n\
  (Ig_u-M_t,          , Ig_u-Ig_t, Ig_u-M_u,            Ig_u-Ig_u)\n\
)"
    },
    {NULL} /* Sentinel */
};

//
// khmer_readaligner_dealloc -- clean up readaligner object
// GRAPHALIGN addition
//
static void khmer_readaligner_dealloc(khmer_ReadAligner_Object* obj)
{
    delete obj->aligner;
    obj->aligner = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}

//
// new_readaligner
//
static PyObject* khmer_ReadAligner_new(PyTypeObject *type, PyObject * args,
                                       PyObject *kwds)
{
    khmer_ReadAligner_Object * self;

    self = (khmer_ReadAligner_Object *)type->tp_alloc(type, 0);

    if (self != NULL) {
        khmer_KCountingHash_Object * ch = NULL;
        unsigned short int trusted_cov_cutoff = 2;
        double bits_theta = 1;
        double scoring_matrix[] = { 0, 0, 0, 0 };
        double * transitions = new double[28];

        if(!PyArg_ParseTuple(
                    args,
                    "O!Hd|(dddd)((dddddd)(dddd)(dddd)(dddddd)(dddd)(dddd))",
                    &khmer_KCountgraph_Type, &ch, &trusted_cov_cutoff,
                    &bits_theta, &scoring_matrix[0], &scoring_matrix[1],
                    &scoring_matrix[2], &scoring_matrix[3], &transitions[0],
                    &transitions[1], &transitions[2], &transitions[3],
                    &transitions[4], &transitions[5], &transitions[6],
                    &transitions[7], &transitions[8], &transitions[9],
                    &transitions[10], &transitions[11], &transitions[12],
                    &transitions[13], &transitions[14], &transitions[15],
                    &transitions[16], &transitions[17], &transitions[18],
                    &transitions[19], &transitions[20], &transitions[21],
                    &transitions[22], &transitions[23], &transitions[24],
                    &transitions[25], &transitions[26], &transitions[27])) {
            Py_DECREF(self);
            return NULL;
        }

        self->aligner = new ReadAligner(ch->counting, trusted_cov_cutoff,
                                        bits_theta, scoring_matrix,
                                        transitions);
    }

    return (PyObject *) self;
}

static PyTypeObject khmer_ReadAlignerType = {
    PyVarObject_HEAD_INIT(NULL, 0) /* init & ob_size */
    "_khmer.ReadAligner",		    /*tp_name*/
    sizeof(khmer_ReadAligner_Object),	    /*tp_basicsize*/
    0,					    /*tp_itemsize*/
    (destructor)khmer_readaligner_dealloc,  /*tp_dealloc*/
    0,                          /*tp_print*/
    0,                          /*tp_getattr*/
    0,                          /*tp_setattr*/
    0,                          /*tp_compare*/
    0,                          /*tp_repr*/
    0,                          /*tp_as_number*/
    0,                          /*tp_as_sequence*/
    0,                          /*tp_as_mapping*/
    0,                          /*tp_hash */
    0,                          /*tp_call*/
    0,                          /*tp_str*/
    0,                          /*tp_getattro*/
    0,                          /*tp_setattro*/
    0,                          /*tp_as_buffer*/
    Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE,         /*tp_flags*/
    "ReadAligner object",           /* tp_doc */
    0,                         /* tp_traverse */
    0,                         /* tp_clear */
    0,                         /* tp_richcompare */
    0,                         /* tp_weaklistoffset */
    0,                         /* tp_iter */
    0,                         /* tp_iternext */
    khmer_ReadAligner_methods, /* tp_methods */
    0,                         /* tp_members */
    0,                         /* tp_getset */
    0,                         /* tp_base */
    0,                         /* tp_dict */
    0,                         /* tp_descr_get */
    0,                         /* tp_descr_set */
    0,                         /* tp_dictoffset */
    0,			               /* tp_init */
    0,                         /* tp_alloc */
    khmer_ReadAligner_new,     /* tp_new */
};

static
PyObject *
hashtable_consume_fasta_and_traverse(khmer_KHashtable_Object * me,
                                     PyObject * args)
{
    Hashtable * hashtable = me->hashtable;

    const char * filename;
    unsigned int radius, big_threshold, transfer_threshold;
    khmer_KCountingHash_Object * counting_o = NULL;

    if (!PyArg_ParseTuple(args, "sIIIO!", &filename,
                          &radius, &big_threshold, &transfer_threshold,
                          &khmer_KCountgraph_Type, &counting_o)) {
        return NULL;
    }

    CountingHash * counting = counting_o->counting;

    try {
        hashtable->consume_fasta_and_traverse(filename, radius, big_threshold,
                                              transfer_threshold, *counting);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

//
// khmer_counting_dealloc -- clean up a counting hash object.
//

static void khmer_counting_dealloc(khmer_KCountingHash_Object * obj)
{
    delete obj->counting;
    obj->counting = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}

//
// khmer_hashbits_dealloc -- clean up a hashbits object.
//
static void khmer_hashbits_dealloc(khmer_KHashbits_Object * obj)
{
    delete obj->hashbits;
    obj->hashbits = NULL;

    Py_TYPE(obj)->tp_free((PyObject*)obj);
}


//
// khmer_subset_dealloc -- clean up a hashbits object.
//

static void khmer_subset_dealloc(khmer_KSubsetPartition_Object * obj)
{
    delete obj->subset;
    obj->subset = NULL;
    Py_TYPE(obj)->tp_free((PyObject*)obj);
}


/***********************************************************************/

//
// KHLLCounter object
//

typedef struct {
    PyObject_HEAD
    khmer::HLLCounter * hllcounter;
} khmer_KHLLCounter_Object;

static PyObject* khmer_hllcounter_new(PyTypeObject * type, PyObject * args,
                                      PyObject * kwds)
{
    khmer_KHLLCounter_Object * self;
    self = (khmer_KHLLCounter_Object *)type->tp_alloc(type, 0);

    if (self != NULL) {
        double error_rate = 0.01;
        WordLength ksize = 20;

        if (!PyArg_ParseTuple(args, "|db", &error_rate, &ksize)) {
            Py_DECREF(self);
            return NULL;
        }

        try {
            self->hllcounter = new HLLCounter(error_rate, ksize);
        } catch (InvalidValue &e) {
            Py_DECREF(self);
            PyErr_SetString(PyExc_ValueError, e.what());
            return NULL;
        }
    }

    return (PyObject *) self;
}

//
// khmer_hllcounter_dealloc -- clean up a hllcounter object.
//

static void khmer_hllcounter_dealloc(khmer_KHLLCounter_Object * obj)
{
    delete obj->hllcounter;
    obj->hllcounter = NULL;

    Py_TYPE(obj)->tp_free((PyObject*)obj);
}

static
PyObject *
hllcounter_add(khmer_KHLLCounter_Object * me, PyObject * args)
{
    const char * kmer_str;

    if (!PyArg_ParseTuple(args, "s", &kmer_str)) {
        return NULL;
    }

    try {
        me->hllcounter->add(kmer_str);
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

static
PyObject *
hllcounter_estimate_cardinality(khmer_KHLLCounter_Object * me, PyObject * args)
{
    if (!PyArg_ParseTuple( args, "" )) {
        return NULL;
    }

    return PyLong_FromLong(me->hllcounter->estimate_cardinality());
}

static
PyObject *
hllcounter_consume_string(khmer_KHLLCounter_Object * me, PyObject * args)
{
    const char * kmer_str;
    unsigned long long n_consumed;

    if (!PyArg_ParseTuple(args, "s", &kmer_str)) {
        return NULL;
    }

    try {
        n_consumed = me->hllcounter->consume_string(kmer_str);
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return NULL;
    }

    return PyLong_FromLong(n_consumed);
}

static PyObject * hllcounter_consume_fasta(khmer_KHLLCounter_Object * me,
        PyObject * args, PyObject * kwds)
{
    const char * filename;
    PyObject * stream_records_o = NULL;

    static const char* const_kwlist[] = {"filename", "stream_records", NULL};
    static char** kwlist = const_cast<char**>(const_kwlist);

    bool stream_records = false;

    if (!PyArg_ParseTupleAndKeywords(args, kwds, "s|O", kwlist,
                                     &filename, &stream_records_o)) {
        return NULL;
    }

    if (stream_records_o != NULL && PyObject_IsTrue(stream_records_o)) {
        stream_records = true;
    }

    // call the C++ function, and trap signals => Python
    unsigned long long  n_consumed    = 0;
    unsigned int        total_reads   = 0;
    try {
        me->hllcounter->consume_fasta(filename, stream_records, total_reads,
                                      n_consumed);
    } catch (khmer_file_exception &exc) {
        PyErr_SetString(PyExc_OSError, exc.what());
        return NULL;
    } catch (khmer_value_exception &exc) {
        PyErr_SetString(PyExc_ValueError, exc.what());
        return NULL;
    }

    return Py_BuildValue("IK", total_reads, n_consumed);
}

static PyObject * hllcounter_merge(khmer_KHLLCounter_Object * me,
                                   PyObject * args);

static
PyObject *
hllcounter_get_erate(khmer_KHLLCounter_Object * me)
{
    return PyFloat_FromDouble(me->hllcounter->get_erate());
}

static
PyObject *
hllcounter_get_ksize(khmer_KHLLCounter_Object * me)
{
    return PyLong_FromLong(me->hllcounter->get_ksize());
}

static
int
hllcounter_set_ksize(khmer_KHLLCounter_Object * me, PyObject *value,
                     void *closure)
{
    if (value == NULL) {
        PyErr_SetString(PyExc_TypeError, "Cannot delete attribute");
        return -1;
    }

    long ksize = 0;
    if (PyLong_Check(value)) {
        ksize = PyLong_AsLong(value);
    } else if (PyInt_Check(value)) {
        ksize = PyInt_AsLong(value);
    } else {
        PyErr_SetString(PyExc_TypeError,
                        "Please use an integer value for k-mer size");
        return -1;
    }

    if (ksize <= 0) {
        PyErr_SetString(PyExc_ValueError, "Please set k-mer size to a value "
                        "greater than zero");
        return -1;
    }

    try {
        me->hllcounter->set_ksize(ksize);
    } catch (ReadOnlyAttribute &e) {
        PyErr_SetString(PyExc_AttributeError, e.what());
        return -1;
    }

    return 0;
}

static
int
hllcounter_set_erate(khmer_KHLLCounter_Object * me, PyObject *value,
                     void *closure)
{
    if (value == NULL) {
        PyErr_SetString(PyExc_TypeError, "Cannot delete attribute");
        return -1;
    }

    if (!PyFloat_Check(value)) {
        PyErr_SetString(PyExc_TypeError,
                        "Please use a float value for k-mer size");
        return -1;
    }

    double erate = PyFloat_AsDouble(value);
    try {
        me->hllcounter->set_erate(erate);
    } catch (InvalidValue &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return -1;
    } catch (ReadOnlyAttribute &e) {
        PyErr_SetString(PyExc_AttributeError, e.what());
        return -1;
    }

    return 0;
}

static
PyObject *
hllcounter_getalpha(khmer_KHLLCounter_Object * me)
{
    return PyFloat_FromDouble(me->hllcounter->get_alpha());
}

static
PyObject *
hllcounter_getcounters(khmer_KHLLCounter_Object * me)
{
    std::vector<int> counters = me->hllcounter->get_M();

    PyObject * x = PyList_New(counters.size());
    for (size_t i = 0; i < counters.size(); i++) {
        PyList_SET_ITEM(x, i, PyLong_FromLong(counters[i]));
    }

    return x;
}

static PyMethodDef khmer_hllcounter_methods[] = {
    {
        "add", (PyCFunction)hllcounter_add,
        METH_VARARGS,
        "Add a k-mer to the counter."
    },
    {
        "estimate_cardinality", (PyCFunction)hllcounter_estimate_cardinality,
        METH_VARARGS,
        "Return the current estimation."
    },
    {
        "consume_string", (PyCFunction)hllcounter_consume_string,
        METH_VARARGS,
        "Break a sequence into k-mers and add each k-mer to the counter."
    },
    {
        "consume_fasta", (PyCFunction)hllcounter_consume_fasta,
        METH_VARARGS | METH_KEYWORDS,
        "Read sequences from file, break into k-mers, "
        "and add each k-mer to the counter. If optional keyword 'stream_out' "
        "is True, also prints each sequence to stdout."
    },
    {
        "merge", (PyCFunction)hllcounter_merge,
        METH_VARARGS,
        "Merge other counter into this one."
    },
    {NULL} /* Sentinel */
};

static PyGetSetDef khmer_hllcounter_getseters[] = {
    {
        (char *)"alpha",
        (getter)hllcounter_getalpha, NULL,
        (char *)"alpha constant for this HLL counter.",
        NULL
    },
    {
        (char *)"error_rate",
        (getter)hllcounter_get_erate, (setter)hllcounter_set_erate,
        (char *)"Error rate for this HLL counter. "
        "Can be changed prior to first counting, but becomes read-only after "
        "that (raising AttributeError)",
        NULL
    },
    {
        (char *)"ksize",
        (getter)hllcounter_get_ksize, (setter)hllcounter_set_ksize,
        (char *)"k-mer size for this HLL counter."
        "Can be changed prior to first counting, but becomes read-only after "
        "that (raising AttributeError)",
        NULL
    },
    {
        (char *)"counters",
        (getter)hllcounter_getcounters, NULL,
        (char *)"Read-only internal counters.",
        NULL
    },
    {NULL} /* Sentinel */
};

static PyTypeObject khmer_KHLLCounter_Type = {
    PyVarObject_HEAD_INIT(NULL, 0)
    "_khmer.KHLLCounter",                       /* tp_name */
    sizeof(khmer_KHLLCounter_Object),          /* tp_basicsize */
    0,                                         /* tp_itemsize */
    (destructor)khmer_hllcounter_dealloc,      /* tp_dealloc */
    0,                                         /* tp_print */
    0,                                         /* tp_getattr */
    0,                                         /* tp_setattr */
    0,                                         /* tp_compare */
    0,                                         /* tp_repr */
    0,                                         /* tp_as_number */
    0,                                         /* tp_as_sequence */
    0,                                         /* tp_as_mapping */
    0,                                         /* tp_hash */
    0,                                         /* tp_call */
    0,                                         /* tp_str */
    0,                                         /* tp_getattro */
    0,                                         /* tp_setattro */
    0,                                         /* tp_as_buffer */
    Py_TPFLAGS_DEFAULT | Py_TPFLAGS_BASETYPE,  /* tp_flags */
    "HyperLogLog counter",                     /* tp_doc */
    0,                                         /* tp_traverse */
    0,                                         /* tp_clear */
    0,                                         /* tp_richcompare */
    0,                                         /* tp_weaklistoffset */
    0,                                         /* tp_iter */
    0,                                         /* tp_iternext */
    khmer_hllcounter_methods,                  /* tp_methods */
    0,                                         /* tp_members */
    khmer_hllcounter_getseters,                /* tp_getset */
    0,                                         /* tp_base */
    0,                                         /* tp_dict */
    0,                                         /* tp_descr_get */
    0,                                         /* tp_descr_set */
    0,                                         /* tp_dictoffset */
    0,                                         /* tp_init */
    0,                                         /* tp_alloc */
    khmer_hllcounter_new,                      /* tp_new */
};

#define is_hllcounter_obj(v)  (Py_TYPE(v) == &khmer_KHLLCounter_Type)

static PyObject * hllcounter_merge(khmer_KHLLCounter_Object * me,
                                   PyObject * args)
{
    khmer_KHLLCounter_Object * other;

    if (!PyArg_ParseTuple(args, "O!", &khmer_KHLLCounter_Type, &other)) {
        return NULL;
    }

    try {
        me->hllcounter->merge(*(other->hllcounter));
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_ValueError, e.what());
        return NULL;
    }

    Py_RETURN_NONE;
}

//////////////////////////////
// standalone functions

static PyObject * forward_hash(PyObject * self, PyObject * args)
{
    const char * kmer;
    WordLength ksize;

    if (!PyArg_ParseTuple(args, "sb", &kmer, &ksize)) {
        return NULL;
    }

    if (ksize > KSIZE_MAX) {
        PyErr_Format(PyExc_ValueError, "k-mer size must be <= %u", KSIZE_MAX);
        return NULL;
    }

    try {
        PyObject * hash;
        hash = PyLong_FromUnsignedLongLong(_hash(kmer, ksize));
        return hash;
    } catch (khmer_exception &e) {
        PyErr_SetString(PyExc_RuntimeError, e.what());
        return NULL;
    }

}

static PyObject * forward_hash_no_rc(PyObject * self, PyObject * args)
{
    const char * kmer;
    WordLength ksize;

    if (!PyArg_ParseTuple(args, "sb", &kmer, &ksize)) {
        return NULL;
    }

    if (ksize > KSIZE_MAX) {
        PyErr_Format(PyExc_ValueError, "k-mer size must be <= %u", KSIZE_MAX);
        return NULL;
    }

    if (strlen(kmer) != ksize) {
        PyErr_SetString(PyExc_ValueError,
                        "k-mer length must equal the k-size");
        return NULL;
    }

    return PyLong_FromUnsignedLongLong(_hash_forward(kmer, ksize));
}

static PyObject * reverse_hash(PyObject * self, PyObject * args)
{
    HashIntoType val;
    WordLength ksize;

    if (!PyArg_ParseTuple(args, "Kb", &val, &ksize)) {
        return NULL;
    }

    if (ksize > KSIZE_MAX) {
        PyErr_Format(PyExc_ValueError, "k-mer size must be <= %u", KSIZE_MAX);
        return NULL;
    }

    return PyUnicode_FromString(_revhash(val, ksize).c_str());
}

static PyObject * murmur3_forward_hash(PyObject * self, PyObject * args)
{
    const char * kmer;

    if (!PyArg_ParseTuple(args, "s", &kmer)) {
        return NULL;
    }

    return PyLong_FromUnsignedLongLong(_hash_murmur(kmer));
}

static PyObject * murmur3_forward_hash_no_rc(PyObject * self, PyObject * args)
{
    const char * kmer;

    if (!PyArg_ParseTuple(args, "s", &kmer)) {
        return NULL;
    }

    return PyLong_FromUnsignedLongLong(_hash_murmur_forward(kmer));
}

//
// technique for resolving literal below found here:
// https://gcc.gnu.org/onlinedocs/gcc-4.9.1/cpp/Stringification.html
//

static
PyObject *
get_version_cpp( PyObject * self, PyObject * args )
{
#define xstr(s) str(s)
#define str(s) #s
    std::string dVersion = xstr(VERSION);
    return PyUnicode_FromString(dVersion.c_str());
}


//
// Module machinery.
//

static PyMethodDef KhmerMethods[] = {
    {
        "forward_hash",     forward_hash,
        METH_VARARGS,       "",
    },
    {
        "forward_hash_no_rc",   forward_hash_no_rc,
        METH_VARARGS,       "",
    },
    {
        "reverse_hash",     reverse_hash,
        METH_VARARGS,       "",
    },
    {
        "hash_murmur3",
        murmur3_forward_hash,
        METH_VARARGS,
        "Calculate the hash value of a k-mer using MurmurHash3 "
        "(with reverse complement)",
    },
    {
        "hash_no_rc_murmur3",
        murmur3_forward_hash_no_rc,
        METH_VARARGS,
        "Calculate the hash value of a k-mer using MurmurHash3 "
        "(no reverse complement)",
    },
    {
        "get_version_cpp", get_version_cpp,
        METH_VARARGS, "return the VERSION c++ compiler option"
    },
    { NULL, NULL, 0, NULL } // sentinel
};

MOD_INIT(_khmer)
{
    using namespace python;

    if (PyType_Ready(&khmer_KHashtable_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    khmer_KCountgraph_Type.tp_base = &khmer_KHashtable_Type;
    if (PyType_Ready(&khmer_KCountgraph_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    if (PyType_Ready(&khmer_PrePartitionInfo_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    khmer_KSubsetPartition_Type.tp_methods = khmer_subset_methods;
    if (PyType_Ready(&khmer_KSubsetPartition_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    khmer_KNodegraph_Type.tp_base = &khmer_KHashtable_Type;
    khmer_KNodegraph_Type.tp_methods = khmer_hashbits_methods;
    if (PyType_Ready(&khmer_KNodegraph_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    khmer_KGraphLabels_Type.tp_base = &khmer_KNodegraph_Type;
    khmer_KGraphLabels_Type.tp_methods = khmer_graphlabels_methods;
    khmer_KGraphLabels_Type.tp_new = khmer_graphlabels_new;
    if (PyType_Ready(&khmer_KGraphLabels_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    if (PyType_Ready(&khmer_KHLLCounter_Type) < 0) {
        return MOD_ERROR_VAL;
    }
    if (PyType_Ready(&khmer_ReadAlignerType) < 0) {
        return MOD_ERROR_VAL;
    }

    _init_ReadParser_Type_constants();
    if (PyType_Ready( &khmer_ReadParser_Type ) < 0) {
        return MOD_ERROR_VAL;
    }

    if (PyType_Ready(&khmer_Read_Type ) < 0) {
        return MOD_ERROR_VAL;
    }

    if (PyType_Ready(&khmer_ReadPairIterator_Type ) < 0) {
        return MOD_ERROR_VAL;
    }

    PyObject * m;

    MOD_DEF(m, "_khmer", "interface for the khmer module low-level extensions",
            KhmerMethods);

    if (m == NULL) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_ReadParser_Type);
    if (PyModule_AddObject( m, "ReadParser",
                            (PyObject *)&khmer_ReadParser_Type ) < 0) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_KCountgraph_Type);
    if (PyModule_AddObject( m, "Countgraph",
                            (PyObject *)&khmer_KCountgraph_Type ) < 0) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_KNodegraph_Type);
    if (PyModule_AddObject(m, "Nodegraph",
                           (PyObject *)&khmer_KNodegraph_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_KGraphLabels_Type);
    if (PyModule_AddObject(m, "GraphLabels",
                           (PyObject *)&khmer_KGraphLabels_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_KHLLCounter_Type);
    if (PyModule_AddObject(m, "HLLCounter",
                           (PyObject *)&khmer_KHLLCounter_Type) < 0) {
        return MOD_ERROR_VAL;
    }

    Py_INCREF(&khmer_ReadAlignerType);
    if (PyModule_AddObject(m, "ReadAligner",
                           (PyObject *)&khmer_ReadAlignerType) < 0) {
        return MOD_ERROR_VAL;
    }

    return MOD_SUCCESS_VAL(m);
}

// vim: set ft=cpp sts=4 sw=4 tw=79: