interfaces.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-
#
# Copyright (C) 2010 Radim Rehurek <radimrehurek@seznam.cz>
# Licensed under the GNU LGPL v2.1 - http://www.gnu.org/licenses/lgpl.html

"""
This module contains basic interfaces used throughout the whole gensim package.

The interfaces are realized as abstract base classes (ie., some optional functionality
is provided in the interface itself, so that the interfaces can be subclassed).
"""

from __future__ import with_statement

import logging
import itertools

from gensim import utils, matutils
from six.moves import xrange


logger = logging.getLogger('gensim.interfaces')


class CorpusABC(utils.SaveLoad):
    """
    Interface (abstract base class) for corpora. A *corpus* is simply an iterable,
    where each iteration step yields one document:

    >>> for doc in corpus:
    >>>     # do something with the doc...

    A document is a sequence of `(fieldId, fieldValue)` 2-tuples:

    >>> for attr_id, attr_value in doc:
    >>>     # do something with the attribute

    Note that although a default :func:`len` method is provided, it is very inefficient
    (performs a linear scan through the corpus to determine its length). Wherever
    the corpus size is needed and known in advance (or at least doesn't change so
    that it can be cached), the :func:`len` method should be overridden.

    See the :mod:`gensim.corpora.svmlightcorpus` module for an example of a corpus.

    Saving the corpus with the `save` method (inherited from `utils.SaveLoad`) will
    only store the *in-memory* (binary, pickled) object representation=the stream
    state, and **not** the documents themselves. See the `save_corpus` static method
    for serializing the actual stream content.
    """
    def __iter__(self):
        """
        Iterate over the corpus, yielding one document at a time.
        """
        raise NotImplementedError('cannot instantiate abstract base class')


    def save(self, *args, **kwargs):
        import warnings
        warnings.warn("corpus.save() stores only the (tiny) iteration object; "
            "to serialize the actual corpus content, use e.g. MmCorpus.serialize(corpus)")
        super(CorpusABC, self).save(*args, **kwargs)

    def __len__(self):
        """
        Return the number of documents in the corpus.

        This method is just the least common denominator and should really be
        overridden when possible.
        """
        raise NotImplementedError("must override __len__() before calling len(corpus)")
#        logger.warning("performing full corpus scan to determine its length; was this intended?")
#        return sum(1 for doc in self) # sum(empty generator) == 0, so this works even for an empty corpus

    @staticmethod
    def save_corpus(fname, corpus, id2word=None, metadata=False):
        """
        Save an existing `corpus` to disk.

        Some formats also support saving the dictionary (`feature_id->word` mapping),
        which can in this case be provided by the optional `id2word` parameter.

        >>> MmCorpus.save_corpus('file.mm', corpus)

        Some corpora also support an index of where each document begins, so
        that the documents on disk can be accessed in O(1) time (see the
        `corpora.IndexedCorpus` base class). In this case, `save_corpus` is automatically
        called internally by `serialize`, which does `save_corpus` plus saves the index
        at the same time, so you want to store the corpus with::

        >>> MmCorpus.serialize('file.mm', corpus) # stores index as well, allowing random access to individual documents

        Calling `serialize()` is preferred to calling `save_corpus()`.

        """
        raise NotImplementedError('cannot instantiate abstract base class')

        # example code:
        logger.info("converting corpus to ??? format: %s" % fname)
        with utils.smart_open(fname, 'wb') as fout:
            for doc in corpus: # iterate over the document stream
                fmt = str(doc) # format the document appropriately...
                fout.write(utils.to_utf8("%s\n" % fmt)) # serialize the formatted document to disk
#endclass CorpusABC


class TransformedCorpus(CorpusABC):
    def __init__(self, obj, corpus, chunksize=None):
        self.obj, self.corpus, self.chunksize = obj, corpus, chunksize
        self.metadata = False

    def __len__(self):
        return len(self.corpus)

    def __iter__(self):
        if self.chunksize:
            for chunk in utils.grouper(self.corpus, self.chunksize):
                for transformed in self.obj.__getitem__(chunk, chunksize=None):
                    yield transformed
        else:
            for doc in self.corpus:
                yield self.obj[doc]

    def __getitem__(self, docno):
        if hasattr(self.corpus, '__getitem__'):
           return self.obj[self.corpus[docno]]
        else:
            raise RuntimeError('Type {} does not support slicing.'.format(type(self.corpus)))
#endclass TransformedCorpus


class TransformationABC(utils.SaveLoad):
    """
    Interface for transformations. A 'transformation' is any object which accepts
    a sparse document via the dictionary notation `[]` and returns another sparse
    document in its stead::

    >>> transformed_doc = transformation[doc]

    or also::

    >>> transformed_corpus = transformation[corpus]

    See the :mod:`gensim.models.tfidfmodel` module for an example of a transformation.

    """

    def __getitem__(self, vec):
        """
        Transform vector from one vector space into another

        **or**

        Transform a whole corpus into another.
        """
        raise NotImplementedError('cannot instantiate abstract base class')


    def _apply(self, corpus, chunksize=None):
        """
        Apply the transformation to a whole corpus (as opposed to a single document)
        and return the result as another corpus.
        """
        return TransformedCorpus(self, corpus, chunksize)
#endclass TransformationABC


class SimilarityABC(utils.SaveLoad):
    """
    Abstract interface for similarity searches over a corpus.

    In all instances, there is a corpus against which we want to perform the
    similarity search.

    For each similarity search, the input is a document and the output are its
    similarities to individual corpus documents.

    Similarity queries are realized by calling ``self[query_document]``.

    There is also a convenience wrapper, where iterating over `self` yields
    similarities of each document in the corpus against the whole corpus (ie.,
    the query is each corpus document in turn).
    """
    def __init__(self, corpus):
        raise NotImplementedError("cannot instantiate Abstract Base Class")


    def get_similarities(self, doc):
        # (Sparse)MatrixSimilarity override this method so that they both use the
        # same  __getitem__ method, defined below
        raise NotImplementedError("cannot instantiate Abstract Base Class")


    def __getitem__(self, query):
        """Get similarities of document `query` to all documents in the corpus.

        **or**

        If `query` is a corpus (iterable of documents), return a matrix of similarities
        of all query documents vs. all corpus document. Using this type of batch
        query is more efficient than computing the similarities one document after
        another.
        """
        is_corpus, query = utils.is_corpus(query)
        if self.normalize:
            # self.normalize only works if the input is a plain gensim vector/corpus (as
            # advertised in the doc). in fact, input can be a numpy or scipy.sparse matrix
            # as well, but in that case assume tricks are happening and don't normalize
            # anything (self.normalize has no effect).
            if matutils.ismatrix(query):
                import warnings
                # warnings.warn("non-gensim input must already come normalized")
            else:
                if is_corpus:
                    query = [matutils.unitvec(v) for v in query]
                else:
                    query = matutils.unitvec(query)
        result = self.get_similarities(query)

        if self.num_best is None:
            return result

        # if the input query was a corpus (=more documents), compute the top-n
        # most similar for each document in turn
        if matutils.ismatrix(result):
            return [matutils.full2sparse_clipped(v, self.num_best) for v in result]
        else:
            # otherwise, return top-n of the single input document
            return matutils.full2sparse_clipped(result, self.num_best)


    def __iter__(self):
        """
        For each index document, compute cosine similarity against all other
        documents in the index and yield the result.
        """
        # turn off query normalization (vectors in the index are assumed to be already normalized)
        norm = self.normalize
        self.normalize = False

        # Try to compute similarities in bigger chunks of documents (not
        # one query = a single document after another). The point is, a
        # bigger query of N documents is faster than N small queries of one
        # document.
        #
        # After computing similarities of the bigger query in `self[chunk]`,
        # yield the resulting similarities one after another, so that it looks
        # exactly the same as if they had been computed with many small queries.
        try:
            chunking = self.chunksize > 1
        except AttributeError:
            # chunking not supported; fall back to the (slower) mode of 1 query=1 document
            chunking = False
        if chunking:
            # assumes `self.corpus` holds the index as a 2-d numpy array.
            # this is true for MatrixSimilarity and SparseMatrixSimilarity, but
            # may not be true for other (future) classes..?
            for chunk_start in xrange(0, self.index.shape[0], self.chunksize):
                # scipy.sparse doesn't allow slicing beyond real size of the matrix
                # (unlike numpy). so, clip the end of the chunk explicitly to make
                # scipy.sparse happy
                chunk_end = min(self.index.shape[0], chunk_start + self.chunksize)
                chunk = self.index[chunk_start : chunk_end]
                if chunk.shape[0] > 1:
                    for sim in self[chunk]:
                        yield sim
                else:
                    yield self[chunk]
        else:
            for doc in self.index:
                yield self[doc]

        # restore old normalization value
        self.normalize = norm
#endclass SimilarityABC