Add Python2 support

README.rst

@@ -4,14 +4,14 @@ DeepCpG: Deep neural networks for predicting single-cell DNA methylation
 
 |Version| |License| |PyPI| |Docs| |Tweet|
 
-.. |Version| image:: https://img.shields.io/badge/python-3.5%2B-green.svg
+.. |Version| image:: https://img.shields.io/badge/python-2.7%2B%2C3.4%2B-green.svg
   :target: https://www.python.org/
 
 .. |License| image:: https://img.shields.io/github/license/mashape/apistatus.svg
   :target: https://github.com/cangermueller/deepcpg/tree/master/LICENSE
 
-.. |PyPI| image:: https://img.shields.io/pypi/v/nine.svg?style=plastic
-  :target: https://pypi.python.org/pypi/deepcpg/1.0.1
+.. |PyPI| image:: https://img.shields.io/badge/pypi-latest-orange.svg
+  :target: https://pypi.python.org/pypi/deepcpg
 
 .. |Docs| image:: https://img.shields.io/badge/docs-up--to--date-green.svg
   :target: http://deepcpg.readthedocs.io

deepcpg/__init__.py

@@ -1 +1 @@
-__version__ = '1.0.1'
+__version__ = '1.0.2'

deepcpg/callbacks.py

@@ -1,12 +1,14 @@
+from __future__ import division
+from __future__ import print_function
+
 from collections import OrderedDict
 import os
-from pkg_resources import parse_version
 from time import time
 
-from keras import backend as K
 from keras.callbacks import Callback
 
 import numpy as np
+import six
 
 from .utils import format_table, EPS
 
@@ -17,8 +19,6 @@ class PerformanceLogger(Callback):
     Stores and prints performance metrics for each batch, epoch, and output.
     """
 
-
-
     def __init__(self, metrics=['loss', 'acc'], log_freq=0.1,
                  precision=4, callbacks=[], verbose=1, logger=print):
         self.metrics = metrics
@@ -78,7 +78,7 @@ def _init_logs(self, logs, train=True):
         for mean_name in metrics:
             logs_dict[mean_name] = []
         # Followed by all output metrics
-        for mean_name, names in metrics.items():
+        for mean_name, names in six.iteritems(metrics):
             for name in names:
                 logs_dict[name] = []
 
@@ -87,7 +87,7 @@ def _init_logs(self, logs, train=True):
     def _update_means(self, logs, metrics):
         """Computes the mean over all outputs, if it does not exist yet."""
 
-        for mean_name, names in metrics.items():
+        for mean_name, names in six.iteritems(metrics):
             # Skip, if mean already exists, e.g. loss.
             if logs[mean_name][-1] is not None:
                 continue
@@ -138,7 +138,7 @@ def on_epoch_end(self, epoch, logs={}):
             self._val_epoch_metrics, self.val_epoch_logs = tmp
 
         # Add new epoch logs to logs table
-        for metric, metric_logs in self.epoch_logs.items():
+        for metric, metric_logs in six.iteritems(self.epoch_logs):
             if metric in logs:
                 metric_logs.append(logs[metric])
             else:
@@ -147,7 +147,7 @@ def on_epoch_end(self, epoch, logs={}):
         self._update_means(self.epoch_logs, self._epoch_metrics)
 
         # Add new validation epoch logs to logs table
-        for metric, metric_logs in self.val_epoch_logs.items():
+        for metric, metric_logs in six.iteritems(self.val_epoch_logs):
             metric_val = 'val_' + metric
             if metric_val in logs:
                 metric_logs.append(logs[metric_val])
@@ -163,15 +163,15 @@ def on_epoch_end(self, epoch, logs={}):
             table[mean_name] = []
         # Show output logs
         if self.verbose:
-            for mean_name, names in self._epoch_metrics.items():
+            for mean_name, names in six.iteritems(self._epoch_metrics):
                 for name in names:
                     table[name] = []
-        for name, logs in self.epoch_logs.items():
+        for name, logs in six.iteritems(self.epoch_logs):
             if name in table:
                 table[name].append(logs[-1])
         if self.val_epoch_logs:
             table['split'].append('val')
-            for name, logs in self.val_epoch_logs.items():
+            for name, logs in six.iteritems(self.val_epoch_logs):
                 if name in table:
                     table[name].append(logs[-1])
         self._log('')
@@ -196,12 +196,12 @@ def on_batch_end(self, batch, logs={}):
             self._totals = OrderedDict()
             # Number of samples up to the current batch
             self._nb_totals = OrderedDict()
-            for name in self._batch_logs.keys():
+            for name in self._batch_logs:
                 if name in logs:
                     self._totals[name] = 0
                     self._nb_totals[name] = 0
 
-        for name, value in logs.items():
+        for name, value in six.iteritems(logs):
             # Skip value if nan, which can occur if the batch size is small.
             if np.isnan(value):
                 continue
@@ -241,11 +241,11 @@ def on_batch_end(self, batch, logs={}):
             for mean_name in self._batch_metrics:
                 table[mean_name] = []
             if self.verbose:
-                for mean_name, names in self._batch_metrics.items():
+                for mean_name, names in six.iteritems(self._batch_metrics):
                     for name in names:
                         table[name] = []
                         precision.append(self.precision)
-            for name, logs in self._batch_logs.items():
+            for name, logs in six.iteritems(self._batch_logs):
                 if name in table:
                     table[name].append(logs[-1])
                     precision.append(self.precision)

deepcpg/data/annotations.py

@@ -1,6 +1,9 @@
+from __future__ import division
+from __future__ import print_function
+
 import pandas as pd
 import numpy as np
-import sys
+from six.moves import range
 
 
 def read_bed(filename, sort=False, usecols=[0, 1, 2], *args, **kwargs):
@@ -130,7 +133,6 @@ def group_overlapping(s, e):
     if n == 0:
         return group
     idx = 0
-    l = s[0]
     r = e[0]
     for i in range(1, n):
         if s[i] > r:

deepcpg/data/dna.py

@@ -1,6 +1,10 @@
+from __future__ import division
+from __future__ import print_function
+
 from collections import OrderedDict
 
 import numpy as np
+from six.moves import range
 
 CHAR_TO_INT = OrderedDict([('A', 0), ('T', 1), ('G', 2), ('C', 3), ('N', 4)])
 INT_TO_CHAR = {v: k for k, v in CHAR_TO_INT.items()}

deepcpg/data/fasta.py

@@ -1,7 +1,12 @@
+from __future__ import division
+from __future__ import print_function
+
 import os
 from glob import glob
 import gzip as gz
 
+from six.moves import range
+
 from ..utils import to_list
 
 

deepcpg/data/feature_extractor.py

@@ -1,4 +1,8 @@
+from __future__ import division
+from __future__ import print_function
+
 import numpy as np
+from six.moves import range
 
 
 class KnnCpgFeatureExtractor(object):

deepcpg/data/hdf.py

@@ -1,7 +1,12 @@
+from __future__ import division
+from __future__ import print_function
+
 import re
 
 import h5py as h5
 import numpy as np
+import six
+from six.moves import range
 
 from ..utils import filter_regex, to_list
 
@@ -38,7 +43,7 @@ def ls(filename, group='/', recursive=False, groups=False,
 def write_data(data, filename):
     is_root = isinstance(filename, str)
     group = h5.File(filename, 'w') if is_root else filename
-    for key, value in data.items():
+    for key, value in six.iteritems(data):
         if isinstance(value, dict):
             key_group = group.create_group(key)
             write_data(value, key_group)
@@ -50,7 +55,7 @@ def write_data(data, filename):
 
 def hnames_to_names(hnames):
     names = []
-    for key, value in hnames.items():
+    for key, value in six.iteritems(hnames):
         if isinstance(value, dict):
             for name in hnames_to_names(value):
                 names.append('%s/%s' % (key, name))
@@ -113,7 +118,7 @@ def reader(data_files, names, batch_size=128, nb_sample=None, shuffle=False,
             # the entire file into memory
             idx = np.arange(nb_sample_file)
             np.random.shuffle(idx)
-            for name, value in data_file.items():
+            for name, value in six.iteritems(data_file):
                 data_file[name] = value[:len(idx)][idx]
 
         nb_batch = int(np.ceil(nb_sample_file / batch_size))
@@ -162,7 +167,7 @@ def read_from(reader, nb_sample=None):
         if not isinstance(data_batch, dict):
             data_batch = _to_dict(data_batch)
             is_dict = False
-        for key, value in data_batch.items():
+        for key, value in six.iteritems(data_batch):
             values = data.setdefault(key, [])
             values.append(value)
         nb_seen += len(list(data_batch.values())[0])
@@ -171,7 +176,7 @@ def read_from(reader, nb_sample=None):
 
     data = stack_dict(data)
     if nb_sample:
-        for key, value in data.items():
+        for key, value in six.iteritems(data):
             data[key] = value[:nb_sample]
 
     if not is_dict:

deepcpg/data/stats.py

@@ -5,6 +5,9 @@
     * [sites, cells, context] for window-based statistics
 """
 
+from __future__ import division
+from __future__ import print_function
+
 import numpy as np
 
 from ..utils import EPS, get_from_module

deepcpg/data/utils.py

@@ -1,10 +1,15 @@
+from __future__ import division
+from __future__ import print_function
+
 import gzip
 import threading
 import re
 
 import h5py as h5
 import numpy as np
 import pandas as pd
+import six
+from six.moves import range
 
 from . import hdf
 
@@ -27,6 +32,9 @@ def __next__(self):
         with self.lock:
             return next(self.it)
 
+    def next(self):
+        return self.__next__()
+
 
 def threadsafe_generator(f):
     """A decorator that takes a generator function and makes it thread-safe.
@@ -37,7 +45,7 @@ def g(*a, **kw):
 
 
 def add_to_dict(src, dst):
-    for key, value in src.items():
+    for key, value in six.iteritems(src):
         if isinstance(value, dict):
             if key not in dst:
                 dst[key] = dict()
@@ -50,7 +58,7 @@ def add_to_dict(src, dst):
 
 def stack_dict(data):
     sdata = dict()
-    for key, value in data.items():
+    for key, value in six.iteritems(data):
         if isinstance(value, dict):
             sdata[key] = stack_dict(value)
         else:

deepcpg/evaluation.py

@@ -1,9 +1,13 @@
+from __future__ import division
+from __future__ import print_function
+
 from collections import OrderedDict
 
 import numpy as np
 import pandas as pd
 import sklearn.metrics as skm
 from scipy.stats import kendalltau
+from six.moves import range
 
 from .data import CPG_NAN, OUTPUT_SEP
 from .utils import get_from_module
@@ -147,7 +151,7 @@ def get_output_metrics(output_name):
 
 def evaluate_outputs(outputs, preds):
     perf = []
-    for output_name in outputs.keys():
+    for output_name in outputs:
         _output_name = output_name.split(OUTPUT_SEP)
         if _output_name[-1] in ['cat_var']:
             tmp = evaluate_cat(outputs[output_name],

deepcpg/metrics.py

@@ -1,3 +1,6 @@
+from __future__ import division
+from __future__ import print_function
+
 from keras import backend as K
 
 from .utils import get_from_module

deepcpg/models/cpg.py

@@ -4,6 +4,9 @@
 cells.
 """
 
+from __future__ import division
+from __future__ import print_function
+
 import inspect
 
 from keras import layers as kl

deepcpg/models/dna.py

@@ -3,6 +3,9 @@
 Provides models trained with DNA sequence windows.
 """
 
+from __future__ import division
+from __future__ import print_function
+
 import inspect
 
 from keras import layers as kl
@@ -197,7 +200,7 @@ class ResNet01(DnaModel):
 
     Parameters: 1.700.000
 
-    He et al., “Identity Mappings in Deep Residual Networks.”
+    He et al., 'Identity Mappings in Deep Residual Networks.'
     """
 
     def _res_unit(self, inputs, nb_filter, size=3, stride=1, stage=1, block=1):
@@ -290,7 +293,7 @@ class ResNet02(ResNet01):
 
     Parameters: 2.000.000
 
-    He et al., “Identity Mappings in Deep Residual Networks.”
+    He et al., 'Identity Mappings in Deep Residual Networks.'
     """
 
     def __call__(self, inputs):
@@ -334,7 +337,7 @@ class ResConv01(ResNet01):
 
     Parameters: 2.800.000
 
-    He et al., “Identity Mappings in Deep Residual Networks.”
+    He et al., 'Identity Mappings in Deep Residual Networks.'
     """
 
     def _res_unit(self, inputs, nb_filter, size=3, stride=1, stage=1, block=1):
@@ -419,8 +422,8 @@ class ResAtrous01(DnaModel):
 
     Parameters: 2.000.000
 
-    He et al., “Identity Mappings in Deep Residual Networks.”
-    Yu and Koltun, “Multi-Scale Context Aggregation by Dilated Convolutions.”
+    He et al., 'Identity Mappings in Deep Residual Networks.'
+    Yu and Koltun, 'Multi-Scale Context Aggregation by Dilated Convolutions.'
     """
 
     def _res_unit(self, inputs, nb_filter, size=3, stride=1, atrous=1,
@@ -515,7 +518,7 @@ def __call__(self, inputs):
 def list_models():
     models = dict()
     for name, value in globals().items():
-        if inspect.isclass(value) and name.lower().find('model') == -1:
+        if inspect.isclass(value) and name.lower().find('model') == 0:
             models[name] = value
     return models
 

deepcpg/models/joint.py

@@ -1,3 +1,6 @@
+from __future__ import division
+from __future__ import print_function
+
 from keras import layers as kl
 from keras import models as km
 from keras import regularizers as kr