updates towards P3 compatibility

setup.cfg

@@ -0,0 +1,2 @@
+[metadata]
+description-file = README.rst

setup.py

@@ -53,19 +53,13 @@
 
 major, minor1, minor2, s, tmp = sys.version_info
 
-# Need to check Python 3 compatibility
-if major == 3:
-    raise SystemExit("""UMI-tools is not fully python3 compatible""")
+# Need to comfirm Python 3 compatibility
+#if major == 3:
+#    raise SystemExit("""UMI-tools is not fully python3 compatible""")
 
 if (major == 2 and minor1 < 7) or major < 2:
     raise SystemExit("""UMI-tools requires Python 2.7 or later.""")
 
-# use requires.txt to identify requirements
-#requires = []
-#for requirement in (
-#        l.strip() for l in open('requires.txt') if not l.startswith("#")):
-#    requires.append(requirement)
-
 umi_tools_packages = ["umi_tools"]
 umi_tools_package_dirs = {'umi_tools': 'umi_tools'}
 
@@ -110,7 +104,7 @@
             "cython>=0.19",
             "numpy>=1.7",
             "pandas>=0.12.0",
-            "pysam>=0.8.4",
+            "pysam==0.8.4",
             "future"]
 setup(
     # package information
@@ -119,19 +113,20 @@
     description='umi-tools: Tools for UMI analyses',
     author='Ian Sudbury',
     author_email='i.sudbery@sheffield.ac.uk',
-    license="BSD",
+    license="MIT",
     platforms=["any"],
     keywords="computational genomics",
     long_description='umi-tools: Tools for UMI analyses',
     classifiers=filter(None, classifiers.split("\n")),
     url="https://github.com/CGATOxford/UMI-tools",
+    download_url="https://github.com/CGATOxford/UMI-tools/tarball/0.0.3",
     # package contents
     packages=umi_tools_packages,
     package_dir=umi_tools_package_dirs,
     include_package_data=True,
     # dependencies
     install_requires=requires,
-    #cmdclass={'build_ext': build_ext},
+    cmdclass={'build_ext': build_ext},
     entry_points={
         'console_scripts': ['umi_tools = umi_tools.umi_tools:main']
     },

umi_tools/Utilities.py

@@ -253,6 +253,7 @@ class method (:func:`cachedmethod`) calls.
 import random
 import uuid
 
+from builtins import bytes, chr
 
 class DefaultOptions:
     stdlog = sys.stdout
@@ -526,12 +527,12 @@ def getParams(options=None):
     if options:
         members = options.__dict__
         for k, v in sorted(members.items()):
-            result.append("# %-40s: %s" % (k, str(v).encode("string_escape")))
+            result.append("# %-40s: %s" % (k, str(v)))
     else:
         vars = inspect.currentframe().f_back.f_locals
         for var in filter(lambda x: re.match("param_", x), vars.keys()):
             result.append("# %-40s: %s" %
-                          (var, str(vars[var]).encode("string_escape")))
+                          (var, str(vars[var])))
 
     if result:
         return "\n".join(result)

umi_tools/dedup.py

@@ -169,6 +169,7 @@
 
 # required to make iteritems python2 and python3 compatible
 from future.utils import iteritems
+from builtins import dict
 
 import pysam
 
@@ -216,7 +217,8 @@ def get_average_umi_distance(umis):
     if len(umis) == 1:
         return -1
 
-    dists = [edit_distance(x, y) for x, y in itertools.combinations(umis, 2)]
+    dists = [edit_distance(x.encode('utf-8'), y.encode('utf-8')) for
+             x, y in itertools.combinations(umis, 2)]
     return float(sum(dists))/(len(dists))
 
 
@@ -299,15 +301,17 @@ def _get_adj_list_adjacency(self, umis, counts, threshold):
         ''' identify all umis within hamming distance threshold'''
 
         return {umi: [umi2 for umi2 in umis if
-                      edit_distance(umi, umi2) <= threshold]
+                      edit_distance(umi.encode('utf-8'),
+                                    umi2.encode('utf-8')) <= threshold]
                 for umi in umis}
 
     def _get_adj_list_directional_adjacency(self, umis, counts, threshold):
         ''' identify all umis within the hamming distance threshold
         and where the counts of the first umi is > (2 * second umi counts)-1'''
 
         return {umi: [umi2 for umi2 in umis if
-                      edit_distance(umi, umi2) == 1 and
+                      edit_distance(umi.encode('utf-8'),
+                                    umi2.encode('utf-8')) == 1 and
                       counts[umi] >= (counts[umi2]*2)-1] for umi in umis}
 
     def _get_adj_list_null(self, umis, counts, threshold):
@@ -530,7 +534,7 @@ def get_bundles(insam, ignore_umi=False, subset=None, quality_threshold=0,
                 out_keys = reads_dict.keys()
 
                 for p in out_keys:
-                    for bundle in reads_dict[p].itervalues():
+                    for bundle in reads_dict[p].values():
                         yield bundle
                     del reads_dict[p]
                     del read_counts[p]
@@ -572,7 +576,7 @@ def get_bundles(insam, ignore_umi=False, subset=None, quality_threshold=0,
                 out_keys = [x for x in reads_dict.keys() if x <= start-1000]
 
                 for p in out_keys:
-                    for bundle in reads_dict[p].itervalues():
+                    for bundle in reads_dict[p].values():
                         yield bundle
                     del reads_dict[p]
                     del read_counts[p]
@@ -632,7 +636,7 @@ def get_bundles(insam, ignore_umi=False, subset=None, quality_threshold=0,
 
     # yield remaining bundles
     for p in reads_dict:
-        for bundle in reads_dict[p].itervalues():
+        for bundle in reads_dict[p].values():
             yield bundle
 
 

umi_tools/version.py

@@ -1,2 +1,2 @@
 
-__version__ = "0.0.3"
+__version__ = "0.0.4"