Merge branch 'master' into fix/consume-filename-or-parser

Makefile

@@ -63,9 +63,9 @@ INCLUDESTRING=$(shell gcc -E -x c++ - -v < /dev/null 2>&1 >/dev/null \
 INCLUDEOPTS=$(shell gcc -E -x c++ - -v < /dev/null 2>&1 >/dev/null \
 	    | grep '^ /' | grep -v cc1plus | awk '{print "-I" $$1 " "}')
 
-PYINCLUDE=$(shell python -c "from __future__ import print_function; \
-	    import sysconfig; flags = ['-I' + sysconfig.get_path('include'), \
-	    '-I' + sysconfig.get_path('platinclude')]; print(' '.join(flags))")
+PYINCLUDE=$(shell python -c "import sysconfig; \
+            flags = ['-I' + sysconfig.get_path('include'), \
+            '-I' + sysconfig.get_path('platinclude')]; print(' '.join(flags))")
 
 CPPCHECK_SOURCES=$(filter-out lib/test%, $(wildcard lib/*.cc khmer/_khmer.cc) )
 CPPCHECK=cppcheck --enable=all \

doc/dev/guidelines-continued-dev.rst

@@ -158,7 +158,6 @@ When wrapping code from liboxli:
 
 For imports,
 
-- `__future__` imports at the top, as usual.
 - `libc` cimports next,
 - then `libcpp` imports and cimports.
 - followed by cimports

examples/python-api/bloom.py

@@ -4,7 +4,6 @@
 # khmer accrues a small false positive rate in order to save substantially on
 # memory requirements.
 
-from __future__ import print_function
 import khmer
 
 ksize = 21

examples/python-api/consume.py

@@ -2,7 +2,6 @@
 
 # A demonstration of khmer's primary sequence loading function.
 
-from __future__ import print_function
 import khmer
 import sys
 

examples/python-api/exact-counting.py

@@ -3,7 +3,6 @@
 # A demonstration of using khmer for exact k-mer counting. The memory required
 # is 4^k, which limits this to small values of k.
 
-from __future__ import print_function
 import khmer
 
 # Note:

include/oxli/hashgraph.hh

@@ -43,7 +43,6 @@ Contact: khmer-project@idyll.org
 #include <fstream>
 #include <iostream>
 #include <list>
-#include <map>
 #include <queue>
 #include <set>
 #include <string>

include/oxli/hashtable.hh

@@ -44,7 +44,6 @@ Contact: khmer-project@idyll.org
 #include <fstream>
 #include <iostream>
 #include <list>
-#include <map>
 #include <queue>
 #include <set>
 #include <string>

include/oxli/oxli.hh

@@ -71,6 +71,7 @@ private:\
 
 #include <set>
 #include <map>
+#include <unordered_map>
 #include <queue>
 #include <list>
 #include <functional>
@@ -129,22 +130,22 @@ typedef void (*CallbackFn)(const char * info, void * callback_data,
 typedef unsigned int PartitionID;
 typedef std::set<HashIntoType> SeenSet;
 typedef std::set<PartitionID> PartitionSet;
-typedef std::map<HashIntoType, PartitionID*> PartitionMap;
-typedef std::map<PartitionID, PartitionID*> PartitionPtrMap;
-typedef std::map<PartitionID, SeenSet*> PartitionsToTagsMap;
+typedef std::unordered_map<HashIntoType, PartitionID*> PartitionMap;
+typedef std::unordered_map<PartitionID, PartitionID*> PartitionPtrMap;
+typedef std::unordered_map<PartitionID, SeenSet*> PartitionsToTagsMap;
 typedef std::set<PartitionID *> PartitionPtrSet;
-typedef std::map<PartitionID, PartitionPtrSet*> ReversePartitionMap;
+typedef std::unordered_map<PartitionID, PartitionPtrSet*> ReversePartitionMap;
 typedef std::queue<HashIntoType> NodeQueue;
-typedef std::map<PartitionID, PartitionID*> PartitionToPartitionPMap;
-typedef std::map<HashIntoType, unsigned int> TagCountMap;
-typedef std::map<PartitionID, unsigned int> PartitionCountMap;
+typedef std::unordered_map<PartitionID, PartitionID*> PartitionToPartitionPMap;
+typedef std::unordered_map<HashIntoType, unsigned int> TagCountMap;
+typedef std::unordered_map<PartitionID, unsigned int> PartitionCountMap;
 typedef std::map<unsigned long long, unsigned long long>
 PartitionCountDistribution;
 
 // types used in @camillescott's sparse labeling extension
 typedef unsigned long long int Label;
-typedef std::multimap<HashIntoType, Label> TagLabelMap;
-typedef std::multimap<Label, HashIntoType> LabelTagMap;
+typedef std::unordered_multimap<HashIntoType, Label> TagLabelMap;
+typedef std::unordered_multimap<Label, HashIntoType> LabelTagMap;
 typedef std::pair<HashIntoType, Label> TagLabelPair;
 typedef std::pair<Label, HashIntoType> LabelTagPair;
 typedef std::set<Label> LabelSet;

include/oxli/storage.hh

@@ -41,12 +41,13 @@ Contact: khmer-project@idyll.org
 #include <cassert>
 #include <array>
 #include <mutex>
+#include <unordered_map>
 using MuxGuard = std::lock_guard<std::mutex>;
 
 #include "gqf.h"
 
 namespace oxli {
-typedef std::map<HashIntoType, BoundedCounterType> KmerCountMap;
+typedef std::unordered_map<HashIntoType, BoundedCounterType> KmerCountMap;
 
 //
 // base Storage class for hashtable-related storage of information in memory.

khmer/__init__.py

@@ -36,7 +36,6 @@
 """This is khmer; please see http://khmer.readthedocs.io/."""
 
 
-from __future__ import print_function
 from collections import namedtuple
 from math import log
 import json

khmer/_oxli/oxli_types.pxd

@@ -1,5 +1,6 @@
 from libcpp cimport bool
 from libcpp.map cimport map
+from libcpp.unordered_map cimport unordered_map
 from libcpp.set cimport set
 from libcpp.string cimport string
 
@@ -13,12 +14,12 @@ cdef extern from "oxli/oxli.hh" namespace "oxli":
 
     ctypedef unsigned int PartitionID
     ctypedef set[PartitionID] PartitionSet
-    ctypedef map[HashIntoType, PartitionID*] PartitionMap
-    ctypedef map[PartitionID, PartitionID*] PartitionPtrMap
-    ctypedef map[PartitionID, HashIntoTypeSet*] PartitionToTagsMap
-    ctypedef map[PartitionID, unsigned int] PartitionCountMap
+    ctypedef unordered_map[HashIntoType, PartitionID*] PartitionMap
+    ctypedef unordered_map[PartitionID, PartitionID*] PartitionPtrMap
+    ctypedef unordered_map[PartitionID, HashIntoTypeSet*] PartitionToTagsMap
+    ctypedef unordered_map[PartitionID, unsigned int] PartitionCountMap
     ctypedef map[unsigned long long, unsigned long long] PartitionCountDistribution
-    ctypedef map[HashIntoType, unsigned int] TagCountMap
+    ctypedef unordered_map[HashIntoType, unsigned int] TagCountMap
     ctypedef unsigned char WordLength
     ctypedef unsigned short int BoundedCounterType
 

khmer/_oxli/parsing.pxd

@@ -1,6 +1,5 @@
 # -*- coding: UTF-8 -*-
 
-from __future__ import unicode_literals
 
 from libc.stdint cimport uintptr_t
 

khmer/_oxli/parsing.pyx

@@ -1,7 +1,5 @@
 # -*- coding: UTF-8 -*-
 
-from __future__ import print_function
-from __future__ import unicode_literals
 
 from cython.operator cimport dereference as deref
 cimport cython

khmer/_oxli/utils.pyx

@@ -1,6 +1,5 @@
 # -*- coding: UTF-8 -*-
 
-from __future__ import unicode_literals
 from cpython.version cimport PY_MAJOR_VERSION
 
 from cython import short, int, long

khmer/kfile.py

@@ -34,7 +34,6 @@
 # Contact: khmer-project@idyll.org
 """File handling/checking utilities for command-line scripts."""
 
-from __future__ import print_function, unicode_literals, division
 
 import os
 import sys

khmer/khmer_args.py

@@ -35,8 +35,6 @@
 # Contact: khmer-project@idyll.org
 """Common argparse constructs."""
 
-from __future__ import unicode_literals
-from __future__ import print_function
 
 import sys
 import argparse

khmer/khmer_logger.py

@@ -33,7 +33,6 @@
 # Contact: khmer-project@idyll.org
 """Lightweight logging framework for khmer."""
 
-from __future__ import print_function, unicode_literals
 import sys
 
 __QUIET__ = False

khmer/thread_utils.py

@@ -35,7 +35,6 @@
 # pylint: disable=missing-docstring,too-few-public-methods
 """Utilities for dealing with multithreaded processing of short reads."""
 
-from __future__ import print_function, unicode_literals
 
 import threading
 import sys

khmer/trimming.py

@@ -32,7 +32,6 @@
 #
 # Contact: khmer-project@idyll.org
 """Common methods for trimming short reads on k-mer abundance."""
-from __future__ import print_function, unicode_literals
 import screed
 
 

khmer/utils.py

@@ -33,7 +33,6 @@
 #
 # Contact: khmer-project@idyll.org
 """Helpful methods for performing common argument-checking tasks in scripts."""
-from __future__ import print_function, unicode_literals
 from khmer._oxli.parsing import (check_is_left, check_is_right, check_is_pair,
                                  UnpairedReadsError, _split_left_right)
 import itertools

oxli/build_graph.py

@@ -43,7 +43,6 @@
 Use '-h' for parameter help.
 """
 
-from __future__ import print_function, absolute_import, unicode_literals
 
 import sys
 

oxli/functions.py

@@ -35,7 +35,6 @@
 
 """A collection of functions for use throughout khmer/oxli."""
 
-from __future__ import print_function
 import threading
 import khmer.utils
 

oxli/partition.py

@@ -7,7 +7,6 @@
 #
 # pylint: disable=missing-docstring,no-member
 """Common functions for partitioning."""
-from __future__ import print_function, absolute_import, unicode_literals
 
 import sys
 import gc

sandbox/assemble-and-track.py

@@ -1,5 +1,4 @@
 #! /usr/bin/env python
-from __future__ import print_function
 import csv
 import screed
 import khmer

sandbox/assemble-on-the-go.py

@@ -1,5 +1,4 @@
 #! /usr/bin/env python
-from __future__ import print_function
 import screed
 import khmer
 import argparse

sandbox/assembly-diff-2.py

@@ -33,7 +33,6 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #
 # Contact: khmer-project@idyll.org
-from __future__ import print_function
 import sys
 import khmer
 import screed

sandbox/assembly-diff.py

@@ -33,8 +33,6 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #
 # Contact: khmer-project@idyll.org
-from __future__ import division
-from __future__ import print_function
 import sys
 import khmer
 import screed

sandbox/assemstats3.py

@@ -41,8 +41,6 @@
 You can obtain screed by running
    pip install screed
 '''
-from __future__ import division
-from __future__ import print_function
 
 import screed
 import sys

sandbox/bloom-count.py

@@ -34,8 +34,6 @@
 #
 # Contact: khmer-project@idyll.org
 # pylint: disable=missing-docstring,no-member
-from __future__ import print_function
-from __future__ import absolute_import
 
 import khmer
 import sys

sandbox/build-sparse-graph.py

@@ -34,7 +34,6 @@
 #
 # Contact: khmer-project@idyll.org
 
-from __future__ import print_function
 import khmer
 import sys
 import screed

sandbox/calc-best-assembly.py

@@ -33,7 +33,6 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #
 # Contact: khmer-project@idyll.org
-from __future__ import print_function
 import screed
 import argparse
 import sys

sandbox/calc-error-profile.py

@@ -41,8 +41,6 @@
 
 Reads FASTQ and FASTA input.
 """
-from __future__ import division
-from __future__ import print_function
 
 import sys
 import argparse

sandbox/calc-median-distribution.py

@@ -33,8 +33,6 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #
 # Contact: khmer-project@idyll.org
-from __future__ import division
-from __future__ import print_function
 import sys
 import khmer
 import argparse

sandbox/collect-reads.py

@@ -43,8 +43,6 @@
 
 Use '-h' for parameter help.
 """
-from __future__ import division
-from __future__ import print_function
 
 import sys
 import textwrap

sandbox/collect-variants.py

@@ -40,7 +40,6 @@
 
 TODO: add to sandbox README
 """
-from __future__ import print_function
 
 import sys
 import screed

sandbox/correct-reads.py

@@ -44,7 +44,6 @@
 
 TODO: add to sandbox/README.
 """
-from __future__ import print_function
 import sys
 import os
 import tempfile

sandbox/count-kmers-single.py

@@ -41,7 +41,6 @@
 
 Use '-h' for parameter help.
 """
-from __future__ import print_function
 
 import sys
 import khmer

sandbox/count-kmers.py

@@ -41,7 +41,6 @@
 
 Use '-h' for parameter help.
 """
-from __future__ import print_function
 
 import sys
 import khmer

sandbox/error-correct-pass2.py

@@ -41,7 +41,6 @@
 
 Use '-h' for parameter help.
 """
-from __future__ import print_function
 import sys
 import os
 import screed

sandbox/estimate_optimal_hash.py

@@ -54,7 +54,6 @@
 Use '-h' for parameter help.
 
 """
-from __future__ import print_function
 import argparse
 import khmer, oxli
 from khmer.khmer_args import info, optimal_size, sanitize_help

sandbox/extract-compact-dbg.py

@@ -1,5 +1,4 @@
 #! /usr/bin/env python
-from __future__ import print_function
 import khmer
 import screed
 import argparse

sandbox/extract-single-partition.py

@@ -33,7 +33,6 @@
 # OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 #
 # Contact: khmer-project@idyll.org
-from __future__ import print_function
 import sys
 from screed.fasta import fasta_iter