Merge pull request #2843 from ryojiysd/ctcdecoder-wheels-win

Support building ctc_decoder wheel package on Windows system

native_client/ctcdecode/Makefile

@@ -15,14 +15,24 @@ else
 GENERATE_DEBUG_SYMS :=
 endif
 
+ifeq ($(findstring _NT,$(OS)),_NT)
+	ARCHIVE_EXT := lib
+else
+	ARCHIVE_EXT := a
+endif
+
+FIRST_PARTY := first_party.$(ARCHIVE_EXT)
+THIRD_PARTY := third_party.$(ARCHIVE_EXT)
+
+
 all: bindings
 
 clean-keep-third-party:
 	rm -rf dist temp_build ds_ctcdecoder.egg-info
-	rm -f swigwrapper_wrap.cpp swigwrapper.py first_party.a
+	rm -f swigwrapper_wrap.cpp swigwrapper.py $(FIRST_PARTY)
 
 clean: clean-keep-third-party
-	rm -f third_party.a
+	rm -f $(THIRD_PARTY)
 	rm workspace_status.cc
 	rm -fr bazel-out/
 

native_client/ctcdecode/build_archive.py

@@ -9,14 +9,23 @@ import sys
 
 from multiprocessing.dummy import Pool
 
-ARGS = ['-DKENLM_MAX_ORDER=6', '-std=c++11', '-Wno-unused-local-typedefs', '-Wno-sign-compare']
-OPT_ARGS = ['-O3', '-DNDEBUG']
-DBG_ARGS = ['-O0', '-g', '-UNDEBUG', '-DDEBUG']
+if sys.platform.startswith('win'):
+    ARGS = ['/nologo', '/D KENLM_MAX_ORDER=6', '/EHsc', '/source-charset:utf-8']
+    OPT_ARGS = ['/O2', '/MT', '/D NDEBUG']
+    DBG_ARGS = ['/Od', '/MTd', '/Zi', '/U NDEBUG', '/D DEBUG']
+    OPENFST_DIR = 'third_party/openfst-1.6.9-win'
+else:
+    ARGS = ['-fPIC', '-DKENLM_MAX_ORDER=6', '-std=c++11', '-Wno-unused-local-typedefs', '-Wno-sign-compare']
+    OPT_ARGS = ['-O3', '-DNDEBUG']
+    DBG_ARGS = ['-O0', '-g', '-UNDEBUG', '-DDEBUG']
+    OPENFST_DIR = 'third_party/openfst-1.6.7'
+
+
 
 INCLUDES = [
     '..',
     '../kenlm',
-    'third_party/openfst-1.6.7/src/include',
+    OPENFST_DIR + '/src/include',
     'third_party/ThreadPool'
 ]
 
@@ -24,7 +33,7 @@ KENLM_FILES = (glob.glob('../kenlm/util/*.cc')
                 + glob.glob('../kenlm/lm/*.cc')
                 + glob.glob('../kenlm/util/double-conversion/*.cc'))
 
-KENLM_FILES += glob.glob('third_party/openfst-1.6.7/src/lib/*.cc')
+KENLM_FILES += glob.glob(OPENFST_DIR + '/src/lib/*.cc')
 
 KENLM_FILES = [
     fn for fn in KENLM_FILES
@@ -42,6 +51,8 @@ CTC_DECODER_FILES = [
 
 def build_archive(srcs=[], out_name='', build_dir='temp_build/temp_build', debug=False, num_parallel=1):
     compiler = os.environ.get('CXX', 'g++')
+    if sys.platform.startswith('win'):
+        compiler = '"{}"'.format(compiler)
     ar = os.environ.get('AR', 'ar')
     libtool = os.environ.get('LIBTOOL', 'libtool')
     cflags = os.environ.get('CFLAGS', '') + os.environ.get('CXXFLAGS', '')
@@ -59,13 +70,19 @@ def build_archive(srcs=[], out_name='', build_dir='temp_build/temp_build', debug
         if os.path.exists(outfile):
             return
 
-        cmd = '{cc} -fPIC -c {cflags} {args} {includes} {infile} -o {outfile}'.format(
+        if sys.platform.startswith('win'):
+            file = '"{}"'.format(file.replace('\\', '/'))
+            output = '/Fo"{}"'.format(outfile.replace('\\', '/'))
+        else:
+            output = '-o ' + outfile
+
+        cmd = '{cc} -c {cflags} {args} {includes} {infile} {output}'.format(
             cc=compiler,
             cflags=cflags,
             args=' '.join(args),
             includes=' '.join('-I' + i for i in INCLUDES),
             infile=file,
-            outfile=outfile,
+            output=output,
         )
         print(cmd)
         subprocess.check_call(shlex.split(cmd))
@@ -82,6 +99,13 @@ def build_archive(srcs=[], out_name='', build_dir='temp_build/temp_build', debug
         )
         print(cmd)
         subprocess.check_call(shlex.split(cmd))
+    elif sys.platform.startswith('win'):
+        cmd = '"lib.exe" /OUT:"{outfile}" {infiles} /MACHINE:X64 /NOLOGO'.format(
+            outfile=out_name,
+            infiles=' '.join(obj_files))
+        cmd = cmd.replace('\\', '/')
+        print(cmd)
+        subprocess.check_call(shlex.split(cmd))
     else:
         cmd = '{ar} rcs {outfile} {infiles}'.format(
             ar=ar,

native_client/ctcdecode/setup.py

@@ -54,8 +54,15 @@ def maybe_rebuild(srcs, out_name, build_dir):
 project_version = read('../../VERSION').strip()
 
 build_dir = 'temp_build/temp_build'
-third_party_build = 'third_party.a'
-ctc_decoder_build = 'first_party.a'
+
+if sys.platform.startswith('win'):
+    archive_ext = 'lib'
+else:
+    archive_ext = 'a'
+
+third_party_build = 'third_party.{}'.format(archive_ext)
+ctc_decoder_build = 'first_party.{}'.format(archive_ext)
+
 
 maybe_rebuild(KENLM_FILES, third_party_build, build_dir)
 maybe_rebuild(CTC_DECODER_FILES, ctc_decoder_build, build_dir)