initial commit, ok inoguchi@
This commit is contained in:
parent
ede447a05a
commit
653add3360
25
inputmethods/libkkc-data/Makefile
Normal file
25
inputmethods/libkkc-data/Makefile
Normal file
|
@ -0,0 +1,25 @@
|
|||
# $OpenBSD: Makefile,v 1.1 2022/01/02 20:45:24 uaa Exp $
|
||||
|
||||
COMMENT = language model data for libkkc
|
||||
|
||||
VERSION = 0.2.7
|
||||
DISTNAME = libkkc-data-${VERSION}
|
||||
|
||||
CATEGORIES = inputmethods japanese
|
||||
|
||||
# GPLv3
|
||||
PERMIT_PACKAGE = Yes
|
||||
|
||||
MASTER_SITES = https://github.com/ueno/libkkc/releases/download/v0.3.5/
|
||||
EXTRACT_SUFX = .tar.xz
|
||||
|
||||
MODULES = lang/python
|
||||
|
||||
BUILD_DEPENDS = textproc/libmarisa,-python
|
||||
RUN_DEPENDS = inputmethods/libkkc
|
||||
|
||||
CONFIGURE_STYLE = gnu
|
||||
|
||||
NO_TEST = Yes
|
||||
|
||||
.include <bsd.port.mk>
|
2
inputmethods/libkkc-data/distinfo
Normal file
2
inputmethods/libkkc-data/distinfo
Normal file
|
@ -0,0 +1,2 @@
|
|||
SHA256 (libkkc-data-0.2.7.tar.xz) = nmeHVaAwBD2mjjekBJqilsKWhp/x+55scAJrJUFZW5k=
|
||||
SIZE (libkkc-data-0.2.7.tar.xz) = 22262552
|
54
inputmethods/libkkc-data/patches/patch-genfilter_py
Normal file
54
inputmethods/libkkc-data/patches/patch-genfilter_py
Normal file
|
@ -0,0 +1,54 @@
|
|||
--- tools/genfilter.py.orig Mon Jul 29 09:53:28 2013
|
||||
+++ tools/genfilter.py Thu Sep 23 10:02:41 2021
|
||||
@@ -1,7 +1,7 @@
|
||||
#!/usr/bin/python
|
||||
|
||||
-# Copyright (C) 2011-2013 Daiki Ueno <ueno@gnu.org>
|
||||
-# Copyright (C) 2011-2013 Red Hat, Inc.
|
||||
+# Copyright (C) 2011-2014 Daiki Ueno <ueno@gnu.org>
|
||||
+# Copyright (C) 2011-2014 Red Hat, Inc.
|
||||
|
||||
# This program is free software: you can redistribute it and/or modify
|
||||
# it under the terms of the GNU General Public License as published by
|
||||
@@ -84,24 +84,24 @@ class FilterGenerator(object):
|
||||
|
||||
def generate(self):
|
||||
size = os.fstat(self.infile.fileno()).st_size
|
||||
- n = size / self.record_size
|
||||
+ n = size // self.record_size
|
||||
m = int(math.ceil(-n*math.log10(ERROR_RATE) /
|
||||
math.pow(math.log10(2), 2)))
|
||||
- m = (m/8 + 1)*8
|
||||
+ m = (m//8 + 1)*8
|
||||
inmem = mmap.mmap(self.infile.fileno(),
|
||||
size,
|
||||
access=mmap.ACCESS_READ)
|
||||
- outmem = bytearray(m/8)
|
||||
- for i in xrange(0, n):
|
||||
+ outmem = bytearray(m//8)
|
||||
+ for i in range(0, n):
|
||||
offset = i*self.record_size
|
||||
b0, b1 = struct.unpack("=LL", inmem[offset:offset+8])
|
||||
- for k in xrange(0, 4):
|
||||
+ for k in range(0, 4):
|
||||
h = murmur_hash3_32(b0, b1, k)
|
||||
h = int(h * (m / float(0xFFFFFFFF)))
|
||||
- outmem[h/8] |= (1 << (h%8))
|
||||
+ outmem[h//8] |= (1 << (h%8))
|
||||
inmem.close()
|
||||
- # Convert bytearray to str, for Python 2.6 compatibility.
|
||||
- self.outfile.write(str(outmem))
|
||||
+ # Convert bytearray to bytes, for Python 3 compatibility.
|
||||
+ self.outfile.write(bytes(outmem))
|
||||
|
||||
if __name__ == '__main__':
|
||||
import sys
|
||||
@@ -110,7 +110,7 @@ if __name__ == '__main__':
|
||||
parser = argparse.ArgumentParser(description='filter')
|
||||
parser.add_argument('infile', type=argparse.FileType('r'),
|
||||
help='input file')
|
||||
- parser.add_argument('outfile', type=argparse.FileType('w'),
|
||||
+ parser.add_argument('outfile', type=argparse.FileType('wb'),
|
||||
help='output file')
|
||||
parser.add_argument('record_size', type=int,
|
||||
help='record size')
|
90
inputmethods/libkkc-data/patches/patch-sortlm_py
Normal file
90
inputmethods/libkkc-data/patches/patch-sortlm_py
Normal file
|
@ -0,0 +1,90 @@
|
|||
--- tools/sortlm.py.orig Mon Jul 29 09:53:28 2013
|
||||
+++ tools/sortlm.py Thu Sep 23 10:02:37 2021
|
||||
@@ -1,7 +1,7 @@
|
||||
#!/usr/bin/python
|
||||
|
||||
-# Copyright (C) 2011-2013 Daiki Ueno <ueno@gnu.org>
|
||||
-# Copyright (C) 2011-2013 Red Hat, Inc.
|
||||
+# Copyright (C) 2011-2014 Daiki Ueno <ueno@gnu.org>
|
||||
+# Copyright (C) 2011-2014 Red Hat, Inc.
|
||||
|
||||
# This program is free software: you can redistribute it and/or modify
|
||||
# it under the terms of the GNU General Public License as published by
|
||||
@@ -40,10 +40,10 @@ class SortedGenerator(object):
|
||||
self.__min_cost = 0.0
|
||||
|
||||
def read(self):
|
||||
- print "reading N-grams"
|
||||
+ print("reading N-grams")
|
||||
self.__read_tries()
|
||||
self.__read_ngrams()
|
||||
- print "min cost = %lf" % self.__min_cost
|
||||
+ print("min cost = %lf" % self.__min_cost)
|
||||
|
||||
def __read_tries(self):
|
||||
while True:
|
||||
@@ -58,7 +58,7 @@ class SortedGenerator(object):
|
||||
line = self.__infile.readline()
|
||||
if line == "":
|
||||
break
|
||||
- line = line.strip()
|
||||
+ line = line.strip('\n')
|
||||
if line == "":
|
||||
break
|
||||
match = self.__ngram_line_regex.match(line)
|
||||
@@ -89,7 +89,7 @@ class SortedGenerator(object):
|
||||
line = self.__infile.readline()
|
||||
if line == "":
|
||||
break
|
||||
- line = line.strip()
|
||||
+ line = line.strip('\n')
|
||||
if line == "":
|
||||
break
|
||||
match = self.__ngram_line_regex.match(line)
|
||||
@@ -125,14 +125,11 @@ class SortedGenerator(object):
|
||||
def quantize(cost, min_cost):
|
||||
return max(0, min(65535, int(cost * 65535 / min_cost)))
|
||||
|
||||
- def cmp_header(a, b):
|
||||
- return cmp(a[0], b[0])
|
||||
-
|
||||
- print "writing 1-gram file"
|
||||
+ print("writing 1-gram file")
|
||||
unigram_offsets = {}
|
||||
unigram_file = open("%s.1gram" % self.__output_prefix, "wb")
|
||||
offset = 0
|
||||
- for ids, value in sorted(self.__ngram_entries[0].iteritems()):
|
||||
+ for ids, value in sorted(self.__ngram_entries[0].items()):
|
||||
unigram_offsets[ids[0]] = offset
|
||||
s = struct.pack("=HHH",
|
||||
quantize(value[0], self.__min_cost),
|
||||
@@ -143,13 +140,13 @@ class SortedGenerator(object):
|
||||
offset += 1
|
||||
unigram_file.close()
|
||||
|
||||
- print "writing 2-gram file"
|
||||
+ print("writing 2-gram file")
|
||||
bigram_offsets = {}
|
||||
bigram_file = open("%s.2gram" % self.__output_prefix, "wb")
|
||||
keys = self.__ngram_entries[1].keys()
|
||||
items = [(struct.pack("=LL", ids[1], unigram_offsets[ids[0]]), ids) for ids in keys]
|
||||
offset = 0
|
||||
- for header, ids in sorted(items, cmp=cmp_header):
|
||||
+ for header, ids in sorted(items, key=lambda x: x[0]):
|
||||
value = self.__ngram_entries[1][ids]
|
||||
bigram_offsets[ids] = offset
|
||||
s = struct.pack("=HH",
|
||||
@@ -160,11 +157,11 @@ class SortedGenerator(object):
|
||||
bigram_file.close()
|
||||
|
||||
if len(self.__ngram_entries[2]) > 0:
|
||||
- print "writing 3-gram file"
|
||||
+ print("writing 3-gram file")
|
||||
trigram_file = open("%s.3gram" % self.__output_prefix, "wb")
|
||||
keys = self.__ngram_entries[2].keys()
|
||||
items = [(struct.pack("=LL", ids[2], bigram_offsets[(ids[0], ids[1])]), ids) for ids in keys]
|
||||
- for header, ids in sorted(items, cmp=cmp_header):
|
||||
+ for header, ids in sorted(items, key=lambda x: x[0]):
|
||||
value = self.__ngram_entries[2][ids]
|
||||
s = struct.pack("=H",
|
||||
quantize(value[0], self.__min_cost))
|
1
inputmethods/libkkc-data/pkg/DESCR
Normal file
1
inputmethods/libkkc-data/pkg/DESCR
Normal file
|
@ -0,0 +1 @@
|
|||
language model data for libkkc
|
12
inputmethods/libkkc-data/pkg/PLIST
Normal file
12
inputmethods/libkkc-data/pkg/PLIST
Normal file
|
@ -0,0 +1,12 @@
|
|||
@comment $OpenBSD: PLIST,v 1.1 2022/01/02 20:45:24 uaa Exp $
|
||||
lib/libkkc/
|
||||
lib/libkkc/models/
|
||||
lib/libkkc/models/sorted3/
|
||||
lib/libkkc/models/sorted3/data.1gram
|
||||
lib/libkkc/models/sorted3/data.1gram.index
|
||||
lib/libkkc/models/sorted3/data.2gram
|
||||
lib/libkkc/models/sorted3/data.2gram.filter
|
||||
lib/libkkc/models/sorted3/data.3gram
|
||||
lib/libkkc/models/sorted3/data.3gram.filter
|
||||
lib/libkkc/models/sorted3/data.input
|
||||
lib/libkkc/models/sorted3/metadata.json
|
Loading…
Reference in New Issue
Block a user