srx-languagetool 0.5.0 → 0.6.0
Sign up to get free protection for your applications and to get access to all the features.
- checksums.yaml +4 -4
- data/.github/workflows/main.yml +1 -1
- data/CHANGELOG.md +4 -0
- data/Gemfile.lock +18 -18
- data/lib/srx/languagetool/version.rb +1 -1
- data/lib/srx/segment.srx +22 -10
- metadata +2 -2
checksums.yaml
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
---
|
2
2
|
SHA256:
|
3
|
-
metadata.gz:
|
4
|
-
data.tar.gz:
|
3
|
+
metadata.gz: 8848393929f09be9f3141298d9fb3f1a33953e5e2253397105a11004c83c53e1
|
4
|
+
data.tar.gz: a21b9fed14c49f2f9da50dbadbe89a5cd00224b827539d9cb6674d77482c5f87
|
5
5
|
SHA512:
|
6
|
-
metadata.gz:
|
7
|
-
data.tar.gz:
|
6
|
+
metadata.gz: 8562133261975230c8fff7f4ef08a7ef31b505cf819080a296b83934176880200677400f6bd1c93ac75e033f4e64af4a0d803810adfd68d06e2d990560b5f520
|
7
|
+
data.tar.gz: b5c97d8b925d425f6832fa5de77f6c98c6cbb483545219da5ee9f9905095bea9e53eda225140e10f409af87a1bd9e2c8bfccaaa08b80c88e048818e19a3ffe29
|
data/.github/workflows/main.yml
CHANGED
data/CHANGELOG.md
CHANGED
data/Gemfile.lock
CHANGED
@@ -1,7 +1,7 @@
|
|
1
1
|
PATH
|
2
2
|
remote: .
|
3
3
|
specs:
|
4
|
-
srx-languagetool (0.
|
4
|
+
srx-languagetool (0.6.0)
|
5
5
|
srx (< 1.0)
|
6
6
|
|
7
7
|
GEM
|
@@ -14,40 +14,40 @@ GEM
|
|
14
14
|
diff-lcs (1.5.0)
|
15
15
|
e2mmap (0.1.0)
|
16
16
|
jaro_winkler (1.5.4)
|
17
|
-
kramdown (2.3.
|
17
|
+
kramdown (2.3.2)
|
18
18
|
rexml
|
19
19
|
kramdown-parser-gfm (1.1.0)
|
20
20
|
kramdown (~> 2.0)
|
21
21
|
minitest (5.15.0)
|
22
|
-
nokogiri (1.
|
22
|
+
nokogiri (1.13.3-x86_64-darwin)
|
23
23
|
racc (~> 1.4)
|
24
|
-
parallel (1.
|
25
|
-
parser (3.
|
24
|
+
parallel (1.22.1)
|
25
|
+
parser (3.1.1.0)
|
26
26
|
ast (~> 2.4.1)
|
27
27
|
racc (1.6.0)
|
28
|
-
rainbow (3.
|
28
|
+
rainbow (3.1.1)
|
29
29
|
rake (13.0.6)
|
30
|
-
regexp_parser (2.2.
|
30
|
+
regexp_parser (2.2.1)
|
31
31
|
reverse_markdown (2.1.1)
|
32
32
|
nokogiri
|
33
33
|
rexml (3.2.5)
|
34
|
-
rspec-expectations (3.
|
34
|
+
rspec-expectations (3.11.0)
|
35
35
|
diff-lcs (>= 1.2.0, < 2.0)
|
36
|
-
rspec-support (~> 3.
|
37
|
-
rspec-support (3.
|
38
|
-
rubocop (1.
|
36
|
+
rspec-support (~> 3.11.0)
|
37
|
+
rspec-support (3.11.0)
|
38
|
+
rubocop (1.26.1)
|
39
39
|
parallel (~> 1.10)
|
40
|
-
parser (>= 3.
|
40
|
+
parser (>= 3.1.0.0)
|
41
41
|
rainbow (>= 2.2.2, < 4.0)
|
42
42
|
regexp_parser (>= 1.8, < 3.0)
|
43
43
|
rexml
|
44
|
-
rubocop-ast (>= 1.
|
44
|
+
rubocop-ast (>= 1.16.0, < 2.0)
|
45
45
|
ruby-progressbar (~> 1.7)
|
46
46
|
unicode-display_width (>= 1.4.0, < 3.0)
|
47
|
-
rubocop-ast (1.
|
48
|
-
parser (>= 3.
|
47
|
+
rubocop-ast (1.16.0)
|
48
|
+
parser (>= 3.1.1.0)
|
49
49
|
ruby-progressbar (1.11.0)
|
50
|
-
solargraph (0.44.
|
50
|
+
solargraph (0.44.3)
|
51
51
|
backport (~> 1.2)
|
52
52
|
benchmark
|
53
53
|
bundler (>= 1.17.2)
|
@@ -64,7 +64,7 @@ GEM
|
|
64
64
|
yard (~> 0.9, >= 0.9.24)
|
65
65
|
srx (0.6.0)
|
66
66
|
nokogiri (~> 1.11)
|
67
|
-
thor (1.1
|
67
|
+
thor (1.2.1)
|
68
68
|
tilt (2.0.10)
|
69
69
|
unicode-display_width (2.1.0)
|
70
70
|
webrick (1.7.0)
|
@@ -85,4 +85,4 @@ DEPENDENCIES
|
|
85
85
|
srx-languagetool!
|
86
86
|
|
87
87
|
BUNDLED WITH
|
88
|
-
2.3.
|
88
|
+
2.3.10
|
data/lib/srx/segment.srx
CHANGED
@@ -1159,7 +1159,7 @@
|
|
1159
1159
|
<afterbreak>D\.?</afterbreak>
|
1160
1160
|
</rule>
|
1161
1161
|
<rule break="no"><!-- min. -->
|
1162
|
-
<beforebreak>\b([Ee]d|pp|[Vv]iz|i\.?[\s\u00A0]*e|[Vvol]|[Rr]col|maj|Lt|[Ff]ig|[Ff]igs|[Vv]iz|[Vv]ols|[Aa]pprox|[Ii]ncl?|[Aa]cc|Pres|[Cc]orp|[Ee]x|[Cc]onn|[Dd]ept|[Mm]in|
|
1162
|
+
<beforebreak>\b([Ee]d|pp|[Vv]iz|i\.?[\s\u00A0]*e|[Vvol]|[Rr]col|maj|Lt|[Ff]ig|[Ff]igs|[Vv]iz|[Vv]ols|[Aa]pprox|[Ii]ncl?|[Aa]cc|Pres|[Cc]orp|[Ee]x|[Cc]onn|[Dd]ept|[Mm]in|[Mm]ax|[Gg]ovt|[Rr]etd|lb|lbf|ft|c\.?[\s\u00A0]*f|vs|dia|lbs|\d+-(:?oz|kc|in|h[rp]|ml)|M?sec)\.[\s\u00A0]</beforebreak>
|
1163
1163
|
<afterbreak>[^\p{Lu}]|I</afterbreak>
|
1164
1164
|
</rule>
|
1165
1165
|
<rule break="no"><!-- hr. -->
|
@@ -1187,7 +1187,11 @@
|
|
1187
1187
|
<afterbreak></afterbreak>
|
1188
1188
|
</rule>
|
1189
1189
|
<rule break="no"><!-- vs. -->
|
1190
|
-
<beforebreak>\
|
1190
|
+
<beforebreak>\b[Vv]s\.[\s\u00A0]</beforebreak>
|
1191
|
+
<afterbreak></afterbreak>
|
1192
|
+
</rule>
|
1193
|
+
<rule break="no"><!-- pp. -->
|
1194
|
+
<beforebreak>\b(pp|PP)\.[\s\u00A0]</beforebreak>
|
1191
1195
|
<afterbreak></afterbreak>
|
1192
1196
|
</rule>
|
1193
1197
|
<rule break="no"><!-- esp. -->
|
@@ -1249,7 +1253,7 @@
|
|
1249
1253
|
</rule>
|
1250
1254
|
<rule break="no"><!-- https://de.wikipedia.org/wiki/VW_ID.3 -->
|
1251
1255
|
<beforebreak>ID.</beforebreak>
|
1252
|
-
<afterbreak>3|Buzz|Crozz</afterbreak>
|
1256
|
+
<afterbreak>3|4|Buzz|Crozz</afterbreak>
|
1253
1257
|
</rule>
|
1254
1258
|
<rule break="no"><!-- Ph.D. (see rule PH_D) -->
|
1255
1259
|
<beforebreak>\bP[Hh]\.?[\s\u00A0]?[Dd]\.[\s\u00A0]</beforebreak>
|
@@ -1550,7 +1554,7 @@
|
|
1550
1554
|
<afterbreak>[A-Za-z0-9\-]+\.(com|net|org|info|de|es|edu|co|eu|nl|io|cn|uk|gov|biz|ca|tk|ru|br|jp|pl)(\.|\b)</afterbreak>
|
1551
1555
|
</rule>
|
1552
1556
|
<rule break="no">
|
1553
|
-
<beforebreak>\b(Drs|Art|Afr|Am|Ar|Br|Cie|Comp|Dhr|([Pp]rof\.)?[Dd]r|Em|Fa|Kon|Bros)\.\s</beforebreak>
|
1557
|
+
<beforebreak>\b(Drs|Art|Afr|Am|Ar|Br|Cie|Comp|Dhr|([Pp]rof\.)?[Dd]r|Em|Fa|Kon|Bros|Stb)\.\s</beforebreak>
|
1554
1558
|
<afterbreak></afterbreak>
|
1555
1559
|
</rule>
|
1556
1560
|
<rule break="no">
|
@@ -1570,6 +1574,10 @@
|
|
1570
1574
|
<afterbreak></afterbreak>
|
1571
1575
|
</rule>
|
1572
1576
|
<rule break="no">
|
1577
|
+
<beforebreak>\b(alc|bro|opm|acc)\.\s</beforebreak>
|
1578
|
+
<afterbreak></afterbreak>
|
1579
|
+
</rule>
|
1580
|
+
<rule break="no">
|
1573
1581
|
<beforebreak>\b(arch|archeol|art|bc|betr|bez|bibl|bijl|bijv)\.\s</beforebreak>
|
1574
1582
|
<afterbreak></afterbreak>
|
1575
1583
|
</rule>
|
@@ -4642,7 +4650,7 @@
|
|
4642
4650
|
<afterbreak>[XIV\d]+\b</afterbreak>
|
4643
4651
|
</rule>
|
4644
4652
|
<rule break="no">
|
4645
|
-
<beforebreak>\b([Ee]ds?|[Cc]oords?|\d+(r|n|t|è|é|ns|es)|masc|fem|sing|pl|adj|adv|g|kg|m|km|cm|ha|u|h|hrs|s|ss|alt|cant|cast|cert|com|dir|gr|nom|parc|pres|set|Sr|Jr|Admón|Adm|Inc|Co|Hnos|Vda|[VU]d[s]?)\.[\p{Pe}\p{Pf}\p{Pd}"']*[\s\u00A0]</beforebreak>
|
4653
|
+
<beforebreak>\b([Ee]ds?|[Cc]oords?|\d+(r|n|t|è|é|a|rs|ns|es)|masc|fem|sing|pl|adj|adv|g|kg|m|km|cm|ha|u|h|hrs|s|ss|alt|cant|cast|cert|com|dir|gr|nom|parc|pres|set|Sr|Jr|Admón|Adm|Inc|Co|Hnos|Vda|[VU]d[s]?)\.[\p{Pe}\p{Pf}\p{Pd}"']*[\s\u00A0]</beforebreak>
|
4646
4654
|
<afterbreak>[\-¡¿«»"'\u2018\u201C\p{Ps}\u2012\u2013\u2014\u2015\u2053]*\p{Ll}</afterbreak>
|
4647
4655
|
</rule>
|
4648
4656
|
<!-- Any word in acronyms like U.S.A.F or F. B. I. or C. or c.s.p. or p. e. -->
|
@@ -4713,6 +4721,10 @@
|
|
4713
4721
|
<beforebreak>Yahoo![\s\u00A0]</beforebreak>
|
4714
4722
|
<afterbreak>\p{Ll}</afterbreak>
|
4715
4723
|
</rule>
|
4724
|
+
<rule break="no">
|
4725
|
+
<beforebreak>40dB.[\s\u00A0]</beforebreak>
|
4726
|
+
<afterbreak>\p{Ll}</afterbreak>
|
4727
|
+
</rule>
|
4716
4728
|
<rule break="yes">
|
4717
4729
|
<beforebreak>\.\[\d+\][\s\u00A0]</beforebreak>
|
4718
4730
|
<afterbreak></afterbreak>
|
@@ -4874,7 +4886,7 @@
|
|
4874
4886
|
</rule>
|
4875
4887
|
<rule break="no"><!-- https://de.wikipedia.org/wiki/VW_ID.3 -->
|
4876
4888
|
<beforebreak>ID.</beforebreak>
|
4877
|
-
<afterbreak>3|Buzz|Crozz</afterbreak>
|
4889
|
+
<afterbreak>3|4|Buzz|Crozz</afterbreak>
|
4878
4890
|
</rule>
|
4879
4891
|
<rule break="no">
|
4880
4892
|
<beforebreak>[1-3]\.[\u00A0\s]</beforebreak>
|
@@ -4978,7 +4990,7 @@
|
|
4978
4990
|
</rule>
|
4979
4991
|
<!-- German abbreviations -->
|
4980
4992
|
<rule break="no">
|
4981
|
-
<beforebreak>\b(ggü|Mag|mtl|versch|d|Übers|usw|Bzw|bzw|Ab[hkst]|abzgl|[Ee]inschl|[Vv]mtl|bezgl|Abzw|[Vv]sl|ahd|Akk|aktual|allg|alltagsspr|altdt|alttest|amerikan|Anh|Ank|Anm|Art|autom|Auftragsnr|Az|Bat|bayr|Bde?|bearb|Bed|Bem|bes|bez|Bez|Bhf|bspw|btto|bw|Dtl|Dez)\.[\u00A0\s]</beforebreak>
|
4993
|
+
<beforebreak>\b(ggü|Mag|mtl|versch|d|Übers|usw|Bzw|bzw|Ab[hkst]|abzgl|[Ee]inschl|[Vv]mtl|Ev|bezgl|Abzw|[Vv]sl|ahd|Akk|aktual|allg|alltagsspr|altdt|alttest|amerikan|Anh|Ank|Anm|Art|autom|Auftragsnr|Az|Bat|bayr|Bde?|bearb|Bed|Bem|bes|bez|Bez|Bhf|bspw|btto|bw|Dtl|Dez)\.[\u00A0\s]</beforebreak>
|
4982
4994
|
<afterbreak></afterbreak>
|
4983
4995
|
</rule>
|
4984
4996
|
<rule break="no">
|
@@ -5399,7 +5411,7 @@
|
|
5399
5411
|
</rule>
|
5400
5412
|
<rule break="no"><!-- https://de.wikipedia.org/wiki/VW_ID.3 -->
|
5401
5413
|
<beforebreak>ID.</beforebreak>
|
5402
|
-
<afterbreak>3|Buzz|Crozz</afterbreak>
|
5414
|
+
<afterbreak>3|4|Buzz|Crozz</afterbreak>
|
5403
5415
|
</rule>
|
5404
5416
|
<rule break="no"><!-- Ph.D. (see rule PH_D) -->
|
5405
5417
|
<beforebreak>\bP[Hh]\.?[\s\u00A0]?[Dd]\.[\s\u00A0]</beforebreak>
|
@@ -5595,9 +5607,9 @@
|
|
5595
5607
|
<beforebreak>\bдол\.[\h\v]*</beforebreak>
|
5596
5608
|
<afterbreak>США</afterbreak>
|
5597
5609
|
</rule>
|
5598
|
-
<!-- п. 10 від 11.10.1933 -->
|
5610
|
+
<!-- п. 10 від 11.10.1933, д. Василь -->
|
5599
5611
|
<rule break="no">
|
5600
|
-
<beforebreak>(?<!т\.[\h\v]?)\b
|
5612
|
+
<beforebreak>(?<!т\.[\h\v]?)\b[пд]\.[\h\v]*</beforebreak>
|
5601
5613
|
<afterbreak></afterbreak>
|
5602
5614
|
</rule>
|
5603
5615
|
<!-- усталені скорочення, що збігаються з нескороченими словами -->
|
metadata
CHANGED
@@ -1,14 +1,14 @@
|
|
1
1
|
--- !ruby/object:Gem::Specification
|
2
2
|
name: srx-languagetool
|
3
3
|
version: !ruby/object:Gem::Version
|
4
|
-
version: 0.
|
4
|
+
version: 0.6.0
|
5
5
|
platform: ruby
|
6
6
|
authors:
|
7
7
|
- Aaron Madlon-Kay
|
8
8
|
autorequire:
|
9
9
|
bindir: exe
|
10
10
|
cert_chain: []
|
11
|
-
date:
|
11
|
+
date: 2022-03-30 00:00:00.000000000 Z
|
12
12
|
dependencies:
|
13
13
|
- !ruby/object:Gem::Dependency
|
14
14
|
name: srx
|