sportdb-parser 0.6.11 → 0.6.12

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (10) hide show
  1. checksums.yaml +4 -4
  2. data/CHANGELOG.md +1 -1
  3. data/Manifest.txt +1 -0
  4. data/config/rounds_en.txt +0 -3
  5. data/config/zones_en.txt +18 -0
  6. data/lib/sportdb/parser/lang.rb +91 -34
  7. data/lib/sportdb/parser/lexer.rb +2 -2
  8. data/lib/sportdb/parser/token-status.rb +8 -40
  9. data/lib/sportdb/parser/version.rb +1 -1
  10. metadata +4 -2
checksums.yaml CHANGED
@@ -1,7 +1,7 @@
1
1
  ---
2
2
  SHA256:
3
- metadata.gz: '0902e4f811d2584b7c7360f00557d979579d5df8b48e3370ace0d3e7d211c45d'
4
- data.tar.gz: e03327cfb9f33e39c3cc44063f9915ccc48d1c78b8076ee9fd7ba511d7b7bdc5
3
+ metadata.gz: 45ae1f806a3734a32c34fd96a32784568443875cd82c9ab1e8ab867bd7358977
4
+ data.tar.gz: 3bc7ed4c867722b5d950ff759afdf02191405949fedffa4609d81c74daf882be
5
5
  SHA512:
6
- metadata.gz: 14e1f06f0fd5137208cce8a26f9a534317f0538a59f1701ae0140914afae9bed2963979df71d1032771798983dcf5ee5197c0105c6fb040c0e72052b86d282b5
7
- data.tar.gz: e8b973f129cf11f951ba5629b81c53896bec02b949b4eb8d7c175b5fae54e33e31788cdfc18985f0d2759dfc473758512523adddf122e4c5293ce2c45d28543f
6
+ metadata.gz: a6d43efebff7d7558d084fea36d83bc6ec9fdb4aed83aa23d8226f79b10018ddbaf6b607539ac6d664b1038d7772062fcf4cf6f22791444e1c87353889e77d91
7
+ data.tar.gz: ae2151c60bb90bc4dd61d3fadd3fed6e44b865739ebfb7b28bd8c4a9e3b373cfb1827a2202ebc740ec8a69103d1073c6fe1caa5241a3580495aef1e503d9f55c
data/CHANGELOG.md CHANGED
@@ -1,4 +1,4 @@
1
- ### 0.6.11
1
+ ### 0.6.12
2
2
  ### 0.0.1 / 2024-07-12
3
3
 
4
4
  * Everything is new. First release.
data/Manifest.txt CHANGED
@@ -7,6 +7,7 @@ config/rounds_en.txt
7
7
  config/rounds_es.txt
8
8
  config/rounds_misc.txt
9
9
  config/rounds_pt.txt
10
+ config/zones_en.txt
10
11
  lib/sportdb/parser.rb
11
12
  lib/sportdb/parser/lang.rb
12
13
  lib/sportdb/parser/lexer.rb
data/config/rounds_en.txt CHANGED
@@ -17,9 +17,6 @@ Major Semi-Final
17
17
  Minor Semi-Final
18
18
 
19
19
 
20
- 1/8 Finals
21
- Qualifying Round
22
-
23
20
 
24
21
 
25
22
  ## keep weirdo matchday ??
data/config/zones_en.txt ADDED
@@ -0,0 +1,18 @@
1
+ #####
2
+ # zone names in english
3
+
4
+
5
+ Western Region
6
+ Eastern Region
7
+
8
+ Western Conference
9
+ Eastern Conference
10
+
11
+ Northern Zone
12
+ Western Zone A
13
+ Western Zone B
14
+ Central Zone
15
+ Central & Eastern Zone
16
+ Southern Zone
17
+
18
+
data/lib/sportdb/parser/lang.rb CHANGED
@@ -31,17 +31,20 @@ end
31
31
 
32
32
 
33
33
 
34
- ROUND_RE = %r{^(
35
- ## add special case for group play-off rounds!
34
+ ROUND_RE = %r{^
35
+ (?:
36
+
37
+ ## add special case for group play-off rounds!
36
38
  ## group 2 play-off (e.g. worldcup 1954, 1958)
37
39
  ##
38
40
  ### note - allow Group ("stand-alone") as "generic" round for now
39
41
  ## BUT do NOT allow Group 1, Group 2, Group A, Group B, etc.
40
42
  (?: Group [ ] [A-Z0-9]+ [ ] Play-?offs? |
41
- Group (?: [ ] phase)? |
43
+ Group (?: [ ] phase|stage)? |
42
44
  League (?: [ ] phase)?
43
45
  )
44
46
  |
47
+
45
48
  # round - note - requiers number e.g. round 1,2, etc.
46
49
  # note - use 1-9 regex (cannot start with 0) - why? why not?
47
50
  # make week 01 or round 01 or matchday 01 possible?
@@ -50,6 +53,9 @@ ROUND_RE = %r{^(
50
53
  Week
51
54
  )
52
55
  [ ] [1-9][0-9]*
56
+ (?: ## note - add optional Matchday 1 of 2 or such
57
+ [ ] of [1-9][0-9]*
58
+ )?
53
59
  )
54
60
  |
55
61
  ## starting with qual(ification)
@@ -58,8 +64,8 @@ ROUND_RE = %r{^(
58
64
  ## Playoff Round 1
59
65
  ## Play-in Round 1
60
66
  (?: (?: Qual \. |
61
- Play-?off |
62
- Play-?in
67
+ Play[ -]?off |
68
+ Play[ -]?in
63
69
  )
64
70
  [ ] Round [ ] [1-9][0-9]* )
65
71
  |
@@ -67,29 +73,45 @@ ROUND_RE = %r{^(
67
73
  ## First Round
68
74
  ## Play-off Round
69
75
  ## Final Round (e.g. Worldcup 1950)
70
- (?:
71
- (?: [1-9][0-9]* \. |
72
- 1st | First |
73
- 2nd | Second |
74
- Play-?off |
75
- Final
76
- )
77
- [ ] Round
76
+ (?: (?:
77
+ Play[ -]?off |
78
+ Final |
79
+ Wildcard |
80
+ Qualifying |
81
+ (?:
82
+ (?:
83
+ [1-9][0-9]* \. |
84
+ 1st | First |
85
+ 2nd | Second |
86
+ 3rd | Third |
87
+ 4th | Fourth |
88
+ 5th | Fifth
89
+ )
90
+ (?: ## with optionals
91
+ [ ] Qualifying
92
+ )?
93
+ )
94
+ )
95
+ [ ] Round
78
96
  )
79
97
  |
80
98
  ## starting with preliminary
81
99
  # e.g. Preliminary round
82
100
  (?: Preliminary [ ]
83
101
  (?: Round |
84
- Semi-?finals |
85
- Final
102
+ Semi[ -]?finals |
103
+ Final |
104
+ Qualifier
86
105
  )
87
106
  )
88
107
  |
89
108
  # more (kockout) rounds
90
- # playoffs - playoff, play-off, play-offs
91
- (?: Play-?offs?
92
- (?: [ ]for[ ]quarter-?finals )?
109
+ # playoffs - playoff, play-off, play-offs &
110
+ # playins
111
+ (?:
112
+ Play[ -]?offs? (?: [ ]for[ ]quarter-?finals )?
113
+ |
114
+ Play[ -]?ins?
93
115
  )
94
116
  |
95
117
  # round32
@@ -98,13 +120,12 @@ ROUND_RE = %r{^(
98
120
  |
99
121
  # round16
100
122
  (?: Round[ ]of[ ]16 |
101
- Last[ ]16 |
102
- 8th[ ]finals )
123
+ Last[ ]16 )
103
124
  |
104
125
  # fifthplace
105
126
  (?:
106
127
  (?: (Fifth|5th)[ -]place
107
- (?: [ ] (?: match|play-?off|final ))?
128
+ (?: [ ] (?: match|play[ -]?off|final ))?
108
129
  ) |
109
130
  (?: Match[ ]for[ ](?: fifth|5th )[ -]place )
110
131
  )
@@ -112,38 +133,44 @@ ROUND_RE = %r{^(
112
133
  # thirdplace
113
134
  (?:
114
135
  (?: (Third|3rd)[ -]place
115
- (?: [ ] (?: match|play-?off|final ))?
136
+ (?: [ ] (?: match|play[ -]?off|final ))?
116
137
  ) |
117
138
  (?: Match[ ]for[ ](?: third|3rd )[ -]place )
118
139
  )
119
140
  |
120
141
  # quarterfinals
121
142
  (?:
122
- Quarter-?finals? |
143
+ ## note - allow quarter-finals/quarter finals/quarterfinals
144
+ Quarter[ -]?finals? |
123
145
  Quarters |
124
- Last[ ]8
146
+ Last[ ]8 |
147
+ 8th[ ]finals |
148
+ 1/8[ ]finals ## check 1/8 finals is same as querter-finals?
125
149
  )
126
150
  |
127
151
  # semifinals
128
152
  (?:
129
- Semi-?finals? |
153
+ Semi[ -]?finals? |
130
154
  Semis |
131
- Last[ ]4
155
+ Last[ ]4 |
156
+ 1/4[ ]finals ## check 1/4 finals is same as semi-finals?
132
157
  )
133
158
  |
134
159
  # final
135
160
  Finals?
136
161
  |
137
- # decider e.g. Entscheidungsspiel
138
- Decider
139
162
  |
140
163
  ## add replays
141
164
  ## e.g. Final Replay
142
165
  ## Quarter-finals replays
143
166
  ## First round replays
144
167
  (?:
145
- (?: First [ ] Round |
146
- Quarter-?finals? |
168
+ (?: (?: 1st | First |
169
+ 2nd | Second |
170
+ 3rd | Third |
171
+ 4th | Fourth |
172
+ 5th | Fifth ) [ ] Round |
173
+ Quarter[ -]?finals? |
147
174
  Finals?
148
175
  )
149
176
  [ ] Replays?
@@ -151,7 +178,8 @@ ROUND_RE = %r{^(
151
178
  |
152
179
  ## more
153
180
  (?:
154
- Reclassification
181
+ Decider | # decider e.g. Entscheidungsspiel
182
+ Reclassification
155
183
  )
156
184
  )$}ix
157
185
 
@@ -196,6 +224,19 @@ def self.more_round_names
196
224
  end
197
225
  end
198
226
 
227
+ def self.zone_names
228
+ @zone_name ||= begin
229
+ names = []
230
+ langs = ['en']
231
+ ## sort names by length??
232
+ langs.each do |lang|
233
+ path = "#{SportDb::Module::Parser.root}/config/zones_#{lang}.txt"
234
+ names += read_names( path )
235
+ end
236
+ names
237
+ end
238
+ end
239
+
199
240
 
200
241
  def self.is_round?( text )
201
242
  ### note - use check for case-insensitive
@@ -208,20 +249,36 @@ def self.is_round?( text )
208
249
  ## maybe in the future use our own unaccent and downcase - wyh? why not?
209
250
  ## note - for now ROUND_RE is also case-insensitive!!
210
251
 
211
- ROUND_RE.match?( text ) || more_round_names.any?{ |str| str.casecmp( text )==0 }
252
+ ROUND_RE.match?( text ) ||
253
+ more_round_names.any?{ |str| str.casecmp( text )==0 }
212
254
  end
213
255
 
256
+ def self.is_zone?( text )
257
+ zone_names.any?{ |str| str.casecmp( text )==0 }
258
+ end
259
+
260
+
214
261
  ##
215
262
  ## keep leg separate (from round) - why? why not?
216
263
  ##
217
264
  LEG_RE = %r{^
218
265
  # leg1
219
- (?: 1st|First)[ ]leg
266
+ (?: 1st|First) [ ] leg
220
267
  |
221
268
  # leg2
222
- (?: 2nd|Second)[ ]leg
269
+ (?: 2nd|Second) [ ] leg
270
+ |
271
+ # leg 1 of 2 / leg 2 of 2
272
+ # note - leg limited to ALWAY 1/2 of 2 for now - why? why not?
273
+ # for more use match 1/2/3 etc.
274
+ (?: leg [ ] [12]
275
+ (?: [ ] of [ ] 2)? )
276
+ |
277
+ (?: match [ ] [1-9][0-9]* )
223
278
  $}ix
224
279
 
280
+
281
+
225
282
  ### Pair matches/games if marked with leg1 n leg2
226
283
  def self.is_leg?( text )
227
284
  LEG_RE.match?( text )
data/lib/sportdb/parser/lexer.rb CHANGED
@@ -26,7 +26,7 @@ end
26
26
  def is_group?( text ) Lang.is_group?( text ); end
27
27
  def is_round?( text ) Lang.is_round?( text ); end
28
28
  def is_leg?( text ) Lang.is_leg?( text ); end
29
-
29
+ def is_zone?( text ) Lang.is_zone?( text ); end
30
30
 
31
31
  ## transforms
32
32
  ##
@@ -187,7 +187,7 @@ def tokenize_with_errors
187
187
  text = t[1]
188
188
  t = if is_group?( text )
189
189
  [:GROUP, text]
190
- elsif is_round?( text ) || is_leg?( text )
190
+ elsif is_round?( text ) || is_leg?( text ) || is_zone?( text )
191
191
  [:ROUND, text]
192
192
  else
193
193
  t ## pass through as-is (1:1)
data/lib/sportdb/parser/token-status.rb CHANGED
@@ -90,53 +90,21 @@ NOTE_RE = %r{
90
90
  # add "top-level" NB: version
91
91
  ## with full (end-of) line note - why? why not?
92
92
  |
93
- (?: originally[ ])? scheduled
94
- ## e.g. [originally scheduled to play in Mexico City]
95
- |
96
93
  rescheduled
97
- ## e.g. [Rescheduled due to earthquake occurred in Mexico on September 19]
94
+ ## e.g. [rescheduled due to earthquake occurred in Mexico on September 19]
95
+ |
96
+ declared
97
+ ## e.g. [declared void]
98
98
  |
99
99
  remaining
100
100
  ## e.g. [remaining 79']
101
101
  ## [remaining 84']
102
102
  ## [remaining 59']
103
103
  ## [remaining 5']
104
- |
105
- played
106
- ## e.g. [played in Macaé-RJ]
107
- ## [played in Caxias do Sul-RS]
108
- ## [played in Sete Lagoas-MG]
109
- ## [played in Uberlândia-MG]
110
- ## [played in Brasília-DF]
111
- ## [played in Vöcklabruck]
112
- ## [played in Pasching]
113
- |
114
- declared
115
- ## e.g. [declared void]
116
- |
117
- inter-group
118
- ## e.g. [inter-group A-B]
119
- ## [inter-group C-D]
120
104
  )
121
105
  [ ]
122
106
  [^\]]+? ## slurp all to next ] - (use non-greedy)
123
107
  )
124
- |
125
- (?:
126
- ## starting with in - do NOT allow digits
127
- ## name starting with in possible - why? why not?
128
- in[ ]
129
- [^0-9\]]+?
130
- ## e.g. [In Estadio La Corregidora]
131
- ## [in Unidad Deportiva Centenario]
132
- ## [in Estadio Olímpico Universitario]
133
- ## [in Estadio Victoria]
134
- ## [in UD José Brindis]
135
- ## [in Colomos Alfredo "Pistache" Torres stadium]
136
- ##
137
- ## TODO/FIX
138
- ## remove in ?? - is same as @ Estadio Victoria and such - why? why not=
139
- )
140
108
  )
141
109
  \]
142
110
  }ix
@@ -148,7 +116,7 @@ SCORE_NOTE_RE = %r{
148
116
  (?<score_note>
149
117
  (?: # plain aet e.g. [aet]
150
118
  aet | a\.e\.t\. |
151
- after [ ] extra [ ] time
119
+ after [ ] extra [ -] time
152
120
  )
153
121
  |
154
122
  (?: # plain penalties e.g. [3-2 pen]
@@ -176,7 +144,7 @@ SCORE_NOTE_RE = %r{
176
144
 
177
145
  (?:
178
146
  (?: # opt 1 - no team listed/named - requires score
179
- wins? [ ] ## note - allow win or wins
147
+ (?: won|wins? ) [ ] ## note - allow won,win or wins
180
148
  (?: ## score
181
149
  \d{1,2}-\d{1,2}
182
150
  [ ]
@@ -190,7 +158,7 @@ SCORE_NOTE_RE = %r{
190
158
  [1-9\p{L}][0-9\p{L} .-]+?
191
159
  [ ]
192
160
  )
193
- wins [ ]
161
+ (?: won|wins? ) [ ] ## won/win/wins
194
162
  (?: ## score optional
195
163
  \d{1,2}-\d{1,2}
196
164
  [ ]
@@ -214,7 +182,7 @@ SCORE_NOTE_RE = %r{
214
182
  [1-9\p{L}][0-9\p{L} .-]+?
215
183
  [ ]
216
184
  )
217
- wins? [ ]
185
+ (?: won|wins? ) [ ] # won/win/wins
218
186
  on [ ] away [ ] goals
219
187
  )
220
188
  ) # score_note ref
data/lib/sportdb/parser/version.rb CHANGED
@@ -4,7 +4,7 @@ module SportDb
4
4
  module Parser
5
5
  MAJOR = 0 ## todo: namespace inside version or something - why? why not??
6
6
  MINOR = 6
7
- PATCH = 11
7
+ PATCH = 12
8
8
  VERSION = [MAJOR,MINOR,PATCH].join('.')
9
9
 
10
10
  def self.version
metadata CHANGED
@@ -1,14 +1,14 @@
1
1
  --- !ruby/object:Gem::Specification
2
2
  name: sportdb-parser
3
3
  version: !ruby/object:Gem::Version
4
- version: 0.6.11
4
+ version: 0.6.12
5
5
  platform: ruby
6
6
  authors:
7
7
  - Gerald Bauer
8
8
  autorequire:
9
9
  bindir: bin
10
10
  cert_chain: []
11
- date: 2025-03-03 00:00:00.000000000 Z
11
+ date: 2025-03-05 00:00:00.000000000 Z
12
12
  dependencies:
13
13
  - !ruby/object:Gem::Dependency
14
14
  name: cocos
@@ -85,6 +85,7 @@ extra_rdoc_files:
85
85
  - config/rounds_es.txt
86
86
  - config/rounds_misc.txt
87
87
  - config/rounds_pt.txt
88
+ - config/zones_en.txt
88
89
  files:
89
90
  - CHANGELOG.md
90
91
  - Manifest.txt
@@ -95,6 +96,7 @@ files:
95
96
  - config/rounds_es.txt
96
97
  - config/rounds_misc.txt
97
98
  - config/rounds_pt.txt
99
+ - config/zones_en.txt
98
100
  - lib/sportdb/parser.rb
99
101
  - lib/sportdb/parser/lang.rb
100
102
  - lib/sportdb/parser/lexer.rb