sportdb-parser 0.6.10 → 0.6.12

data/lib/sportdb/parser/racc_tree.rb

@@ -169,10 +169,12 @@ end
 
 
 MatchLine   = Struct.new( :ord, :date, :time, :wday,
-                          :team1, :team2, :score,
-                          :status, 
+                          :team1, :team2, 
+                          :score, :score_note,
+                          :status,  
                           :geo,
-                          :timezone )  do   ## change to geos - why? why not?
+                          :timezone,
+                          :note )  do   ## change to geos - why? why not?
 
   def pretty_print( printer )
     printer.text( "<MatchLine " )

data/lib/sportdb/parser/token-status.rb

@@ -90,63 +90,104 @@ NOTE_RE = %r{
           # add "top-level" NB: version
           ##   with full (end-of) line note - why? why not?
           |
-          (?: originally[ ])? scheduled
-          ## e.g. [originally scheduled to play in Mexico City] 
-          |
           rescheduled
-          ## e.g.  [Rescheduled due to earthquake occurred in Mexico on September 19]
+          ## e.g.  [rescheduled due to earthquake occurred in Mexico on September 19]
+          |
+          declared
+          ## e.g.  [declared void]
           |
           remaining
           ## e.g. [remaining 79']   
           ##      [remaining 84'] 
           ##      [remaining 59']   
           ##      [remaining 5']
-          |
-          played  
-          ## e.g. [played in Macaé-RJ]
-          ##      [played in Caxias do Sul-RS]
-          ##      [played in Sete Lagoas-MG]
-          ##      [played in Uberlândia-MG]
-          ##      [played in Brasília-DF]
-          ##      [played in Vöcklabruck]
-          ##      [played in Pasching]
-          |
-          declared
-          ## e.g.  [declared void]
-          |
-          inter-group
-          ## e.g. [inter-group A-B]
-          ##      [inter-group C-D]
        )
       [ ]
       [^\]]+?    ## slurp all to next ] - (use non-greedy) 
      )
+   )
+   \] 
+}ix    
+
+
+
+SCORE_NOTE_RE = %r{
+    \[ 
+    (?<score_note>
+      (?:   # plain aet e.g. [aet]
+             aet | a\.e\.t\. |
+             after [ ] extra [ -] time
+       )
+      |
+       (?:  # plain penalties e.g. [3-2 pen]
+             \d{1,2}-\d{1,2}
+                [ ]* (?: p|pen )
+       )
+      |
+        (?:  # plain aet with penalties e.g. [aet; 4-3 pen] or [aet, 4-3p]
+              aet [ ]* [,;]
+                [ ]*
+              \d{1,2}-\d{1,2}
+                [ ]* (?: p|pen )
+         )
       |
-     (?:
-       ## starting with in  - do NOT allow digits
-       ##   name starting with in possible - why? why not?
-           in[ ]
-            [^0-9\]]+?
-       ## e.g. [In Estadio La Corregidora] 
-       ##      [in Unidad Deportiva Centenario]
-       ##      [in Estadio Olímpico Universitario]
-       ##      [in Estadio Victoria]
-       ##      [in UD José Brindis]
-       ##      [in Colomos Alfredo "Pistache" Torres stadium]
-     )
-       |
       (?:
          ## e.g. Spain wins on penalties
          ##       1860 München wins on penalties etc.
          ##   must start with digit 1-9 or letter
          ##     todo - add more special chars - why? why not?
-            [1-9\p{L}][0-9\p{L} .-]+?    
-            [ ]wins[ ]on[ ]penalties
-             [^\]]*?   ## use non-greedy
-      )
-   )
-   \] 
-}ix    
+         ##     
+               (?:
+                    aet [ ]*   ## allow space here - why? why not
+                       [,;][ ]
+                )?
+           
+              (?:
+              (?:  # opt 1 - no team listed/named - requires score
+                 (?: won|wins? ) [ ]     ## note - allow won,win or wins
+                (?:   ## score
+                   \d{1,2}-\d{1,2}
+                   [ ]
+                ) 
+                on [ ]  (?: pens | penalties |
+                          aggregate  )   
+               )
+              |
+              (?:  # opt 2 - team required; score optional
+                (?:  ## team required
+                      [1-9\p{L}][0-9\p{L} .-]+?    
+                     [ ]
+                 )
+                 (?: won|wins? ) [ ]     ## won/win/wins
+                 (?:   ## score optional
+                    \d{1,2}-\d{1,2}
+                    [ ]
+                  )?            
+                  on [ ] (?:  pens | penalties |
+                              aggregate  )
+             ###  [^\]]*?   ## allow more? use non-greedy
+          )
+        ))
+         |
+         (?:  ## e.g. agg 3-2 etc.
+             agg [ ] \d{1,2}-\d{1,2}
+         )
+         |
+         (?:   ## e.g. agg 4-4, Ajax win on away goals
+              (?:   ## agg 4-4, optional for now - why? why not? 
+                 agg [ ] \d{1,2}-\d{1,2} 
+                 [ ]*[,;][ ]
+               )?
+             (?:  ## team required
+                      [1-9\p{L}][0-9\p{L} .-]+?    
+                     [ ]
+              )
+              (?: won|wins? ) [ ]     # won/win/wins
+              on [ ] away [ ] goals
+         )
+      )   # score_note ref
+    \]
+}ix
 
 
 end  #  class Lexer

data/lib/sportdb/parser/token.rb

@@ -133,6 +133,7 @@ ANY_RE = %r{
 
 RE = Regexp.union(
                     STATUS_RE,
+                    SCORE_NOTE_RE,
                     NOTE_RE,
                     DURATION_RE,  # note - duration MUST match before date
                     DATE_RE,  ## note - date must go before time (e.g. 12.12. vs 12.12)

data/lib/sportdb/parser/version.rb

@@ -4,7 +4,7 @@ module SportDb
     module Parser
   MAJOR = 0    ## todo: namespace inside version or something - why? why not??
   MINOR = 6
-  PATCH = 10
+  PATCH = 12
   VERSION = [MAJOR,MINOR,PATCH].join('.')
 
   def self.version

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sportdb-parser
 version: !ruby/object:Gem::Version
-  version: 0.6.10
+  version: 0.6.12
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-03-03 00:00:00.000000000 Z
+date: 2025-03-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cocos
@@ -85,6 +85,7 @@ extra_rdoc_files:
 - config/rounds_es.txt
 - config/rounds_misc.txt
 - config/rounds_pt.txt
+- config/zones_en.txt
 files:
 - CHANGELOG.md
 - Manifest.txt
@@ -95,6 +96,7 @@ files:
 - config/rounds_es.txt
 - config/rounds_misc.txt
 - config/rounds_pt.txt
+- config/zones_en.txt
 - lib/sportdb/parser.rb
 - lib/sportdb/parser/lang.rb
 - lib/sportdb/parser/lexer.rb