sportdb-parser 0.3.5 → 0.3.7

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 11c0a0d9844b8412682ef69d487331ddb0bd962ea0fb24cbe7ab885612a5786d
-  data.tar.gz: 8914380e7f5f2379c145fee6134672311c5502362e64529f3b06023e1e39a74f
+  metadata.gz: 6323a39fa903163b57c2416f732be300590217e6818ffbb782c30af06bf52052
+  data.tar.gz: 6491846e993a62bd815853acca1a4790ee2210aec5ab804c788332280f91d861
 SHA512:
-  metadata.gz: d437fb355775c14171aceddef1c901f11db578a774befd1f5d0f297305ad792269a7e5547e0fd35f533a65ae317588ecf72d6c8ca1e73e477e2ea5eddf7dd0b1
-  data.tar.gz: d58d18af259e4d14f7c3c02e09681885bec21500d04d2122d3d50afc4bafc016aa675fbf46580bf344055bee5ad4c4e869b794e7222b7755d7438965cd602143
+  metadata.gz: 70ea59920f73fef1255a4bcd07c3e58571b5ad5a21a35d3775183ca254dfd64d048106066d6cbb53b65692bdb523a08a7d0e2c210e2ef6ee68be1efd2287677b
+  data.tar.gz: a2d61dc09cc2cfdbe3669d7e4376594d63440a50b09c5adc1dfa4e4c3ce983374d5157db1c84a6b44425d4068b041041d7650071f50a1de2f652cc2ab022edc9

data/CHANGELOG.md

@@ -1,4 +1,4 @@
-### 0.3.5
+### 0.3.7
 
 ### 0.0.1 / 2024-07-12
 

data/Manifest.txt

@@ -3,6 +3,11 @@ Manifest.txt
 README.md
 Rakefile
 bin/fbtok
+config/rounds_de.txt
+config/rounds_en.txt
+config/rounds_es.txt
+config/rounds_misc.txt
+config/rounds_pt.txt
 lib/sportdb/parser.rb
 lib/sportdb/parser/fbtok/main.rb
 lib/sportdb/parser/lang.rb

data/config/rounds_de.txt

@@ -0,0 +1,29 @@
+###############
+#  rounds in deutsch (de) / german
+
+Vorrunde
+Gruppenphase
+
+
+Aufstieg
+Endrunde
+Aufstiegsrunde
+Aufstiegsrunde Zone A
+Entscheidung Zone B
+1. Aufstieg
+1. Aufstieg Zone A
+1. Aufstieg Zone B
+2. Aufstieg Zone A
+2. Aufstieg Zone B
+2. Aufstieg 1. Phase
+2. Aufstieg 2. Phase
+2. Aufstieg 3. Phase
+Direkter Aufstieg
+Direkter Abstieg
+5. Platz
+7. Platz
+9. Platz
+11. Platz
+
+Zwischenrunde Gr. B        ## move to group_de - why? why not?
+

data/config/rounds_en.txt

@@ -0,0 +1,12 @@
+##########
+#  note - more english rounds here
+#    remove here if added to regex!!!
+
+Play-in round
+First semifinal
+Second semifinal
+
+Conference Semifinals
+Conference Finals
+Wildcard
+

data/config/rounds_es.txt

@@ -0,0 +1,9 @@
+#########
+#  rounds in español (es) / spanish
+
+Recalificación
+Reclasificación
+
+
+Preclasificación Nacional B
+

data/config/rounds_misc.txt

@@ -0,0 +1,14 @@
+#######################
+#  more rounds  misc(ellaneous)
+
+District West I
+District West II
+District Noord
+District Oost
+District Zuid I
+District Zuid II
+
+Tussenronde
+Replay achtste finale
+Replay kwartfinale
+

data/config/rounds_pt.txt

@@ -0,0 +1,4 @@
+############################
+#  rounds in   português (pt) / portuguese
+
+Troféu do Interior

data/lib/sportdb/parser/lang.rb

@@ -32,7 +32,8 @@ ROUND_RE = %r{^(
    ### note - allow Group ("stand-alone") as "generic" round for now
    ##      BUT do NOT allow Group 1, Group 2, Group A, Group B, etc.
      (?: Group [ ] [A-Z0-9]+ [ ] Play-?offs?  |
-         Group
+         Group (?: [ ] phase)?  |
+         League (?: [ ] phase)?
      )
         |
    # round  - note - requiers number e.g. round 1,2, etc.
@@ -127,6 +128,9 @@ ROUND_RE = %r{^(
    # final
          Finals?
          |
+   # decider e.g. Entscheidungsspiel
+         Decider
+         |
     ## add replays
     ##  e.g. Final Replay
     ##       Quarter-finals replays
@@ -146,8 +150,50 @@ ROUND_RE = %r{^(
 )$}ix
 
 
+####
+#  add more round names in different languages
+#    via txt files
+#
+#  for now must match case - maybe make caseinsensitive later - why? why not?
+def self.read_names( path )
+     txt = read_text( path )
+     names = [] # array of lines (with words)
+     txt.each_line do |line|
+       line = line.strip
+
+       next if line.empty?
+       next if line.start_with?( '#' )   ## skip comments too
+
+       ## strip inline (until end-of-line) comments too
+       ##   e.g. Janvier  Janv  Jan  ## check janv in use??
+       ##   =>   Janvier  Janv  Jan
+
+       line = line.sub( /#.*/, '' ).strip
+       ## pp line
+
+       names << line
+     end
+     names
+end
+
+
+def self.more_round_names
+   @more_round_name ||= begin
+                           names = []
+                           langs = ['en', 'de', 'es', 'pt', 'misc']
+                           ## sort names by length??
+                           langs.each do |lang|
+                             path = "#{SportDb::Module::Parser.root}/config/rounds_#{lang}.txt"
+                             names += read_names( path )
+                           end
+                           names
+                        end
+end
+
+
 def is_round?( text )
-    ROUND_RE.match?( text )
+    ROUND_RE.match?( text ) ||
+    self.class.more_round_names.include?( text )
 end
 
 ##

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

@@ -1,13 +1,13 @@
-module SportDb 
+module SportDb
 class Parser
-  
+
 
     ## todo/check: use ‹› (unicode chars) to mark optional parts in regex constant name - why? why not?
 
     #####
     #  english helpers (penalty, extra time, ...)
     ##   note - p must go last (shortest match)
-    #     pso = penalty shootout 
+    #     pso = penalty shootout
     P_EN  =  '(?: pso | pen\.? | p\.? )'     # e.g. p., p, pen, pen., PSO, etc.
     ET_EN =  '(?: aet | a\.e\.t\.? )'     # note: make last . optional (e.g a.e.t) allowed too
 
@@ -26,7 +26,20 @@ class Parser
             (?<et1>\d{1,2}) - (?<et2>\d{1,2})
                [ ]* #{ET_EN}
                (?=[ \]]|$)
-        )}ix  
+        )}ix
+                ## todo/check:  remove loakahead assertion here - why require space?
+                ## note: \b works only after non-alphanum e.g. )
+
+
+    ##  note: allow SPECIAL with penalty only
+    ##      3-4 pen.
+    SCORE__P__RE = %r{
+        (?<score>
+           \b
+              (?<p1>\d{1,2}) - (?<p2>\d{1,2})
+                [ ]* #{P_EN}
+                (?=[ \]]|$)
+         )}ix
                 ## todo/check:  remove loakahead assertion here - why require space?
                 ## note: \b works only after non-alphanum e.g. )
 
@@ -89,8 +102,8 @@ class Parser
 
     ## e.g. 2-1 (1-1) or
     ##      2-1
- 
-    SCORE__FT_HT__RE = %r{      
+
+    SCORE__FT_HT__RE = %r{
             (?<score>
               \b
               (?<ft1>\d{1,2}) - (?<ft2>\d{1,2})
@@ -104,18 +117,18 @@ class Parser
                     ## note: \b works only after non-alphanum e.g. )
 
 
-                    
+
 #############################################
-# map tables 
+# map tables
 #  note: order matters; first come-first matched/served
 
-SCORE_RE = Regexp.union( 
+SCORE_RE = Regexp.union(
   SCORE__P_ET_FT_HT__RE,  # e.g. 5-1 pen. 2-2 a.e.t. (1-1, 1-0)
   SCORE__P_FT_HT__RE,     # e.g. 5-1 pen. (1-1)
   SCORE__P_ET__RE,        # e.g. 2-2 a.e.t.  or  5-1 pen. 2-2 a.e.t.
-  SCORE__FT_HT__RE        # e.g. 1-1 (1-0)
+  SCORE__P__RE,           # e.g. 5-1 pen.
+  SCORE__FT_HT__RE,        # e.g. 1-1 (1-0) or 1-1  -- note - must go last!!!
 )
 
 end  #  class Parser
-end  # module SportDb 
-  
+end  # module SportDb

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

@@ -57,13 +57,18 @@ TEXT_RE = %r{
                   )?
                 (?:
                   \p{L} |
-                  [&/']
+                  [&/'°]
                     |
                  (?:
                    \d+
-                   (?![0-9.:h'/+-])
+                   (?!
+                     [0-9h'+-] |    ## protected break on 12h / 12' / 1-1
+                                    ##  check usege for 3+4 - possible? where ? why?
+                     (?:[.:]\d)     ## protected/exclude/break on 12.03 / 12:03
+                    )
                    ## negative lookahead for numbers
                    ##   note - include digits itself!!!
+                   ##   note - remove / (slash) e.g. allows UDI'19/Beter Bed
                  )|
                  \.
                )

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 = 3
-  PATCH = 5
+  PATCH = 7
   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: 0.3.5
+  version: 0.3.7
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-10-09 00:00:00.000000000 Z
+date: 2024-10-13 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cocos
@@ -81,12 +81,22 @@ extra_rdoc_files:
 - CHANGELOG.md
 - Manifest.txt
 - README.md
+- config/rounds_de.txt
+- config/rounds_en.txt
+- config/rounds_es.txt
+- config/rounds_misc.txt
+- config/rounds_pt.txt
 files:
 - CHANGELOG.md
 - Manifest.txt
 - README.md
 - Rakefile
 - bin/fbtok
+- config/rounds_de.txt
+- config/rounds_en.txt
+- config/rounds_es.txt
+- config/rounds_misc.txt
+- config/rounds_pt.txt
 - lib/sportdb/parser.rb
 - lib/sportdb/parser/fbtok/main.rb
 - lib/sportdb/parser/lang.rb