sportdb-parser 0.2.0 → 0.2.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: c94dcd42fc13a7043f6b926ca1d947df3199877693b22e53e4f50b5aa522bf5d
-  data.tar.gz: 33eb689dcfb2bab0728c19b7d706da1556ddefafbfbcc6e424ac5bcbe3bccef6
+  metadata.gz: 3657cedc5125ee2515efa8be4a1838d05b7290523dd893f7eba5b87024e71238
+  data.tar.gz: caf6d7e909e17fa0dcabf659ab8f5046ca1940d8f7c1c6f5312e485dc0089384
 SHA512:
-  metadata.gz: 97ef8d76ffa26312d66359f364588af3d7c76a3b0cebd3644b1f1ae775463defa9cb9552b267f26677c2c6f4e9b7b9fe62479dd34a7211fd1a4a3c1b5e9af830
-  data.tar.gz: ca9b56c6c02c132f3924fb40c293e90379812b830a2899e2be02c1d6469a278456c6d68db7f73d5f5fd69b372c958953e3fefd829ac1120cf56b0944176a2b87
+  metadata.gz: 4063565aada304a1eb96009b6fe542392f41a55d4ad4d21b5de156004bd69a055c5f86b076bed1defbe50423c8c891dd538931ea6ca9b8ec41e237c23e699219
+  data.tar.gz: 91f6476810cb6617dfcd703ada57592cd38b87f3b4b9fc6fd4468a9457ff0e6ae6337a4e4f5c782e1b80f5f6b6015d5ce26ed6330915cd67a5fb6606f665017f

data/CHANGELOG.md

@@ -1,4 +1,4 @@
-### 0.2.0
+### 0.2.1
 
 ### 0.0.1 / 2024-07-12
 

data/Rakefile

@@ -26,6 +26,6 @@ Hoe.spec 'sportdb-parser' do
   ]
 
   self.spec_extras = {
-    required_ruby_version: '>= 2.2.2'
+    required_ruby_version: '>= 3.1.0'
   }
 end

data/bin/fbt

@@ -55,7 +55,7 @@ p args
 
 paths =  if args.empty?
             [
-              '../../../openfootball/euro/2020--europe/euro.txt',
+              '../../../openfootball/euro/2021--europe/euro.txt',
               '../../../openfootball/euro/2024--germany/euro.txt',
             ]
          else

data/lib/sportdb/parser/lang.rb

@@ -27,6 +27,12 @@ end
 
 ROUND_RE = %r{^(
 
+   ## add special case for group play-off rounds!
+   ##  group 2 play-off   (e.g. worldcup 1954, 1958)
+     (?:   Group [ ] [a-z0-9]+ [ ]
+           Play-?offs?
+     )
+        |
    # round  - note - requiers number e.g. round 1,2, etc.
    #   note - use 1-9 regex (cannot start with 0) - why? why not?
    #             make week 01 or round 01 or matchday 01 possible?
@@ -46,17 +52,23 @@ ROUND_RE = %r{^(
        |
    ## 1. Round / 2. Round / 3. Round / etc.
    ##  Play-off Round
+   ##  First Round
+   ##  Final Round   (e.g. Worldcup 1950)
       (?:
-           (?: [1-9][0-9]* \.
-                |
-                Play-?off
+           (?: [1-9][0-9]* \.  |
+                Play-?off   |
+                1st | First   |
+                2nd | Second  |
+                Final
            )
              [ ] Round
        )
        |
   ## starting with preliminary
+  #   e.g.  Preliminary round
      (?:  Preliminary  [ ]
-           (?:  Semi-?finals |
+           (?:  Round |
+                Semi-?finals |
                 Final
            )
      )
@@ -110,10 +122,15 @@ ROUND_RE = %r{^(
          Finals?
          |
     ## add replays
-    ##  Final Replay
+    ##  e.g. Final Replay
+    ##       Quarter-finals replays
+    ##       First round replays
      (?:
-        Final
-        [ ] Replay
+        (?: First [ ] Round |
+            Quarter-?finals? |
+            Finals?
+         )
+        [ ] Replays?
       )
 )$}ix
 

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

@@ -1,6 +1,6 @@
-module SportDb 
+module SportDb
 class Parser
-   
+
 
 
 def self.parse_names( txt )
@@ -47,8 +47,8 @@ def self.build_map( lines, downcase: false )
   ##   "may" => 5,
   ##   "june" => 6,     "jun" => 6, ...
   lines.each_with_index.reduce( {} ) do |h,(line,i)|
-    line.each do |name| 
-       h[ downcase ? name.downcase : name ] = i+1 
+    line.each do |name|
+       h[ downcase ? name.downcase : name ] = i+1
     end  ## note: start mapping with 1 (and NOT zero-based, that is, 0)
     h
   end
@@ -109,28 +109,56 @@ DAY_MAP   = build_map( DAY_LINES, downcase: true )
 ## todo - add more date variants !!!!
 
 # e.g. Fri Aug/9  or Fri Aug 9
-DATE_RE = %r{
+DATE_I_RE = %r{
 (?<date>
   \b
      ## optional day name
      ((?<day_name>#{DAY_NAMES})
           [ ]
-     )?    
+     )?
      (?<month_name>#{MONTH_NAMES})
          (?: \/|[ ] )
      (?<day>\d{1,2})
      ## optional year
      (  [ ]
         (?<year>\d{4})
-     )?   
-  \b   
+     )?
+  \b
+)}ix
+
+
+# e.g. 3 June  or 10 June
+DATE_II_RE = %r{
+(?<date>
+  \b
+     ## optional day name
+     ((?<day_name>#{DAY_NAMES})
+          [ ]
+     )?
+     (?<day>\d{1,2})
+         [ ]
+     (?<month_name>#{MONTH_NAMES})
+     ## optional year
+     (  [ ]
+        (?<year>\d{4})
+     )?
+  \b
 )}ix
 
 
+#############################################
+# map tables
+#  note: order matters; first come-first matched/served
+DATE_RE = Regexp.union(
+   DATE_I_RE,
+   DATE_II_RE
+)
+
+
 ###
-#  date duration  
+#  date duration
 #   use - or + as separator
-#    in theory plus( +) only if dates 
+#    in theory plus( +) only if dates
 #     are two days next to each other
 #
 #   otherwise  define new dates type in the future? why? why not?
@@ -147,7 +175,7 @@ DATE_RE = %r{
 #  Jun/25 .. 26        - why? why not???
 #  Jun/25 to 26        - why? why not???
 #  Jun/25 + 26        - add - why? why not???
-#  Sun-Wed Jun/23-26  -  add - why? why not??? 
+#  Sun-Wed Jun/23-26  -  add - why? why not???
 #  Wed+Thu Jun/26+27 2024  -  add - why? why not???
 #
 #  maybe use comman and plus for list of dates
@@ -157,39 +185,89 @@ DATE_RE = %r{
 #   add back optional comma (before) year - why? why not?
 
 
-DURATION_RE =  %r{
+##
+#   todo add plus later on - why? why not?
+
+DURATION_I_RE =  %r{
 (?<duration>
     \b
    ## optional day name
    ((?<day_name1>#{DAY_NAMES})
       [ ]
-   )?    
+   )?
    (?<month_name1>#{MONTH_NAMES})
       (?: \/|[ ] )
    (?<day1>\d{1,2})
    ## optional year
    ( [ ]
       (?<year1>\d{4})
-   )?   
+   )?
 
    ## support + and -  (add .. or such - why??)
-   [ ]*[+-][ ]*   
-  
+   [ ]*[-][ ]*
+
    ## optional day name
    ((?<day_name2>#{DAY_NAMES})
       [ ]
-   )?    
+   )?
    (?<month_name2>#{MONTH_NAMES})
       (?: \/|[ ] )
    (?<day2>\d{1,2})
    ## optional year
    ( [ ]
       (?<year2>\d{4})
-   )?   
-   \b   
+   )?
+   \b
+)}ix
+
+
+###
+#   variant ii
+# e.g. 26 July - 27 July
+
+DURATION_II_RE =  %r{
+(?<duration>
+    \b
+   ## optional day name
+   ((?<day_name1>#{DAY_NAMES})
+      [ ]
+   )?
+   (?<day1>\d{1,2})
+      [ ]
+   (?<month_name1>#{MONTH_NAMES})
+   ## optional year
+   ( [ ]
+      (?<year1>\d{4})
+   )?
+
+   ## support + and -  (add .. or such - why??)
+   [ ]*[-][ ]*
+
+   ## optional day name
+   ((?<day_name2>#{DAY_NAMES})
+      [ ]
+   )?
+   (?<day2>\d{1,2})
+      [ ]
+   (?<month_name2>#{MONTH_NAMES})
+   ## optional year
+   ( [ ]
+      (?<year2>\d{4})
+   )?
+   \b
 )}ix
 
 
+#############################################
+# map tables
+#  note: order matters; first come-first matched/served
+DURATION_RE = Regexp.union(
+   DURATION_I_RE,
+   DURATION_II_RE
+)
+
+
+
 end  #   class Parser
-end  # module SportDb 
-   
+end  # module SportDb
+

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 = 2
-  PATCH = 0
+  PATCH = 1
   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: 0.2.0
+  version: 0.2.1
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-08-22 00:00:00.000000000 Z
+date: 2024-08-24 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cocos
@@ -112,7 +112,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 2.2.2
+      version: 3.1.0
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="