sportdb-formats 1.0.2 → 1.0.3

data/lib/sportdb/formats/score/score_formats.rb

@@ -93,12 +93,51 @@ module ScoreFormats
              (?=[ \]]|$)/xi    ## todo/check:  remove loakahead assertion here - why require space?
                                ## note: \b works only after non-alphanum e.g. )
 
+
     #####
     #  deutsch / german helpers (penalty, extra time, ...)
     ## todo add more marker e.g. im Elf. or such!!!
     P_DE  =  '(?: ie | i\.e\.? )'     # e.g. iE, i.E., i.E etc.
     ET_DE =  '(?: nv | n\.v\.? )'     # e.g. nV, n.V., n.V etc.
 
+
+    ## support alternate all-in-one score e.g.
+    ##     i.E. 2:4, n.V. 3:3 (1:1, 1:1)  or
+    ##               n.V. 3:2 (2:2, 1:2)
+    DE__P_ET_FT_HT__RE = /\b
+                     (?:
+                     #{P_DE}
+                      [ ]*
+                     (?<score1p>\d{1,2})
+                      [ ]* : [ ]*
+                     (?<score2p>\d{1,2})
+                      [ ]* (?:, [ ]*)?
+                     )?   # note: make penalty (P) score optional for now
+                      #{ET_DE}
+                      [ ]*
+                     (?<score1et>\d{1,2})
+                      [ ]* : [ ]*
+                     (?<score2et>\d{1,2})
+                      [ ]*
+                    \(
+                 [ ]*
+             (?<score1>\d{1,2})
+                  [ ]* : [ ]*
+             (?<score2>\d{1,2})
+                  [ ]*
+               (?:
+                   , [ ]*
+                   (?:
+                    (?<score1i>\d{1,2})
+                      [ ]* : [ ]*
+                    (?<score2i>\d{1,2})
+                      [ ]*
+                   )?
+               )?    # note: make half time (HT) score optional for now
+             \)
+            (?=[ \]]|$)
+              /xi
+
     ## support all-in-one "literal form e.g.
     #  2:2 (1:1, 1:0) n.V. 5:1 i.E.   or
     #  2-2 (1-1, 1-0) n.V. 5-1 i.E.
@@ -168,7 +207,8 @@ FORMATS_EN = [
 ]
 
 FORMATS_DE = [
-  [ DE__ET_FT_HT_P__RE, '[SCORE.DE__ET_(FT_HT?)_P?]' ], # e.g.  2:2 (1:1, 1:0) n.V. 5:1 i.E.
+  [ DE__ET_FT_HT_P__RE, '[SCORE.DE__ET_(FT_HT?)_P?]' ], # e.g. 2:2 (1:1, 1:0) n.V. 5:1 i.E.
+  [ DE__P_ET_FT_HT__RE, '[SCORE.DE__P?_ET_(FT_HT?)]' ], # e.g. i.E. 2:4, n.V. 3:3 (1:1, 1:1)
   [ DE__FT_HT__RE,      '[SCORE.DE__FT_(HT)?]'       ], # e.g. 1:1 (1:0)
 ]
 

data/lib/sportdb/formats/team/club_reader_props.rb

@@ -27,7 +27,7 @@ class ClubPropsReader
   def parse
     recs = parse_csv( @txt )
     recs.each do |rec|
-      name = rec[:name]
+      name = rec['Name']
       if name.nil?
         puts "** !!! ERROR !!! Name column required / missing / NOT found in row:"
         pp rec
@@ -62,8 +62,8 @@ class ClubPropsReader
         ## todo/fix:   only updated "on-demand" from in-memory struct/records!!!!
 
         ##  update attributes
-        club_rec.key  = rec[:key]      if rec[:key]
-        club_rec.code = rec[:code]     if rec[:code]
+        club_rec.key  = rec['Key']      if rec['Key']
+        club_rec.code = rec['Code']     if rec['Code']
         ## todo/fix: add (some) more props e.g. address, web, etc.
       end
     end

data/lib/sportdb/formats/version.rb

@@ -2,11 +2,12 @@
 
 
 module SportDb
+  module Module
 module Formats
 
   MAJOR = 1    ## todo: namespace inside version or something - why? why not??
   MINOR = 0
-  PATCH = 2
+  PATCH = 3
   VERSION = [MAJOR,MINOR,PATCH].join('.')
 
   def self.version
@@ -22,4 +23,5 @@ module Formats
   end
 
 end # module Formats
+end # module Module
 end # module SportDb

data/test/helper.rb

@@ -147,7 +147,7 @@ def parse_json( str )
 end
 
 def read_test( path )
-  blocks = read_blocks( "../football.txt/#{path}" )
+  blocks = read_blocks( "#{SportDb::Test.data_dir}/football.txt/#{path}" )
 
   if blocks.size == 2
     [blocks[0], parse_json( blocks[1] )]

data/test/test_club_reader_props.rb

@@ -29,8 +29,8 @@ TXT
     recs = parse_csv( ENG_CLUBS_PROPS_TXT )
     pp recs
 
-    assert_equal [{ key: 'chelsea', name: 'Chelsea FC', code: 'CHE' },
-                  { key: 'arsenal', name: 'Arsenal FC', code: 'ARS' }], recs[0..1]
+    assert_equal [{ 'Key' => 'chelsea', 'Name' => 'Chelsea FC', 'Code' => 'CHE' },
+                  { 'Key' => 'arsenal', 'Name' => 'Arsenal FC', 'Code' => 'ARS' }], recs[0..1]
   end
 
   CLUBS = SportDb::Import.catalog.clubs

data/test/test_csv_match_parser.rb

@@ -0,0 +1,114 @@
+# encoding: utf-8
+
+###
+#  to run use
+#     ruby -I ./lib -I ./test test/test_csv_match_parser.rb
+
+
+require 'helper'
+
+
+class TestCsvMatchParser < MiniTest::Test
+
+##
+# Div,Date,HomeTeam,AwayTeam,FTHG,FTAG,FTR,HTHG,HTAG,HTR,
+#   Referee,HS,AS,HST,AST,HF,AF,HC,AC,HY,AY,HR,AR,
+#    B365H,B365D,B365A,BWH,BWD,BWA,IWH,IWD,IWA,LBH,LBD,LBA,PSH,PSD,PSA,
+#    WHH,WHD,WHA,VCH,VCD,VCA,
+#    Bb1X2,BbMxH,BbAvH,BbMxD,BbAvD,BbMxA,BbAvA,BbOU,BbMx>2.5,BbAv>2.5,BbMx<2.5,BbAv<2.5,
+#    BbAH,BbAHh,BbMxAHH,BbAvAHH,BbMxAHA,BbAvAHA,PSCH,PSCD,PSCA
+  def test_eng_filters
+    path = "#{SportDb::Test.data_dir}/england/2017-18/E0.csv"
+
+    matches = SportDb::CsvMatchParser.read( path,
+                                            filters: { 'HomeTeam' => 'Arsenal' } )
+
+    pp path
+    pp matches[0..2]
+
+    m=matches[0]
+    assert_equal '2017-08-11', m.date
+    assert_equal 4, m.score1
+    assert_equal 3, m.score2
+    assert_equal 2, m.score1i
+    assert_equal 2, m.score2i
+    assert_equal 'Arsenal',    m.team1
+    assert_equal 'Leicester',  m.team2
+
+    m=matches[1]
+    assert_equal '2017-09-09', m.date
+    assert_equal 3, m.score1
+    assert_equal 0, m.score2
+    assert_equal 2, m.score1i
+    assert_equal 0, m.score2i
+    assert_equal 'Arsenal',     m.team1
+    assert_equal 'Bournemouth', m.team2
+  end
+
+
+  def test_eng_headers
+    path = "#{SportDb::Test.data_dir}/england/2017-18/E0.csv"
+
+    headers = { team1:  'HomeTeam',
+                team2:  'AwayTeam',
+                date:   'Date',
+                score1: 'FTHG',
+                score2: 'FTAG' }
+
+    matches = SportDb::CsvMatchParser.read( path, headers: headers )
+
+    pp path
+    pp matches[0..2]
+
+    m=matches[0]
+    assert_equal '2017-08-11', m.date
+    assert_equal 4, m.score1
+    assert_equal 3, m.score2
+    assert_nil   m.score1i      ## todo/fix: missing half time (ht) score !!!!
+    assert_nil   m.score2i      ## todo/fix: missing half time (ht) score !!!!
+    assert_equal 'Arsenal',   m.team1
+    assert_equal 'Leicester', m.team2
+
+    m=matches[1]
+    assert_equal '2017-08-12', m.date
+    assert_equal 0, m.score1
+    assert_equal 2, m.score2
+    assert_nil   m.score1i      ## todo/fix: missing half time (ht) score !!!!
+    assert_nil   m.score2i      ## todo/fix: missing half time (ht) score !!!!
+    assert_equal 'Brighton', m.team1
+    assert_equal 'Man City', m.team2
+  end
+
+
+###
+# Country,League,Season,Date,Time,Home,Away,HG,AG,
+#  Res,PH,PD,PA,MaxH,MaxD,MaxA,AvgH,AvgD,AvgA
+  def test_at
+    path = "#{SportDb::Test.data_dir}/austria/AUT.csv"
+
+    matches = SportDb::CsvMatchParser.read( path, filters: { 'Season' => '2017/2018' } )
+
+    pp matches[0..2]
+    pp path
+
+    m=matches[0]
+    assert_equal '2017-07-22', m.date
+    assert_equal 2, m.score1
+    assert_equal 2, m.score2
+    assert_nil      m.score1i  ## todo/fix: missing half time (ht) score !!!!
+    assert_nil      m.score2i  ## todo/fix: missing half time (ht) score !!!!
+    assert_equal 'Rapid Vienna', m.team1
+    assert_equal 'Mattersburg',  m.team2
+
+    m=matches[1]
+    assert_equal '2017-07-22', m.date
+    assert_equal 0, m.score1
+    assert_equal 2, m.score2
+    assert_nil   m.score1i    ## todo/fix: missing half time (ht) score !!!!
+    assert_nil   m.score2i    ## todo/fix: missing half time (ht) score !!!!
+    assert_equal 'AC Wolfsberger', m.team1
+    assert_equal 'Salzburg',       m.team2
+  end
+
+
+end # class TestCsvMatchParser

data/test/test_csv_match_parser_utils.rb

@@ -0,0 +1,20 @@
+# encoding: utf-8
+
+###
+#  to run use
+#     ruby -I ./lib -I ./test test/test_csv_match_parser_utils.rb
+
+
+require 'helper'
+
+class TestCsvMatchParserUtils < MiniTest::Test
+
+  def test_find_seasons
+     pp = SportDb::CsvMatchParser.find_seasons( "#{SportDb::Test.data_dir}/dl/AUT.csv" )  ## defaults to col: 'Season', col_sep: ','
+     pp = SportDb::CsvMatchParser.find_seasons( "#{SportDb::Test.data_dir}/dl/Bundesliga_1963_2014.csv", col: 'Saison', sep: ';'  )
+     assert true
+  end
+
+end # class TestCsvMatchParserUtils
+
+

data/test/test_csv_reader.rb

@@ -13,7 +13,7 @@ class TestCsvReader < MiniTest::Test
     recs = parse_csv( <<TXT )
 ### World Countries
 
-key, fifa, name
+Key, Code, Name
 af,  AFG,  Afghanistan
 al,  ALB,  Albania
 dz,  ALG,  Algeria
@@ -21,10 +21,10 @@ as,  ASA,  American Samoa (US)
 TXT
 
    pp recs
-   assert_equal  [{ key:'af', fifa:'AFG', name:'Afghanistan'},
-                  { key:'al', fifa:'ALB', name:'Albania'},
-                  { key:'dz', fifa:'ALG', name:'Algeria'},
-                  { key:'as', fifa:'ASA', name:'American Samoa (US)'},
+   assert_equal  [{ 'Key' => 'af', 'Code' => 'AFG', 'Name' => 'Afghanistan'},
+                  { 'Key' => 'al', 'Code' => 'ALB', 'Name' => 'Albania'},
+                  { 'Key' => 'dz', 'Code' => 'ALG', 'Name' => 'Algeria'},
+                  { 'Key' => 'as', 'Code' => 'ASA', 'Name' => 'American Samoa (US)'},
                  ], recs[0..3]
   end
 

data/test/test_datafile.rb

@@ -27,36 +27,4 @@ class TestDatafile < MiniTest::Test
     assert Datafile::Package.match_exclude?( '/.build/leagues.txt' )
     assert Datafile::Package.match_exclude?( '/.git/leagues.txt' )
   end
-
-
-  CLUBS_DIR   = '../../../openfootball/clubs'    ## source repo directory path
-  LEAGUES_DIR = '../../../openfootball/leagues'
-
-  def test_bundle
-    datafiles = SportDb::Package.find_clubs( CLUBS_DIR )
-    pp datafiles
-
-    ## todo/fix: turn into Datafile::Bundle.new  and Bundle#write/save -why? why not?
-    bundle = Datafile::Bundle.new( './tmp/clubs.txt' )
-    bundle.write <<TXT
-##########################################
-# auto-generated all-in-one single datafile clubs.txt bundle
-#    on #{Time.now} from #{datafiles.size} datafile(s)
-TXT
-    bundle.write datafiles
-    bundle.close
-  end
-
-  def test_bundle_old
-    datafiles = SportDb::Package.find_leagues( LEAGUES_DIR )
-    pp datafiles
-
-    Datafile.write_bundle( './tmp/leagues.txt',
-                           datafiles: datafiles,
-                           header: <<TXT )
-##########################################
-# auto-generated all-in-one single datafile leagues.txt bundle
-#    on #{Time.now} from #{datafiles.size} datafile(s)
-TXT
-  end
 end   # class TestDatafile

data/test/test_datafile_package.rb

@@ -0,0 +1,46 @@
+# encoding: utf-8
+
+###
+#  to run use
+#     ruby -I ./lib -I ./test test/test_datafile_package.rb
+
+
+require 'helper'
+
+
+class TestDatafilePackage < MiniTest::Test
+
+  def test_read
+    [Datafile::DirPackage.new( '../../../openfootball/england' ),
+     Datafile::ZipPackage.new( 'tmp/england-master.zip' )
+    ].each do |eng|
+       assert eng.find( '2015-16/.conf.txt' ).read.start_with?( '= English Premier League 2015/16' )
+       assert eng.find( '2017-18/.conf.txt' ).read.start_with?( '= English Premier League 2017/18' )
+       assert eng.find( '2015-16/1-premierleague-i.txt' ).read.start_with?( '= English Premier League 2015/16' )
+    end
+
+    [Datafile::DirPackage.new( '../../../openfootball/austria' ),
+     Datafile::ZipPackage.new( 'tmp/austria-master.zip' )
+    ].each do |at|
+       assert at.find( '2018-19/.conf.txt' ).read.start_with?( '= Österr. Bundesliga 2018/19' )
+    end
+
+    puts "DirPackage:"
+    pack = Datafile::DirPackage.new( '../../../openfootball/austria' )
+    puts pack.name
+    puts "  entries:"
+    pack.each( pattern: /\.txt$/ ) do |entry|
+      puts entry.name
+    end
+
+    puts "ZipPackage:"
+    pack = Datafile::ZipPackage.new( 'tmp/austria-master.zip' )
+    puts pack.name
+    puts "  entries:"
+    pack.each( pattern: /\.txt$/ ) do |entry|
+      puts entry.name
+    end
+
+  end  # method test_read
+
+end  # class TestDatafilePackage

data/test/test_package.rb

@@ -7,40 +7,72 @@
 
 require 'helper'
 
-
 class TestPackage < MiniTest::Test
 
-  def test_read
-    [Datafile::DirPackage.new( '../../../openfootball/england' ),
-     Datafile::ZipPackage.new( 'tmp/england-master.zip' )
-    ].each do |eng|
-       assert eng.find( '2015-16/.conf.txt' ).read.start_with?( '= English Premier League 2015/16' )
-       assert eng.find( '2017-18/.conf.txt' ).read.start_with?( '= English Premier League 2017/18' )
-       assert eng.find( '2015-16/1-premierleague-i.txt' ).read.start_with?( '= English Premier League 2015/16' )
-    end
+  AUSTRIA_DIR     = '../../../openfootball/austria'
+  AUSTRIA_CSV_DIR = '../../../footballcsv/austria'
 
-    [Datafile::DirPackage.new( '../../../openfootball/austria' ),
-     Datafile::ZipPackage.new( 'tmp/austria-master.zip' )
-    ].each do |at|
-       assert at.find( '2018-19/.conf.txt' ).read.start_with?( '= Österr. Bundesliga 2018/19' )
-    end
+  ENGLAND_DIR     = '../../../openfootball/england'
+  ENGLAND_CSV_DIR = '../../../footballcsv/england'
+
+
+  def test_match_by
 
-    puts "DirPackage:"
-    pack = Datafile::DirPackage.new( '../../../openfootball/austria' )
-    puts pack.name
-    puts "  entries:"
-    pack.each( pattern: /\.txt$/ ) do |entry|
-      puts entry.name
+    [ AUSTRIA_CSV_DIR,
+      ENGLAND_CSV_DIR,
+      "#{SportDb::Test.data_dir}/packages/test-levels",
+    ].each do |path|
+      puts
+      puts "match (csv) in #{path}:"
+      pack = SportDb::Package.new( path )
+      pp pack.match_by_season_dir( format: 'csv' )
+      puts "---"
+      pp pack.match_by_season( format: 'csv' )
     end
 
-    puts "ZipPackage:"
-    pack = Datafile::ZipPackage.new( 'tmp/austria-master.zip' )
-    puts pack.name
-    puts "  entries:"
-    pack.each( pattern: /\.txt$/ ) do |entry|
-      puts entry.name
+    [ AUSTRIA_DIR,
+      ENGLAND_DIR,
+      "#{SportDb::Test.data_dir}/packages/test-levels",
+    ].each do |path|
+      puts
+      puts "match (txt) in #{path}:"
+      pack = SportDb::Package.new( path )
+      pp pack.match_by_season_dir
+      puts "---"
+      pp pack.match_by_season( start: '2000' )   ## note: try with start season filter!!
     end
+  end
+
+
+  CLUBS_DIR   = '../../../openfootball/clubs'    ## source repo directory path
+  LEAGUES_DIR = '../../../openfootball/leagues'
+
+  def test_bundle
+    datafiles = SportDb::Package.find_clubs( CLUBS_DIR )
+    pp datafiles
+
+    ## todo/fix: turn into Datafile::Bundle.new  and Bundle#write/save -why? why not?
+    bundle = Datafile::Bundle.new( './tmp/clubs.txt' )
+    bundle.write <<TXT
+##########################################
+# auto-generated all-in-one single datafile clubs.txt bundle
+#    on #{Time.now} from #{datafiles.size} datafile(s)
+TXT
+    bundle.write datafiles
+    bundle.close
+  end
+
+  def test_bundle_old
+    datafiles = SportDb::Package.find_leagues( LEAGUES_DIR )
+    pp datafiles
 
-  end  # method test_read
+    Datafile.write_bundle( './tmp/leagues.txt',
+                           datafiles: datafiles,
+                           header: <<TXT )
+##########################################
+# auto-generated all-in-one single datafile leagues.txt bundle
+#    on #{Time.now} from #{datafiles.size} datafile(s)
+TXT
+  end
 
-end  # class TestPackage
+end # class TestPackage