sportdb-formats 0.4.0 → 1.0.0

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sportdb-formats
 version: !ruby/object:Gem::Version
-  version: 0.4.0
+  version: 1.0.0
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2020-04-19 00:00:00.000000000 Z
+date: 2020-05-10 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: alphabets
@@ -16,28 +16,28 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.1.3
+        version: 1.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.1.3
+        version: 1.0.0
 - !ruby/object:Gem::Dependency
   name: date-formats
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.2.4
+        version: 1.0.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.2.4
+        version: 1.0.0
 - !ruby/object:Gem::Dependency
   name: csvreader
   requirement: !ruby/object:Gem::Requirement
@@ -58,14 +58,14 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.0.1
+        version: 0.1.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.0.1
+        version: 0.1.0
 - !ruby/object:Gem::Dependency
   name: rubyzip
   requirement: !ruby/object:Gem::Requirement
@@ -122,32 +122,52 @@ files:
 - README.md
 - Rakefile
 - lib/sportdb/formats.rb
+- lib/sportdb/formats/config.rb
 - lib/sportdb/formats/datafile.rb
+- lib/sportdb/formats/datafile_package.rb
 - lib/sportdb/formats/goals.rb
+- lib/sportdb/formats/match/conf_parser.rb
+- lib/sportdb/formats/match/mapper.rb
+- lib/sportdb/formats/match/mapper_teams.rb
+- lib/sportdb/formats/match/match_parser.rb
+- lib/sportdb/formats/match/match_parser_auto_conf.rb
+- lib/sportdb/formats/name_helper.rb
 - lib/sportdb/formats/outline_reader.rb
 - lib/sportdb/formats/package.rb
-- lib/sportdb/formats/scores.rb
+- lib/sportdb/formats/parser_helper.rb
+- lib/sportdb/formats/score/score_formats.rb
+- lib/sportdb/formats/score/score_parser.rb
 - lib/sportdb/formats/season_utils.rb
-- lib/sportdb/formats/structs/club.rb
 - lib/sportdb/formats/structs/country.rb
+- lib/sportdb/formats/structs/group.rb
 - lib/sportdb/formats/structs/league.rb
 - lib/sportdb/formats/structs/match.rb
 - lib/sportdb/formats/structs/matchlist.rb
 - lib/sportdb/formats/structs/round.rb
 - lib/sportdb/formats/structs/season.rb
 - lib/sportdb/formats/structs/standings.rb
+- lib/sportdb/formats/structs/team.rb
 - lib/sportdb/formats/structs/team_usage.rb
 - lib/sportdb/formats/version.rb
 - test/helper.rb
-- test/test_club_helpers.rb
 - test/test_clubs.rb
+- test/test_conf.rb
 - test/test_csv_reader.rb
 - test/test_datafile.rb
-- test/test_datafile_match.rb
 - test/test_goals.rb
 - test/test_match.rb
+- test/test_match_auto.rb
+- test/test_match_auto_champs.rb
+- test/test_match_auto_euro.rb
+- test/test_match_auto_worldcup.rb
+- test/test_match_champs.rb
+- test/test_match_eng.rb
+- test/test_match_euro.rb
+- test/test_match_worldcup.rb
+- test/test_name_helper.rb
 - test/test_outline_reader.rb
 - test/test_package.rb
+- test/test_package_match.rb
 - test/test_scores.rb
 - test/test_season.rb
 homepage: https://github.com/sportdb/sport.db

data/lib/sportdb/formats/scores.rb

@@ -1,253 +0,0 @@
-# encoding: utf-8
-
-module SportDb
-
-class ScoresFinder
-
-  include LogUtils::Logging
-
-
-  ## e.g. 3-4 pen. 2-2 a.e.t. (1-1, 1-1)  or
-  ##      3-4 pen. 2-2 a.e.t. (1-1, )
-  EN__P_ET_FT_HT__RE = /\b
-             (?<score1p>\d{1,2})
-                   -
-             (?<score2p>\d{1,2})
-                 \s*                # allow optional spaces
-             (?:p|pen\.?|pso)       # e.g. pen, pen., PSO, p etc.
-                 \s*                # allow optional spaces
-             (?<score1et>\d{1,2})
-                   -
-             (?<score2et>\d{1,2})
-                 \s*                # allow optional spaces
-             (?:aet|a\.e\.t\.)
-                 \s*                # allow optional spaces
-                 \(
-            (?<score1>\d{1,2})
-                 -
-            (?<score2>\d{1,2})
-                \s*
-                ,
-                \s*
-              (?:
-                  (?<score1i>\d{1,2})
-                    -
-                  (?<score2i>\d{1,2})
-              )?    # note: make half time (HT) score optional for now
-            \)
-           (?=[\s\]]|$)/xi    ## todo/check:  remove loakahead assertion here - why require space?
-                             ## note: \b works only after non-alphanum e.g. )
-
-
-  ## e.g. 2-1 a.e.t. (1-1, 0-0)  or
-  ##      2-1 a.e.t. (1-1, )
-  EN__ET_FT_HT__RE = /\b
-             (?<score1et>\d{1,2})
-                   -
-             (?<score2et>\d{1,2})
-                 \s*                # allow optional spaces
-             (?:aet|a\.e\.t\.)
-                 \s*                # allow optional spaces
-                 \(
-            (?<score1>\d{1,2})
-                 -
-            (?<score2>\d{1,2})
-                \s*
-                ,
-                \s*
-              (?:
-                (?<score1i>\d{1,2})
-                  -
-                (?<score2i>\d{1,2})
-              )?    # note: make half time (HT) score optional for now
-            \)
-           (?=[\s\]]|$)/xi    ## todo/check:  remove loakahead assertion here - why require space?
-                             ## note: \b works only after non-alphanum e.g. )
-
-
-  ## e.g. 2-1 (1-1)
-  EN__FT_HT__RE = /\b
-            (?<score1>\d{1,2})
-               -
-            (?<score2>\d{1,2})
-             \s*
-             \(
-            (?<score1i>\d{1,2})
-               -
-            (?<score2i>\d{1,2})
-               \)
-           (?=[\s\]]|$)/x    ## todo/check:  remove loakahead assertion here - why require space?
-                             ## note: \b works only after non-alphanum e.g. )
-
-
-
-###################
-# more
-
-  # e.g. 1:2 or 0:2 or 3:3  or
-  #      1-1 or 0-2 or 3-3  or
-  #      1x1 or 1X1 or 0x2 or 3x3   -- used in Brazil / Portugal
-  FT_RE = /\b
-            (?<score1>\d{1,2})
-               [:\-xX]
-            (?<score2>\d{1,2})
-           \b/x
-
-
-  # e.g. 1:2nV  => after extra time a.e.t
-
-  # note: possible ending w/ . -> thus cannot use /b will not work w/ .; use zero look-ahead
-  ET_RE = /\b
-               (?<score1>\d{1,2})
-                  [:\-xX]
-               (?<score2>\d{1,2})
-                  \s?                # allow optional space
-                (?:nv|n\.v\.|aet|a\.e\.t\.)        # allow optional . e.g. nV or n.V.
-               (?=[\s\)\]]|$)/xi
-
-  ## todo: add/allow english markers e.g. pen or p  ??
-
-  # e.g. 5:4iE  => penalty / after penalty a.p
-
-
-  # note: possible ending w/ . -> thus cannot use /b will not work w/ .; use zero look-ahead
-  P_RE = /\b
-              (?<score1>\d{1,2})
-                  [:\-xX]
-              (?<score2>\d{1,2})
-                  \s?                # allow optional space
-                (?:iE|i\.E\.|p|pen|PSO)       # allow optional . e.g. iE or i.E.
-              (?=[\s\)\]]|$)/xi
-
-
-  ## todo: allow all-in-one "literal form a la kicker" e.g.
-  #  2:2 (1:1, 1:0) n.V. 5:1 i.E.
-
-  def initialize
-    # nothing here for now
-  end
-
-  def find!( line, opts={} )
-
-    ### fix: add and match all-in-one literal first, followed by
-
-    # note: always call after find_dates !!!
-    #  scores match date-like patterns!!  e.g. 10-11  or 10:00 etc.
-    #   -- note: score might have two digits too
-
-    ### fix: depending on language allow 1:1 or 1-1
-    ##   do NOT allow mix and match
-    ##  e.g. default to en is  1-1
-    ##    de is 1:1 etc.
-
-
-    # extract score from line
-    # and return it
-    # note: side effect - removes date from line string
-
-
-    score1i   = nil    # half time (ht) scores
-    score2i   = nil
-
-    score1   = nil    # full time (ft) scores
-    score2   = nil
-
-    score1et = nil    # extra time (et) scores
-    score2et = nil
-
-    score1p  = nil   # penalty (p) scores
-    score2p  = nil
-
-
-    if (m = EN__P_ET_FT_HT__RE.match( line ))
-      if m[:score1i] && m[:score2i]   ## note: half time (HT) score is optional now
-        score1i   = m[:score1i].to_i
-        score2i   = m[:score2i].to_i
-      end
-
-      score1    = m[:score1].to_i
-      score2    = m[:score2].to_i
-      score1et  = m[:score1et].to_i
-      score2et  = m[:score2et].to_i
-      score1p   = m[:score1p].to_i
-      score2p   = m[:score2p].to_i
-
-      logger.debug "   score.en__p_et_ft_ht: >#{score1p}-#{score2p} pen. #{score1et}-#{score2et} a.e.t. (#{score1}-#{score2}, #{score1i}-#{score2i})<"
-
-      line.sub!( m[0], '[SCORES.EN__P_ET_FT_HT]' )
-
-    elsif (m = EN__ET_FT_HT__RE.match( line ))
-      if m[:score1i] && m[:score2i]   ## note: half time (HT) score is optional now
-        score1i   = m[:score1i].to_i
-        score2i   = m[:score2i].to_i
-      end
-
-      score1    = m[:score1].to_i
-      score2    = m[:score2].to_i
-      score1et  = m[:score1et].to_i
-      score2et  = m[:score2et].to_i
-
-      logger.debug "   score.en__et_ft_ht: >#{score1et}-#{score2et} a.e.t. (#{score1}-#{score2}, #{score1i}-#{score2i})<"
-
-      line.sub!( m[0], '[SCORES.EN__ET_FT_HT]' )
-
-    elsif (m = EN__FT_HT__RE.match( line ))
-      score1i = m[:score1i].to_i
-      score2i = m[:score2i].to_i
-      score1  = m[:score1].to_i
-      score2  = m[:score2].to_i
-
-      logger.debug "   score.en__ft_ht: >#{score1}-#{score2} (#{score1i}-#{score2i})<"
-
-      line.sub!( m[0], '[SCORES.EN__FT_HT]' )
-    else
-      #######################################################
-      ## try "standard" generic patterns for fallbacks
-
-      if (m = ET_RE.match( line ))
-        score1et = m[:score1].to_i
-        score2et = m[:score2].to_i
-
-        logger.debug "   score.et: >#{score1et}-#{score2et}<"
-
-        line.sub!( m[0], '[SCORE.ET]' )
-      end
-
-      if (m = P_RE.match( line ))
-        score1p = m[:score1].to_i
-        score2p = m[:score2].to_i
-
-        logger.debug "   score.p: >#{score1p}-#{score2p}<"
-
-        line.sub!( m[0], '[SCORE.P]' )
-      end
-
-      ## let full time (ft) standard regex go last - has no marker
-
-      if (m = FT_RE.match( line ))
-        score1 = m[:score1].to_i
-        score2 = m[:score2].to_i
-
-        logger.debug "   score: >#{score1}-#{score2}<"
-
-        line.sub!( m[0], '[SCORE]' )
-      end
-    end
-
-    ## todo: how to handle game w/o extra time
-    #   but w/ optional penalty ???  e.g. used in copa liberatores, for example
-    #    retrun 0,0 or nil,nil for extra time score ?? or -1, -1 ??
-    #    for now use nil,nil
-
-    scores = []
-    scores += [score1i, score2i]     if score1p || score2p || score1et || score2et || score1 || score2 || score1i || score2i
-    scores += [score1, score2]       if score1p || score2p || score1et || score2et || score1 || score2
-    scores += [score1et, score2et]   if score1p || score2p || score1et || score2et
-    scores += [score1p,  score2p]    if score1p || score2p
-
-    scores
-  end
-
-end # class ScoresFinder
-
-end # module SportDb

data/lib/sportdb/formats/structs/club.rb

@@ -1,213 +0,0 @@
-# encoding: utf-8
-
-module SportDb
-  module Import
-
-########
-# more attribs - todo/fix - also add "upstream" to struct & model!!!!!
-#   district, geos, year_end, country, etc.
-
-class Club
-
-  def self.create( **kwargs )
-    new.update( kwargs )
-  end
-
-  def update( **kwargs )
-    @name        = kwargs[:name]       if kwargs.has_key? :name
-    @alt_names   = kwargs[:alt_names]  if kwargs.has_key? :alt_names
-    @city        = kwargs[:city]       if kwargs.has_key? :city
-    ## todo/fix:  use city struct - why? why not?
-    ## todo/fix: add country too  or report unused keywords / attributes - why? why not?
-
-    self   ## note - MUST return self for chaining
-  end
-
-
-  ##  todo: use just names for alt_names - why? why not?
-  attr_accessor :key, :name, :alt_names,
-                :code,    ## code == abbreviation e.g. ARS etc.
-                :year, :year_end,   ## todo/fix: change year_end to end_year (like in season)!!!
-                :ground
-
-
-  alias_method :title, :name  ## add alias/compat - why? why not
-
-  def names
-    ## todo/check: add alt_names_auto too? - why? why not?
-    [@name] + @alt_names
-  end   ## all names
-
-
-  ## special import only attribs
-  attr_accessor :alt_names_auto    ## auto-generated alt names
-  attr_accessor :wikipedia   # wikipedia page name (for english (en))
-
-
-  def historic?()  @year_end ? true : false; end
-  alias_method  :past?, :historic?
-
-
-  attr_accessor :a, :b
-  def a?()  @a == nil; end  ## is a (1st) team / club (i)?            if a is NOT set
-  def b?()  @a != nil; end  ## is b (2nd/reserve/jr) team / club (ii) if a is     set
-
-  ## note: delegate/forward all geo attributes for team b for now (to team a) - keep - why? why not?
-  attr_writer  :city, :district, :country, :geos
-  def city()      @a == nil ?  @city     : @a.city;     end
-  def district()  @a == nil ?  @district : @a.district; end
-  def country()   @a == nil ?  @country  : @a.country;  end
-  def geos()      @a == nil ?  @geos     : @a.geos;     end
-
-
-  def wikipedia?()  @wikipedia; end
-  def wikipedia_url
-    if @wikipedia
-      ##  note: replace spaces with underscore (-)
-      ##  e.g. Club Brugge KV => Club_Brugge_KV
-      ##  todo/check/fix:
-      ##    check if "plain" dash (-) needs to get replaced with typographic dash??
-      "https://en.wikipedia.org/wiki/#{@wikipedia.gsub(' ','_')}"
-    else
-      nil
-    end
-  end
-
-
-  def initialize
-    @alt_names      = []
-    @alt_names_auto = []
-  end
-
-
-  ## helper methods for import only
-  ## check for duplicates
-  def duplicates?
-    names = [name] + alt_names + alt_names_auto
-    names = names.map { |name| normalize( sanitize(name) ) }
-
-    names.size != names.uniq.size
-  end
-
-  def duplicates
-    names = [name] + alt_names + alt_names_auto
-
-    ## calculate (count) frequency and select if greater than one
-    names.reduce( Hash.new ) do |h,name|
-       norm = normalize( sanitize(name) )
-       h[norm] ||= []
-       h[norm] << name; h
-    end.select { |norm,names| names.size > 1 }
-  end
-
-  def add_variants( name_or_names )
-    names = name_or_names.is_a?(Array) ? name_or_names : [name_or_names]
-    names.each do |name|
-      name = sanitize( name )
-      self.alt_names_auto += variants( name )
-    end
-  end
-
-
-   ###################################
-   # "global" helper - move to ___ ? why? why not?
-
-   ## note: allow placeholder years to e.g. (-___) or (-????)
-   ##    for marking missing (to be filled in) years
-   YEAR_REGEX = /\([0-9, ?_-]+?\)/    # note: non-greedy (minimum/first) match
-   def self.strip_year( name )
-     ## check for year(s) e.g. (1887-1911), (-2013),
-     ##                        (1946-2001, 2013-) etc.
-     name.gsub( YEAR_REGEX, '' ).strip
-   end
-
-   def self.has_year?( name ) name =~ YEAR_REGEX; end
-
-   LANG_REGEX = /\[[a-z]{1,2}\]/   ## note also allow [a] or [d] or [e] - why? why not?
-   def self.strip_lang( name )
-     name.gsub( LANG_REGEX, '' ).strip
-   end
-
-   def self.has_lang?( name ) name =~ LANG_REGEX; end
-
-  def self.sanitize( name )
-    ## check for year(s) e.g. (1887-1911), (-2013),
-    ##                        (1946-2001,2013-) etc.
-    name = strip_year( name )
-    ## check lang codes e.g. [en], [fr], etc.
-    name = strip_lang( name )
-    name
-  end
-
-
-  ## note: also add (),’,−  etc. e.g.
-  ##   Estudiantes (LP) => Estudiantes LP
-  ##   Saint Patrick’s Athletic FC => Saint Patricks Athletic FC
-  ##   Myllykosken Pallo −47 => Myllykosken Pallo 47
-
-  NORM_REGEX =  %r{
-                    [.'’º/()−-]
-                  }x   # note: in [] dash (-) if last doesn't need to get escaped
-  ## note: remove all dots (.), dash (-), ', º, /, etc.
-  #   .  U+002E (46) - FULL STOP
-  #   '  U+0027 (39) - APOSTROPHE
-  #   ’  U+2019 (8217) - RIGHT SINGLE QUOTATION MARK
-  #   º  U+00BA (186) - MASCULINE ORDINAL INDICATOR
-  #   /  U+002F (47) - SOLIDUS
-  #   (  U+0028 (40) - LEFT PARENTHESIS
-  #   )  U+0029 (41) - RIGHT PARENTHESIS
-  #   −  U+2212 (8722) - MINUS SIGN
-  #   -  U+002D (45) - HYPHEN-MINUS
-
-  ##         for norm(alizing) names
-  def self.strip_norm( name )
-    name.gsub( NORM_REGEX, '' )
-  end
-
-  def self.normalize( name )
-    # note: do NOT call sanitize here (keep normalize "atomic" for reuse)
-    name = strip_norm( name )
-    name = name.gsub( ' ', '' )  # note: also remove all spaces!!!
-
-    ## todo/fix: use our own downcase - why? why not?
-    name = downcase_i18n( name )     ## do NOT care about upper and lowercase for now
-    name
-  end
-
-
-  def self.strip_wiki( name )     # todo/check: rename to strip_wikipedia_en - why? why not?
-    ## note: strip disambiguationn qualifier from wikipedia page name if present
-    ##        note: only remove year and foot... for now
-    ## e.g. FC Wacker Innsbruck (2002) => FC Wacker Innsbruck
-    ##      Willem II (football club)  => Willem II
-    ##
-    ## e.g. do NOT strip others !! e.g.
-    ##   América Futebol Clube (MG)
-    ##  only add more "special" cases on demand (that, is) if we find more
-    name = name.gsub( /\([12][^\)]+?\)/, '' ).strip  ## starting with a digit 1 or 2 (assuming year)
-    name = name.gsub( /\(foot[^\)]+?\)/, '' ).strip  ## starting with foot (assuming football ...)
-    name
-  end
-
-
-private
-  ## private "shortcut" convenience helpers
-  def sanitize( name )    self.class.sanitize( name ); end
-  def normalize( name )   self.class.normalize( name ); end
-
-  def variants( name )  Variant.find( name ); end
-end # class Club
-
-
-
-############
-#  convenience
-#   Club and Team are for now alias
-#     in the future make
-#        Club > Team
-#        NationalTeam > Team  - why? why not?
-Team = Club
-
-
-end   # module Import
-end   # module SportDb