worldfootball 0.2.6 → 0.3.1

data/lib/worldfootball/page_schedule.rb

@@ -12,6 +12,15 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
   end
 
 
+  PLACEHOLDERS = [
+     'N.N.',
+     'Verlierer HF 1',
+     'Verlierer HF 2',
+     'Sieger HF 1',
+     'Sieger HF 2',
+    ]
+  def placeholder?( str )  PLACEHOLDERS.include?( str ); end
+
 
   def matches
     @matches ||= begin
@@ -20,10 +29,12 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
  # <table class="standard_tabelle" cellpadding="3" cellspacing="1">
 
  ## note: use > for "strict" sibling (child without any in-betweens)
- table = doc.css( 'div.data > table.standard_tabelle' ).first    ## get table
+ tables = doc.css( 'div.data > table.standard_tabelle' )    ## get table
  # puts table.class.name  #=> Nokogiri::XML::Element
  # puts table.text
 
+ assert( tables.size==1, "expected one table.standard_tabelle; got #{tables.size}" )
+ table = tables.first
  assert( table, 'no table.standard_tabelle found in schedule page!!')
 
  trs   = table.css( 'tr' )
@@ -77,8 +88,10 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
     ths =  tr.css( 'th' )
     tds =  tr.css( 'td' )
 
-   if tr.text.strip =~ /Spieltag/ ||
-      tr.text.strip =~ /[1-9]\.[ ]Runde|
+    tr_text = squish( tr.text )
+
+   if tr_text =~ /Spieltag/ ||
+      tr_text =~ /[1-9]\.[ ]Runde|
                            Qual\.[ ][1-9]\.[ ]Runde|  # see EL or CL Quali
                            Qualifikation|     # see CA Championship
                            Sechzehntelfinale|   # see EL
@@ -110,25 +123,24 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
      if debug?
        puts
        print '[%03d] ' % i
-       ## print squish( tr.text )
-       print "round >#{tr.text.strip}<"
+       print "round >#{tr_text}<"
        print "\n"
      end
 
-     last_round = tr.text.strip
+     last_round = tr_text
    elsif ths.count > 0 &&
          tds.count == 0
        ## check for round NOT yet configured!!!
-       puts "!! WARN: found unregistered round line >#{tr.text.strip}<"
-       log( "!! WARN: found unregistered round line >#{tr.text.strip}< in page #{title}" )
+       puts "!! WARN: found unregistered round line >#{tr_text}<"
+       log( "!! WARN: found unregistered round line >#{tr_text}< in page #{title}" )
 
-       last_round = tr.text.strip
+       last_round = tr_text
    else   ## assume table row (tr) is match line
 
      date_str  = squish( tds[0].text )
      time_str  = squish( tds[1].text )
 
-     date_str = last_date_str    if date_str.empty?
+     date_str = last_date_str    if date_str.empty? && last_date_str
 
      if debug?
        ## note: for debugging - print as we go along (parsing)
@@ -148,7 +160,7 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
      else
        team1_str    = squish( tds[2].text )
        team1_ref    = nil
-       puts "!! WARN: no team1_ref for >#{team1_str}< found"
+       puts "!! WARN: no team1_ref for >#{team1_str}< found"  unless placeholder?( team1_str )
      end
 
      if debug?
@@ -170,9 +182,10 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
      else
        team2_str    = squish( tds[4].text )
        team2_ref    = nil
-       puts "!! WARN: no team2_ref for >#{team2_str}< found"
+       puts "!! WARN: no team2_ref for >#{team2_str}< found"  unless placeholder?( team2_str )
      end
 
+
      if debug?
        ## note: for debugging - print as we go along (parsing)
        print "%-22s | " % team2_str
@@ -198,8 +211,12 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
      ##   <img src="https://s.hs-data.com/bilder/shared/live/2.png" /></a>
      ## </td>
      img = tds[6].css( 'img' )[0]
+
+
+
      if img && img[:src].index( '/live/')
-       puts "!! WARN: live match badge, resetting score from #{score_str} to -:-"
+       ## puts "!! WARN: live match badge, resetting score from #{score_str} to -:-"
+       print " LIVE BADGE "
        score_str = '-:-'  # note: -:- gets replaced to ---
      end
 
@@ -217,8 +234,12 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
      ## special case for '00.00.0000'
      ##   CANNOT parse
      ##   use empty date - why? why not?
+     ##   if start with 00.00. e.g. 00.00.1939
+
 
-     date     = if date_str == '00.00.0000'
+     date     = if date_str == '00.00.0000' ||
+                   date_str.start_with?( '00.00.' ) ||
+                   date_str.empty?
                   nil
                 else
                   Date.strptime( date_str, '%d.%m.%Y' )
@@ -237,7 +258,9 @@ class Schedule < Page  ## note: use nested class for now - why? why not?
                report_ref: score_ref
              }
 
-     last_date_str = date_str
+     ## note - only update last date if date present
+     ##             might be empty (not available) in the beginning
+     last_date_str = date_str      if !date_str.empty?
    end
   end # each tr (table row)
 

data/lib/worldfootball/version.rb

@@ -1,8 +1,8 @@
 
 module Worldfootball
   MAJOR = 0    ## todo: namespace inside version or something - why? why not??
-  MINOR = 2
-  PATCH = 6
+  MINOR = 3
+  PATCH = 1
   VERSION = [MAJOR,MINOR,PATCH].join('.')
 
   def self.version

data/lib/worldfootball.rb

@@ -6,6 +6,9 @@ require 'webget'              ## incl. webget, webcache, webclient, etc.
 require 'nokogiri'
 
 
+require 'sportdb/writers'
+
+
 ###
 # our own code
 require_relative 'worldfootball/version'
@@ -38,6 +41,14 @@ module Worldfootball
    def self.debug=(value) @debug = value; end
    def self.debug?()      @debug ||= false; end  ## note: default is FALSE
 
+   def self.log( msg )  ### append to log
+      File.open( './logs.txt', 'a:utf-8' ) do |f|
+        f.write( msg )
+        f.write( "\n" )
+      end
+   end
+
+
 
 
 class Configuration
@@ -47,10 +58,16 @@ class Configuration
      def out_dir()       @out_dir || './o'; end
      def out_dir=(value) @out_dir = value; end
   end
+  class Generate
+     def out_dir()       @out_dir || './o'; end
+     def out_dir=(value) @out_dir = value; end
+ end
 
- def convert()  @convert ||= Convert.new; end
+ def convert()  @convert  ||= Convert.new; end
+ def generate() @generate ||= Generate.new; end
 end # class Configuration
 
+
 ## lets you use
 ##   Worldfootball.configure do |config|
 ##      config.convert.out_dir = './o'
@@ -62,6 +79,60 @@ end   # module Worldfootball
 
 
 
+###
+# todo - move generate to generate file!!!
+module Worldfootball
+def self.generate( league:, season:,
+                   overwrite: true )
+   season = Season( season )  ## cast (ensure) season class (NOT string, integer, etc.)
+
+   league = find_league!( league )
+   pages  = league.pages!( season: season )
+
+
+   out_path = if season >= Season( '2000' )
+                "#{config.generate.out_dir}/#{season.to_path}/#{league.key}.txt"
+              else
+                decade = season.start_year - (season.start_year%10)
+                ## use archive-style before 2000!!!
+                "#{config.generate.out_dir}/archive/#{decade}s/#{season.to_path}/#{league.key}.txt"
+              end
+
+   ## check if output exists already
+   if !overwrite && File.exist?( out_path )
+     ## skip generation
+     puts "  OK #{league.key} #{season}   (do NOT overwrite)"
+     return
+   end
+
+
+   ## get matches
+   path = "#{config.convert.out_dir}/#{season.to_path}/#{league.key}.csv"
+   puts "  ---> reading matches in #{path} ..."
+   matches = SportDb::CsvMatchParser.read( path )
+   puts "     #{matches.size} matches"
+
+   ## build
+   txt = SportDb::TxtMatchWriter.build( matches )
+   puts txt
+
+
+   buf = String.new
+   ## note - use league key for league name for now!!
+   buf << "= #{league.key.upcase.gsub('.', ' ')} #{season.key}\n\n"
+   buf << txt
+
+   puts "   writing to >#{out_path}<..."
+   write_text( out_path, buf )
+
+   ## add to tmp too for debugging
+   out_path2 = "#{config.generate.out_dir}/tmp/#{league.key}/#{season.to_path}.txt"
+   puts "   writing to >#{out_path2}<..."
+   write_text( out_path2, buf )
+end
+end   # module Worldfootball
+
+
 
 
 puts Worldfootball.banner    ## say hello

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: worldfootball
 version: !ruby/object:Gem::Version
-  version: 0.2.6
+  version: 0.3.1
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2024-10-05 00:00:00.000000000 Z
+date: 2024-12-31 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: football-timezones
@@ -66,6 +66,20 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: '0'
+- !ruby/object:Gem::Dependency
+  name: sportdb-writers
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: '0'
 - !ruby/object:Gem::Dependency
   name: rdoc
   requirement: !ruby/object:Gem::Requirement
@@ -92,19 +106,24 @@ dependencies:
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '4.1'
+        version: '4.2'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '4.1'
+        version: '4.2'
 description: worldfootball - get world football (leagues, cups & more) match data
   via the worldfootball.net/weltfussball.de pages
 email: gerald.bauer@gmail.com
 executables:
 - wfb
+- wfbconf
+- wfbconv
+- wfbdump
+- wfbgen
+- wfbup
 extensions: []
 extra_rdoc_files:
 - CHANGELOG.md
@@ -116,12 +135,17 @@ files:
 - README.md
 - Rakefile
 - bin/wfb
-- config/leagues_africa.csv
-- config/leagues_america.csv
-- config/leagues_asia.csv
-- config/leagues_europe.csv
-- config/leagues_middle_east.csv
-- config/leagues_pacific.csv
+- bin/wfbconf
+- bin/wfbconv
+- bin/wfbdump
+- bin/wfbgen
+- bin/wfbup
+- config/leagues/africa.csv
+- config/leagues/america.csv
+- config/leagues/asia.csv
+- config/leagues/europe.csv
+- config/leagues/middle_east.csv
+- config/leagues/pacific.csv
 - config/rounds.csv
 - config/stages.csv
 - lib/worldfootball.rb
@@ -161,7 +185,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubygems_version: 3.4.10
+rubygems_version: 3.5.22
 signing_key:
 specification_version: 4
 summary: worldfootball - get world football (leagues, cups & more) match data via