football-sources 0.1.0 → 0.2.0

data/lib/football-sources/worldfootball/convert_reports.rb

@@ -1,107 +0,0 @@
-module Worldfootball
-
-
-def self.convert_reports( league:, season: )
-  season = Season( season )  ## cast (ensure) season class (NOT string, integer, etc.)
-
-  league = find_league( league )
-
-   ## note: use only first part from key for lookup
-   ##    e.g. at.1  => at
-   ##         eng.1 => eng
-   ##     and so on
-   mods = MODS[ league.key.split('.')[0] ] || {}
-
-
-
-  pages = league.pages( season: season )
-
-  recs = []
-
-  ## if single (simple) page setup - wrap in array
-  pages = pages.is_a?(Array) ? pages : [pages]
-  pages.each do |page_meta|  # note: use page_info for now (or page_rec or page_meta or such)
-
-    page = Page::Schedule.from_cache( page_meta[:slug] )
-    print "  page title=>#{page.title}<..."
-    print "\n"
-
-    matches = page.matches
-
-    puts "matches - #{matches.size} rows:"
-    pp matches[0]
-
-    puts "#{page.generated_in_days_ago}  - #{page.generated}"
-
-
-    matches.each_with_index do |match,i|
-
-      report_ref = match[:report_ref]
-      if report_ref.nil?
-        puts "!! WARN: no match report ref found for match:"
-        pp match
-        next
-      end
-
-      puts "reading #{i+1}/#{matches.size} - #{report_ref}..."
-      report = Page::Report.from_cache( report_ref )
-
-      puts
-      puts report.title
-      puts report.generated
-
-      rows = report.goals
-      puts "goals - #{rows.size} records"
-      ## pp rows
-
-
-      if rows.size > 0
-        ## add goals
-        date = Date.strptime( match[:date], '%Y-%m-%d')
-
-        team1 = match[:team1]
-        team2 = match[:team2]
-
-        ## clean team name (e.g. remove (old))
-        ##   and asciify (e.g. ’ to ' )
-        team1 = norm_team( team1 )
-        team2 = norm_team( team2 )
-
-        team1 = mods[ team1 ]   if mods[ team1 ]
-        team2 = mods[ team2 ]   if mods[ team2 ]
-
-        match_id = "#{team1} - #{team2} | #{date.strftime('%b %-d %Y')}"
-
-
-        rows.each do |row|
-          extra = if row[:owngoal]
-                   '(og)'  ## or use OG or O.G.- why? why not?
-                  elsif row[:penalty]
-                   '(pen)' ## or use P or PEN - why? why not?
-                  else
-                    ''
-                  end
-
-          rec = [match_id,
-                row[:score],
-                "#{row[:minute]}'",
-                extra,
-                row[:player],
-                row[:notes]]
-          recs << rec
-        end
-      end
-     end #  each match
-    end # each page
-
-  ## pp recs
-
-  out_path = "#{config.convert.out_dir}/#{season.path}/#{league.key}~goals.csv"
-
-  headers  = ['Match', 'Score', 'Minute', 'Extra', 'Player', 'Notes']
-
-  puts "write #{out_path}..."
-  Cache::CsvMatchWriter.write( out_path, recs, headers: headers )
-end
-end  # module Worldfootballl
-

data/lib/football-sources/worldfootball/jobs.rb

@@ -1,76 +0,0 @@
-
-
-module Worldfootball
-################################
-# add more helpers
-#  move upstream for (re)use - why? why not?
-
-## todo/check: what to do: if league is both included and excluded?
-##   include forces include? or exclude has the last word? - why? why not?
-##  Excludes match before includes,
-##   meaning that something that has been excluded cannot be included again
-
-## todo - find "proper/classic" timezone ("winter time")
-
-##  Brasilia - Distrito Federal, Brasil  (GMT-3)  -- summer time?
-##  Ciudad de México, CDMX, México       (GMT-5)  -- summer time?
-##  Londres, Reino Unido (GMT+1)
-##   Madrid -- ?
-##   Lisboa -- ?
-##   Moskow -- ?
-##
-## todo/check - quick fix timezone offsets for leagues for now
-##   - find something better - why? why not?
-## note: assume time is in GMT+1
-OFFSETS = {
-  'eng.1' => -1,
-  'eng.2' => -1,
-  'eng.3' => -1,
-  'eng.4' => -1,
-  'eng.5' => -1,
-
-  'es.1'  => -1,
-  'es.2'  => -1,
-
-  'pt.1'  => -1,
-  'pt.2'  => -1,
-
-  'br.1'  => -5,
-  'mx.1'  => -7,
-}
-
-
-
-class Job     ## todo/check: use a module (NOT a class) - why? why not?
-def self.download( datasets )
-  datasets.each_with_index do |dataset,i|
-    league  = dataset[0]
-    seasons = dataset[1]
-
-    puts "downloading [#{i+1}/#{datasets.size}] #{league}..."
-    seasons.each_with_index do |season,j|
-      puts "  season [#{j+1}/#{season.size}] #{league} #{season}..."
-      Worldfootball.schedule( league: league,
-                              season: season )
-    end
-  end
-end
-
-def self.convert( datasets )
-  datasets.each_with_index do |dataset,i|
-    league  = dataset[0]
-    seasons = dataset[1]
-
-    puts "converting [#{i+1}/#{datasets.size}] #{league}..."
-    seasons.each_with_index do |season,j|
-      puts "  season [#{j+1}/#{season.size}] #{league} #{season}..."
-      Worldfootball.convert( league: league,
-                             season: season,
-                             offset: OFFSETS[ league ] )
-    end
-  end
-end
-end  # class Job
-
-end # module Worldfootball
-

data/lib/football-sources/worldfootball/mods.rb

@@ -1,72 +0,0 @@
-#### todo/check: move MODS and SCORE_ERRORS out-of-lib
-##               and into config or such - why? why not?
-
-
-module Worldfootball
-
-
-######
-# "global" helpers
-def self.norm_team( team )
-   ## clean team name and asciify (e.g. ’->' )
-   team = team.sub( '(old)', '' ).strip
-   team = team.gsub( '’', "'" )     ## e.g. Hawke’s Bay United FC
-   team
-end
-
-
-
-MODS = {
- 'at' => {
-    ## AT 1
-    'SC Magna Wiener Neustadt' => 'SC Wiener Neustadt', # in 2010/11
-    'KSV Superfund'            => 'Kapfenberger SV',    # in 2010/11
-    'Kapfenberger SV 1919'     => 'Kapfenberger SV',    # in 2011/12
-    'FC Trenkwalder Admira'    => 'FC Admira Wacker',    # in 2011/12
-    ## AT 2
-    'Austria Wien (A)'         => 'Young Violets',  # in 2019/20
-    'FC Wacker Innsbruck (A)'  => 'FC Wacker Innsbruck II',   # in 2018/19
-    ## AT CUP
-    'Rapid Wien (A)'           => 'Rapid Wien II',  # in 2011/12
-    'Sturm Graz (A)'           => 'Sturm Graz II',
-    'Kapfenberger SV 1919 (A)' => 'Kapfenberger SV II',
-    'SV Grödig (A)'            => 'SV Grödig II',
-    'RB Salzburg (A)'          => 'RB Salzburg II',
-    'SR WGFM Donaufeld'        => 'SR Donaufeld Wien',
-    'FC Trenkwalder Admira (A)' => 'FC Admira Wacker II',
-    ## AT 3.O (Regionalliga Ost)
-    'FC Admira Wacker (A)'     => 'FC Admira Wacker II',  # in 2020/21
-  },
- 'nz' => {
-    ## NZ 1
-   'Wellington Phoenix (R)' => 'Wellington Phoenix Reserves',
-  },
-}
-
-
-
-## fix/patch known score format errors in at/de cups
-##   new convention
-##   for a fix require league, date, and team1 & team2 for now!!!!
-##    - do NOT use some "generic" fix / patch!!!!
-##
-## old de/at patches/fixes:
-##  '0-1 (0-0, 0-0, 0-0) n.V.' => '0-1 (0-0, 0-0) n.V.',       # too long
-##  '2-1 (1-1, 1-1, 1-0) n.V.' => '2-1 (1-1, 1-1) n.V.',
-##  '4-2 (0-0, 0-0) i.E.'      => '4-2 (0-0, 0-0, 0-0) i.E.',  # too short
-
-
-SCORE_ERRORS = {
-  'ro.1' => {
-    ## 2013/14
-    '2013-07-29' => [ 'FC Brașov', 'Săgeata Năvodari', ['1-1 (0-0, 0-1)', '1-1 (0-0)']],
-  },
-  'gr.1' => {
-    ## 2010/11
-    '2010-11-24' => [ 'Ergotelis',    'Olympiakos Piräus', ['0-2 (0-0, 0-0, 0-0)', '0-2 (0-0)']],
-    '2010-11-28' => [ 'Panserraikos', 'Aris Saloniki',     ['1-0 (1-0, 0-0, 0-0)', '1-0 (1-0)']],
-  }
-}
-
-
-end # module Worldfootball

data/lib/football-sources/worldfootball/vacuum.rb

@@ -1,66 +0,0 @@
-module Worldfootball
-
-
-MAX_HEADERS = [
-'Stage',
-'Round',
-'Date',
-'Time',
-'Team 1',
-'FT',
-'HT',
-'Team 2',
-'ET',
-'P',
-'Comments']    ## e.g. awarded, cancelled/canceled, etc.
-
-MIN_HEADERS = [   ## always keep even if all empty
-'Date',
-'Team 1',
-'FT',
-'Team 2'
-]
-
-def self.vacuum( rows, headers: MAX_HEADERS, fixed_headers: MIN_HEADERS )
-  ## check for unused columns and strip/remove
-  counter = Array.new( MAX_HEADERS.size, 0 )
-  rows.each do |row|
-     row.each_with_index do |col, idx|
-       counter[idx] += 1  unless col.nil? || col.empty?
-     end
-  end
-
-  pp counter
-
-  ## check empty columns
-  headers       = []
-  indices       = []
-  empty_headers = []
-  empty_indices = []
-
-  counter.each_with_index do |num, idx|
-     header = MAX_HEADERS[ idx ]
-     if num > 0 || (num == 0 && fixed_headers.include?( header ))
-       headers << header
-       indices << idx
-     else
-       empty_headers << header
-       empty_indices << idx
-     end
-  end
-
-  if empty_indices.size > 0
-    rows = rows.map do |row|
-             row_vacuumed = []
-             row.each_with_index do |col, idx|
-               ## todo/fix: use values or such??
-               row_vacuumed << col   unless empty_indices.include?( idx )
-             end
-             row_vacuumed
-         end
-    end
-
-  [rows, headers]
-end
-
-end # module Worldfootball

data/lib/football-sources/worldfootball.rb

@@ -1,19 +0,0 @@
-###########################
-#  note: split code in two parts
-#    metal  - "bare" basics - no ref to sportdb
-#    and rest / convert  with sportdb references / goodies
-
-
-## our own code
-require_relative 'worldfootball/config'
-require_relative 'worldfootball/mods'
-require_relative 'worldfootball/vacuum'
-require_relative 'worldfootball/build'
-require_relative 'worldfootball/convert'
-require_relative 'worldfootball/convert_reports'
-
-
-require_relative 'worldfootball/jobs'
-
-
-

data/test/helper.rb

@@ -1,8 +0,0 @@
-$LOAD_PATH.unshift( '../webget/lib')
-
-require 'minitest/autorun'
-
-
-## our own code
-require 'football/sources'     ## or require 'football-sources'
-

data/test/test_version.rb

@@ -1,16 +0,0 @@
-###
-#  to run use
-#     ruby -I ./lib -I ./test test/test_version.rb
-
-
-require 'helper'
-
-class TestVersion < MiniTest::Test
-
-  def test_version
-    pp FootballSources::VERSION
-    pp FootballSources.banner
-    pp FootballSources.root
-  end
-
-end # class TestVersion