rsssf 0.1.0 → 0.2.0

data/lib/rsssf/reports/page.rb

@@ -1,25 +1,31 @@
-# encoding: utf-8
 
 
 module Rsssf
 
 class PageReport
 
+
+def self.build( files, title:  )   
+  stats = []
+  files.each do |file|
+    page = Page.read_txt( file )
+    stats << page.build_stat
+  end
+
+  new( stats, title: title )
+end
+
+
 attr_reader :title
 
-def initialize( stats, opts )
+def initialize( stats, title: )
   @stats = stats
-  @opts  = opts
-    
-  @title = opts[:title] || 'Your Title Here' 
+  @title = title
 end
 
-def save( path )
-  ### save report as README.md in repo
-  File.open( path, 'w' ) do |f|
-    f.write build_summary
-  end
-end
+### save report as README.md in repo
+def save( path ) write_text( path, build_summary ); end
+
 
 def build_summary
 
@@ -33,19 +39,26 @@ def build_summary
 
 football.db RSSSF Archive Data Summary for #{title}
 
-_Last Update: #{Time.now}_
-
 EOS
 
+## no longer add last update
+##  _Last Update: #{Time.now}_
+
+
   txt = ''
   txt << header
 
-  txt << "| Season | File   | Authors  | Last Updated | Lines (Chars) | Sections |\n"
-  txt << "| :----- | :----- | :------- | :----------- | ------------: | :------- |\n"
+  txt << "| File   | Authors  | Last Updated | Lines (Chars) | Sections |\n"
+  txt << "| :----- | :------- | :----------- | ------------: | :------- |\n"
+
+## note - removed season (no longer tracked here)
 
   stats.each do |stat|
-    txt << "| #{stat.season} "
-    txt << "| [#{stat.basename}.txt](#{stat.basename}.txt) "
+    ## get basename from source url
+    url_path  = URI.parse( stat.source ).path
+    basename  = File.basename( url_path, File.extname( url_path ) )  ## e.g. duit92.txt or duit92.html => duit92
+  
+    txt << "| [#{basename}.txt](#{basename}.txt) "
     txt << "| #{stat.authors} "
     txt << "| #{stat.last_updated} "
     txt << "| #{stat.line_count} (#{stat.char_count}) "
@@ -60,5 +73,3 @@ end  # method build_summary
 end  ## class PageReport
 end  ## module Rsssf
 
-## add (shortcut) alias
-RsssfPageReport = Rsssf::PageReport

data/lib/rsssf/reports/schedule.rb

@@ -1,30 +1,72 @@
-# encoding: utf-8
+
 
 module Rsssf
-     
+
+  
+ScheduleStat = Struct.new(
+    :path,     ## path to .txt file
+    :errors   ## array or nil
+)
+
+
+
 class ScheduleReport
 
-attr_reader :title
+  include Utils       ## e.g. year_from_file, etc.
+
+##
+##  quick hack?  pass along (optional) patch
+
+def self.build( files, title:, 
+                       patch: nil )
+  linter = Parser::Linter.new
 
-def initialize( stats, opts )
-  @stats = stats
-  @opts  = opts
+  stats = []
+  files.each_with_index do |file,i|
+
+    puts "==> [#{i+1}/#{files.size}] reading >#{file}<..."  
+
+    txt = read_text( file )
+
+    if patch && patch.respond_to?(:on_parse)
+      season_dir = File.basename(File.dirname(file)) 
+      season     = Season( season_dir )
+      basename   = File.basename(file, File.extname(file)) 
+      puts "  [debug] before  patch.on_parse #{basename}, #{season}"
+      txt = patch.on_parse( txt, basename, season )
+    end
+  
+    linter.parse( txt, parse: true,
+                       path:  file  )   ## todo/fix - change path to file/filename - why? why not?
     
-  @title = opts[:title] || 'Your Title Here' 
+    stat = ScheduleStat.new
+    stat.path   = file
+    stat.errors = linter.errors
+     
+    stats << stat
+  end
+
+  new( stats, title: title )
 end
 
-def save( path )
-  ### save report as README.md in repo
-  File.open( path, 'w' ) do |f|
-    f.write build_summary
-  end
+
+attr_reader :title
+
+def initialize( stats,  title: )
+  @stats = stats 
+  @title = title
 end
 
+### save report as README.md in repo
+def save( path ) write_text( path, build_summary ); end
+
+
 def build_summary
-  ## sort start by season (latest first) than by name (e.g. 1-bundesliga, cup, etc.)
+  ## sort start 1) by season (latest first) than 
+  ##            2) by name (e.g. 1-bundesliga, cup, etc.)
   stats = @stats.sort do |l,r|
-    v =  r.season   <=> l.season
-    v =  l.filename <=> r.filename  if v == 0  ## same season
+    v =  File.basename(File.dirname(r.path)) <=> File.basename(File.dirname(l.path))
+    v =  File.basename(l.path) <=> File.basename(r.path)    if v == 0  ## same season
     v
   end
 
@@ -35,11 +77,14 @@ def build_summary
 football.db RSSSF (Rec.Sport.Soccer Statistics Foundation) Archive Data for
 #{title}
 
-_Last Update: #{Time.now}_
-
 EOS
 
+## no longer add last update
+## _Last Update: #{Time.now}_
+##
 
+
+=begin
   footer =<<EOS
 
 ## Questions? Comments?
@@ -48,30 +93,71 @@ Send them along to the
 [Open Sports & Friends Forum](http://groups.google.com/group/opensport).
 Thanks!
 EOS
+=end
+
 
+  errors = []
 
-  txt = ''
+
+  txt = String.new
   txt << header
   
-  txt << "| Season | League, Cup | Rounds |\n"
+  txt << "| Season | League, Cup | Errors |\n"
   txt << "| :----- | :---------- | -----: |\n"
 
+  
+  stats.each_with_index do |stat,i|
+ 
+      path = stat.path
+      season_dir = File.basename(File.dirname( path ))
+      filename   = File.basename( path ) ## incl. extension !!
+
+      season = Season( season_dir )
+      ## note - use archive_dir_for_season for archive path
+      
+
+      txt << "| #{season_dir} "
+      txt << "| [#{filename}](#{archive_dir_for_season(season)}/#{filename}) "
+    
+      txt <<   if stat.errors.size > 0
+                 "|  **!! #{stat.errors.size}**  "
+               else
+                 "|  OK  "
+               end
+      txt << "|\n"
+  
+      errors += stat.errors  if stat.errors.size > 0 
+  end
+
+   if errors.size > 0
+     txt << "\n\n"
+     txt << "#{errors.size} errors in #{stats.size} datafile(s)\n\n"
+
+     txt << "```\n"
+     errors.each do |path, msg, line|
+        season_dir = File.basename(File.dirname( path ))
+        filename   = File.basename( path ) ## incl. extension !!
+  
+        txt <<"#{season_dir}/#{filename} -- #{msg}\n"
+        txt << "     in line >#{line}<\n"    unless line.empty?
+     end
+     txt << "```\n"
+   end
+
+=begin
   stats.each do |stat|
     txt << "| #{stat.season} "
     txt << "| [#{stat.filename}](#{stat.path}/#{stat.filename}) "
     txt << "| #{stat.rounds} "
     txt << "|\n"
   end
+=end
 
-  txt << "\n\n" 
-
-  txt << footer
+ 
+  ## txt << footer
   txt
 end  # method build_summary
 
 end  ## class ScheduleReport
 end  ## module Rsssf
 
-## add (shortcut) alias
-RsssfScheduleReport = Rsssf::ScheduleReport
-

data/lib/rsssf/schedule.rb

@@ -1,31 +1,21 @@
-# encoding: utf-8
 
 module Rsssf
 
 class Schedule
 
-def self.from_string( txt )
-  self.new( txt )
-end
 
-attr_accessor :rounds     # track no of rounds
+#  attr_accessor :rounds     # track no of rounds - why? why not?
   
 def initialize( txt )
   @txt = txt
   
-  @rounds = nil   # undefined
+  ## @rounds = nil   # undefined
 end
 
 
-def save( path )
-  File.open( path, 'w' ) do |f|
-    f.write @txt
-  end    
+def save( path, header: )
+  write_text( path, header + @txt )
 end
 
 end  ## class Schedule
 end  ## module Rsssf
-
-## add (shortcut) alias
-RsssfSchedule = Rsssf::Schedule
-

data/lib/rsssf/utils.rb

@@ -1,8 +1,10 @@
-# encoding: utf-8
 
 module Rsssf
 module Utils
 
+
+## move to Page - why? why not?
+
 def year_from_file( path )
   extname  = File.extname( path )
   basename = File.basename( path, extname )  ## e.g. duit92.txt or duit92.html => duit92
@@ -32,44 +34,23 @@ def year_from_name( name )
 end  # method year_from_name
 
 
-def year_to_season( year )
-
-  ## todo: require four digit years? why? why not??
-
-  ## e.g. 64   => 1963-64
-  ##      2011 => 2010-11   etc.
-
-  if year <= 16  ## assume 20xx for now from 00..16
-    year += 2000
-  elsif year <= 99
-    year += 1900
-  else
-    # use as is; assume four digit year
-  end  
 
-  year_prev = year-1
+def archive_dir_for_season( season )
+  season = Season( season )
 
-  "%4d-%02d" % [year_prev, year%100]   ## e.g. return 1974-75
-end
-
-
-def archive_dir_for_year( year )
-  season = year_to_season( year )
-  if year <= 2010   # e.g. season 2009-10
+  if season < Season('2010')   # e.g. season 2009-10 
     ## use archive folder (w/ 1980s etc)
     ## get decade folder
-    decade  = year-1
+    decade  = season.start_year     ## 1999/2000 2000
     decade -= decade % 10   ## turn 1987 into 1980 etc
-    "archive/#{decade}s/#{season}"
+    "archive/#{decade}s/#{season.to_path}"
   else
-    season 
+    season.to_path
   end
 end
 
 
+
 end  # module Utils
 end  # module Rsssf
 
-## add (shortcut) alias
-RsssfUtils = Rsssf::Utils
-

data/lib/rsssf/version.rb

@@ -1,10 +1,8 @@
-# encoding: utf-8
-
 
 module Rsssf
 
   MAJOR = 0
-  MINOR = 1
+  MINOR = 2
   PATCH = 0
   VERSION = [MAJOR,MINOR,PATCH].join('.')
 
@@ -13,11 +11,11 @@ module Rsssf
   end
 
   def self.banner
-    "rsssf/#{VERSION} on Ruby #{RUBY_VERSION} (#{RUBY_RELEASE_DATE}) [#{RUBY_PLATFORM}]"
+    "rsssf/#{VERSION} on Ruby #{RUBY_VERSION} (#{RUBY_RELEASE_DATE}) [#{RUBY_PLATFORM}] in (#{root})"
   end
 
   def self.root
-    "#{File.expand_path( File.dirname(File.dirname(File.dirname(__FILE__))) )}"
+     File.expand_path( File.dirname(File.dirname(File.dirname(__FILE__))) )
   end
 
 end # module Rsssf

data/lib/rsssf.rb

@@ -1,34 +1,57 @@
-# encoding: utf-8
 
-## stdlibs
-require 'pp'
-require 'yaml'
-require 'uri'
+## 3rd party (our own)
+require 'season/formats'   ## add season support
+require 'webget'           ## incl. webget, webcache, webclient, etc.
+
+require 'cocos'
+
+
+## (old) 3rd party libs
+##  require 'textutils'      ## used for File.read_utf8 etc.
+## require 'fetcher'        ## used for Fetcher::Worker.new.fetch etc.
+
+
+#######
+##   add RsssfParser too
+require 'rsssf/parser'    ## from rsssf-parser gem
 
 
-## 3rd party libs
-require 'textutils'      ## used for File.read_utf8 etc.
-require 'fetcher'        ## used for Fetcher::Worker.new.fetch etc.
 
 
 ## our own code
-require 'rsssf/version'    # note: let version always go first
+require_relative 'rsssf/version'    # note: let version always go first
+
+require_relative 'rsssf/utils'      # include Utils - goes first
+
+require_relative 'rsssf/download'
+
+require_relative 'rsssf/convert'
+require_relative 'rsssf/page'
+require_relative 'rsssf/schedule'
+
+require_relative 'rsssf/reports/schedule'
+require_relative 'rsssf/reports/page'
+
+require_relative 'rsssf/repo'
+
 
-require 'rsssf/utils'      # include Utils - goes first
-require 'rsssf/html2txt'   # include Filters - goes first
 
-require 'rsssf/fetch'
-require 'rsssf/page'
-require 'rsssf/schedule'
-require 'rsssf/patch'
 
-require 'rsssf/reports/schedule'
-require 'rsssf/reports/page'
+#############
+## add (shortcut) alias(es)
+RsssfPage           = Rsssf::Page
+RsssfPageConverter  = Rsssf::PageConverter
+RsssfPageStat       = Rsssf::PageStat
+RsssfPageReport     = Rsssf::PageReport
 
-require 'rsssf/repo'
+RsssfSchedule       = Rsssf::Schedule
+RsssfScheduleStat   = Rsssf::ScheduleStat
+RsssfScheduleReport = Rsssf::ScheduleReport
 
+RsssfRepo           = Rsssf::Repo
+RsssfUtils          = Rsssf::Utils
 
 
 
 ## say hello
-puts Rsssf.banner   if defined?($RUBYLIBS_DEBUG) && $RUBYLIBS_DEBUG
+puts Rsssf.banner   ##  if defined?($RUBYLIBS_DEBUG) && $RUBYLIBS_DEBUG

metadata

@@ -1,17 +1,17 @@
 --- !ruby/object:Gem::Specification
 name: rsssf
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.2.0
 platform: ruby
 authors:
 - Gerald Bauer
-autorequire: 
+autorequire:
 bindir: bin
 cert_chain: []
-date: 2015-09-15 00:00:00.000000000 Z
+date: 2024-07-22 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
-  name: logutils
+  name: cocos
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -25,7 +25,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: textutils
+  name: season-formats
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -39,7 +39,7 @@ dependencies:
       - !ruby/object:Gem::Version
         version: '0'
 - !ruby/object:Gem::Dependency
-  name: fetcher
+  name: rsssf-parser
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
@@ -56,63 +56,65 @@ dependencies:
   name: rdoc
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '4.0'
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: '7'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - "~>"
+    - - ">="
       - !ruby/object:Gem::Version
         version: '4.0'
+    - - "<"
+      - !ruby/object:Gem::Version
+        version: '7'
 - !ruby/object:Gem::Dependency
   name: hoe
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.13'
+        version: '4.1'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - "~>"
       - !ruby/object:Gem::Version
-        version: '3.13'
+        version: '4.1'
 description: rsssf - tools 'n' scripts for RSSSF (Rec.Sport.Soccer Statistics Foundation)
   archive data
-email: opensport@googlegroups.com
+email: gerald.bauer@gmail.com
 executables: []
 extensions: []
 extra_rdoc_files:
-- HISTORY.md
+- CHANGELOG.md
 - Manifest.txt
 - README.md
 files:
-- ".gemtest"
-- HISTORY.md
+- CHANGELOG.md
 - Manifest.txt
 - README.md
 - Rakefile
 - lib/rsssf.rb
-- lib/rsssf/fetch.rb
-- lib/rsssf/html2txt.rb
+- lib/rsssf/convert.rb
+- lib/rsssf/download.rb
 - lib/rsssf/page.rb
-- lib/rsssf/patch.rb
 - lib/rsssf/repo.rb
 - lib/rsssf/reports/page.rb
 - lib/rsssf/reports/schedule.rb
 - lib/rsssf/schedule.rb
 - lib/rsssf/utils.rb
 - lib/rsssf/version.rb
-- test/helper.rb
-- test/test_utils.rb
-homepage: https://github.com/sportdb/rsssf
+homepage: https://github.com/sportdb/sport.db.sources
 licenses:
 - Public Domain
 metadata: {}
-post_install_message: 
+post_install_message:
 rdoc_options:
 - "--main"
 - README.md
@@ -122,16 +124,15 @@ required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
-      version: 1.9.2
+      version: 2.2.2
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
-rubyforge_project: 
-rubygems_version: 2.2.3
-signing_key: 
+rubygems_version: 3.4.10
+signing_key:
 specification_version: 4
 summary: rsssf - tools 'n' scripts for RSSSF (Rec.Sport.Soccer Statistics Foundation)
   archive data

data/lib/rsssf/fetch.rb

@@ -1,80 +0,0 @@
-# encoding: utf-8
-
-module Rsssf
-
-class PageFetcher
-
-  include Filters   # e.g. html2text, sanitize etc.
-
-
-def initialize
-  @worker = Fetcher::Worker.new
-end
-  
-def fetch( src_url )
-
-  ## note: assume plain 7-bit ascii for now
-  ##  -- assume rsssf uses ISO_8859_15 (updated version of ISO_8859_1) -- does NOT use utf-8 character encoding!!!
-  html = @worker.read( src_url )  
-
-  ### todo/fix: first check if html is all ascii-7bit e.g.
-  ## includes only chars from 64 to 127!!!
-
-  ## normalize newlines
-  ##   remove \r (form feed) used by Windows; just use \n (new line)
-  html = html.gsub( "\r", '' )
-
-  ## note:
-  ##   assume (default) to ISO 3166-15 (an updated version of ISO 3166-1) for now
-  ##
-  ##  other possible alternatives - try:
-  ##  - Windows CP 1562  or
-  ##  - ISO 3166-2  (for eastern european languages )
-  ##
-  ## note: german umlaut use the same code (int)
-  ##    in ISO 3166-1/15 and 2 and Windows CP1562  (other chars ARE different!!!)
-
-  html = html.force_encoding( Encoding::ISO_8859_15 )
-  html = html.encode( Encoding::UTF_8 )    # try conversion to utf-8
-
-  ## check for html entities
-  html = html.gsub( "ä", 'ä' )
-  html = html.gsub( "ö", 'ö' )
-  html = html.gsub( "ü", 'ü' )
-  html = html.gsub( "Ä", 'Ä' )
-  html = html.gsub( "Ö", 'Ö' )
-  html = html.gsub( "Ü", 'Ü' )
-  html = html.gsub( "ß", 'ß' )
-
-  html = html.gsub( "&oulm;", 'ö' )    ## support typo in entity (ö)
-  html = html.gsub( "&slig;", "ß" )    ## support typo in entity (ß)
-  
-  html = html.gsub( "É", 'É' )
-  html = html.gsub( "ø", 'ø' )
-  
-  ## check for more entities
-  html = html.gsub( /&[^;]+;/) do |match|
-    puts "*** found unencoded html entity #{match}"
-    match   ## pass through as is (1:1)
-  end
-  ## todo/fix: add more entities
-
-
-  txt   = html_to_txt( html )
-
-  header = <<EOS
-<!--
-   source: #{src_url}
-  -->
-
-EOS
-
-  header+txt  ## return txt w/ header
-end  ## method fetch
-
-end  ## class PageFetcher
-end  ## module Rsssf
-
-## add (shortcut) alias
-RsssfPageFetcher = Rsssf::PageFetcher
-