fbtok 0.3.3 → 0.4.0

data/lib/fbtok/pathspec.rb

@@ -0,0 +1,230 @@
+
+
+module SportDb
+
+class Pathspec
+
+  def self.debug=(value) @@debug = value; end
+  def self.debug?()      @@debug ||= false; end  ## note: default is FALSE
+
+
+    SEASON_RE = %r{ (?:
+                          (?<season>\d{4}-\d{2})
+                        | (?<season>\d{4})
+                    )
+                  }x
+
+    ## note: if pattern includes directory add here
+    ##     (otherwise move to more "generic" datafile) - why? why not?
+    ##   update - note include/allow dot (.) too
+    ##     BUT NOT as first character!!! (e.g. exclude .confg.txt !!!)
+    ##               e.g. 2024-25/at.1.txt
+    ##                        change to at_1 or uefa_cl or such - why? why not?
+    MATCH_RE = %r{
+                    ## "classic" variant i)  with season folder
+                    ##     e.g.  /1930/cup.txt
+                    (?:
+                      (?: ^|/ )    # beginning (^) or beginning of path (/)
+
+                       #{SEASON_RE}
+                          (?:  --[a-z0-9_-]+
+                          )?
+                         /
+                       [a-z0-9][a-z0-9._-]* \.txt   ## e.g /1-premierleague.txt
+                         $
+                    )
+
+                      |
+                    ## "compact" variant ii) with season in filename
+                    ##         e.g. 1930.txt or 2024_br.txt etc.
+                    (?:
+                       (?: ^|/ )      # beginning (^) or beginning of path (/)
+                         #{SEASON_RE}
+                         (?:   _   ## allow more than one underscore - why? why not?
+                            [a-z0-9][a-z0-9._-]*
+                          )?
+                            \.txt
+                            $
+                    )
+                }x
+
+    ### add support for matchdatafile with season NOT in directory
+    ##      but starting filename e.g. 2024_friendlies.txt or 2024-25_bundesliga.txt
+
+
+def self._norm_seasons( seasons )
+    seasons.map {|season| Season(season) }
+end
+
+
+## todo/check - rename to glob or expand or such - why? why not?
+def self._find( path, seasons: nil )
+    ## check - rename dir
+    ##          use root_dir or work_dir or cd or such - why? why not?
+
+
+    ## note: normalize path - use File.expand_path ??
+    ##    change all backslash to slash for now
+    ## path = path.gsub( "\\", '/' )
+    path =  File.expand_path( path )
+
+    ####
+    ## note - make sure path exists; raise error if not
+    raise Errno::ENOENT, "No such directory - #{path})"  unless Dir.exist?( path )
+
+
+    if seasons && seasons.size > 0
+       seasons = _norm_seasons( seasons )    ## norm seasons (string, integer => Season obj)
+    end
+
+
+    ## check all txt files
+    ## note: incl. files starting with dot (.)) as candidates
+    ##     (normally excluded with just *)
+    ##  was:  Dir.glob( "#{path}/**/{*,.*}.txt" )
+
+    candidates = Dir.glob( "#{path}/**/*.txt" )
+    ## pp candidates
+
+
+    datafiles = []
+    candidates.each do |candidate|
+       if m = MATCH_RE.match( candidate )
+
+          ## check for seasons filter
+          next   if seasons && seasons.size > 0 &&
+                    !seasons.include?( Season.parse( m[:season] ))
+
+          ### exclude squad
+          ##   and .v2 or .v2603
+          ##
+          ##    - worldcup/more/1930_squads.txt   =>   squads
+          ##    - 2014--brazil/cup.v2.txt",
+          ##    - 2014--brazil/cup.v260318_164934.txt",
+
+          basename = File.basename( candidate, File.extname( candidate ))
+
+          next   if /squad/i.match( basename )
+          next   if /\.v[0-9][0-9_]*/i.match( basename )
+
+
+          datafiles << candidate
+       end
+    end
+
+    ## pp datafiles
+    datafiles
+end
+
+
+
+
+
+def self.read( src )
+    recs = read_csv( src )
+    pp recs     if debug?
+
+    ##  note - make pathspecs relative to passed in file arg!!!
+    basedir = File.dirname( src )
+
+    recs.each do |rec|
+        path = rec['path']
+        fullpath = File.expand_path( path, basedir )
+        datafiles =   _find( fullpath )
+
+        ## add (new) datafiles column (from expanded pathspec)
+        rec['datafiles'] = datafiles
+    end
+
+    recs
+end
+
+
+
+
+end  # class Pathspec
+
+
+
+##
+#  PathspecReport (aka/formerly BatchReport)
+
+class PathspecReport
+  def initialize( specs, title: )
+     @specs = specs
+     @title = title
+  end
+
+  def build
+     buf = String.new
+     buf << "# #{@title} - #{@specs.size} dataset(s)\n\n"
+
+     @specs.each_with_index do |rec,i|
+       datafiles  = rec['datafiles']
+       errors     = rec['errors']
+
+       if errors.size > 0
+         buf << "!! #{errors.size} ERROR(S)  "
+       else
+         buf << "   OK          "
+       end
+       buf << "%-20s" % rec['path']
+       buf << " - #{datafiles.size} datafile(s)"
+       buf << "\n"
+
+       if errors.size > 0
+         buf << errors.pretty_inspect
+         buf << "\n"
+       end
+     end
+
+     buf
+  end   # method build
+end  # class BatchReport
+
+BatchReport = PathspecReport
+
+end  # module Sportdb
+
+
+
+##
+##  keep helpers as global functions - why? why not?
+
+
+
+def build_pathspecs( args )    ### note:   was expand_args
+   specs = []
+
+   ## note - collect all "loose/standalone" files (NOT directories)
+   ##             in single default pathspec node
+   more = []
+
+   args.each do |arg|
+    ## check if directory
+    if Dir.exist?( arg )
+          datafiles = SportDb::Pathspec._find( arg )
+          specs << { 'path'       => arg,
+                     'datafiles'  => datafiles }
+    elsif File.file?( arg )  ## assume it's a file
+        ## make sure path exists; raise error if not
+        ##   (auto-)expand path to normalize - why? why not?
+          more << arg
+    else
+        raise Errno::ENOENT, "No such file or directory - #{arg}"
+    end
+  end
+
+   if more.size > 0
+      specs << { 'path'      => '<input>',
+                 'datafiles' =>  more }
+   end
+
+  specs
+end
+
+
+
+def read_pathspecs( src )
+    SportDb::Pathspec.read( src )
+end

data/lib/fbtok.rb

@@ -1,13 +1,6 @@
 
-require 'sportdb/formats'
-
-
-## more stdlibs
-require 'optparse'   ## check - already auto-required in cocos? keep? why? why not?
+require 'sportdb/quick'
 
 
 ## our own code
-require_relative 'fbtok/opts'
-require_relative 'fbtok/linter'
-
-
+require_relative 'fbtok/pathspec'

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: fbtok
 version: !ruby/object:Gem::Version
-  version: 0.3.3
+  version: 0.4.0
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-02-25 00:00:00.000000000 Z
+date: 2026-05-26 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: sportdb-parser
@@ -16,28 +16,28 @@ dependencies:
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.5.8
+        version: 0.7.1
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 0.5.8
+        version: 0.7.1
 - !ruby/object:Gem::Dependency
-  name: sportdb-formats
+  name: sportdb-quick
   requirement: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.1.2
+        version: 0.7.0
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
     - - ">="
       - !ruby/object:Gem::Version
-        version: 2.1.2
+        version: 0.7.0
 - !ruby/object:Gem::Dependency
   name: rdoc
   requirement: !ruby/object:Gem::Requirement
@@ -75,8 +75,7 @@ dependencies:
 description: fbtok - football.txt lint tools incl. tokenizer, parser & more
 email: gerald.bauer@gmail.com
 executables:
-- fbchk
-- fbt
+- fbquick
 - fbtok
 - fbtree
 - fbx
@@ -90,14 +89,12 @@ files:
 - Manifest.txt
 - README.md
 - Rakefile
-- bin/fbchk
-- bin/fbt
+- bin/fbquick
 - bin/fbtok
 - bin/fbtree
 - bin/fbx
 - lib/fbtok.rb
-- lib/fbtok/linter.rb
-- lib/fbtok/opts.rb
+- lib/fbtok/pathspec.rb
 homepage: https://github.com/sportdb/footty
 licenses:
 - Public Domain

data/bin/fbchk

@@ -1,158 +0,0 @@
-#!/usr/bin/env ruby
-
-## tip: to test run:
-##   ruby -I ./lib bin/fbchk
-
-
-## local hack for debugging
-##   use local versions if present
-$LOAD_PATH.unshift( File.expand_path( '/sports/sportdb/sport.db/sportdb-structs/lib' ))
-$LOAD_PATH.unshift( File.expand_path( '/sports/sportdb/sport.db/sportdb-catalogs/lib' ))
-$LOAD_PATH.unshift( File.expand_path( '/sports/sportdb/sport.db/sportdb-search/lib' ))
-$LOAD_PATH.unshift( File.expand_path( '/sports/sportdb/sport.db/sportdb-formats/lib' ))
-
-
-## our own code
-require 'fbtok'
-
-
-
-## local hack
-##   if exists up-to-date catalog db (use local version NOT built-in)
-catalog_path = '/sports/sportdb/sport.db/catalog/catalog.db'
-if File.exist?( catalog_path )
-  SportDb::Import.config.catalog_path = catalog_path
-end
-
-
- args = ARGV
- opts = { debug: false,
-          file:  nil,
-          teams: true,   ## check/lint teams (name errros etc.)
-        }
-
- parser = OptionParser.new do |parser|
-  parser.banner = "Usage: #{$PROGRAM_NAME} [options]"
-
-##
-## check if git has a offline option?? (use same)
-##             check for other tools - why? why not?
-#    parser.on( "-q", "--quiet",
-#                 "less debug output/messages - default is (#{!opts[:debug]})" ) do |debug|
-#      opts[:debug] = false
-#    end
-   parser.on( "--verbose", "--debug",
-               "turn on verbose / debug output (default: #{opts[:debug]})" ) do |debug|
-     opts[:debug] = true
-   end
-
-    parser.on( "-f FILE", "--file FILE",
-                    "read datafiles (pathspecs) via .csv file") do |file|
-      opts[:file] = file
-    end
-
-    parser.on( "--[no-]teams",
-                  "turn on/off team name checks (default: #{opts[:teams]})") do |teams|
-      opts[:teams] = teams
-    end
-end
-parser.parse!( args )
-
-
-if opts[:debug]
-  puts "OPTS:"
-  p opts
-  puts "ARGV:"
-  p args
-
-  SportDb::QuickMatchLinter.debug = true
-  SportDb::QuickMatchReader.debug = true
-  SportDb::MatchParser.debug      = true
-else
-  SportDb::QuickMatchLinter.debug = false
-  SportDb::QuickMatchReader.debug = false
-  SportDb::MatchParser.debug      = false
-  LogUtils::Logger.root.level = :info
-end
-
-
-
-
-## todo/check - use packs or projects or such
-##                instead of specs - why? why not?
-specs =  if opts[:file]
-            SportDb::Parser::Opts.read_pathspecs( opts[:file] )
-         else
-           paths =  if args.empty?
-                      [] 
-                    else
-                       ## check for directories
-                       ##   and auto-expand
-                       SportDb::Parser::Opts.expand_args( args )
-                    end
-             ## always return array of specs
-            [SportDb::Parser::Opts.build_pathspec( paths: paths)]  
-         end
-
-
-
-
-  specs.each_with_index do |spec,i|
-    paths = spec.paths
-    rec   = spec.rec
-
-    errors = []
-    paths.each_with_index do |path,j|
-      puts "==> [#{i+1}/#{specs.size}, #{j+1}/#{paths.size}] reading >#{path}<..."
-      quick = SportDb::QuickMatchLinter.new( read_text( path ),
-                                              check_teams: opts[:teams] )
-      matches = quick.parse
-
-
-      if quick.errors?
-        puts "!! #{quick.errors.size} error(s):"
-        pp quick.errors
-
-        quick.errors.each do |err|
-          errors << [ path, *err ]   # note: use splat (*) to add extra values (starting with msg)
-        end
-      end
-      puts "  #{matches.size} match(es)"
-    end
-
-    if errors.size > 0
-      puts
-      puts "!! #{errors.size} PARSE ERRORS in #{paths.size} datafile(s)"
-      pp errors
-    else
-      puts
-      puts "  OK - no parse errors in #{paths.size} datafile(s)"
-    end
-
-     ## add errors to rec via rec['errors'] to allow
-     ##   for further processing/reporting
-     rec['errors'] = errors
-  end
-
-
-
-###
-## generate a report if --file option used
-if opts[:file]
-  buf = SportDb::Parser::BatchReport.new( 
-            specs, 
-            title: 'fbchk summary report' ).build
-
-  puts
-  puts "SUMMARY:"
-  puts buf
-
-  basedir  = File.dirname( opts[:file] )
-  basename = File.basename( opts[:file], File.extname( opts[:file] ))
-  outpath  = "#{basedir}/fbchk.#{basename}.txt"
-  write_text( outpath, buf )
-end
-
-
-puts "bye"
-

data/lib/fbtok/linter.rb

@@ -1,119 +0,0 @@
-
-module SportDb
-class Parser
-
-###
-## note - Linter for now nested inside Parser - keep? why? why not?
-class Linter
-
-def self.debug=(value) @@debug = value; end
-def self.debug?()      @@debug ||= false; end  ## note: default is FALSE
-def debug?()  self.class.debug?; end
-
-##  turn warn(ings) into errors (default is false)
-def self.warn=(value) @@warn = value; end
-def self.warn?()      @@warn ||= false; end  ## note: default is FALSE
-def warn?()   self.class.warn?; end
-
-
-attr_reader :errors
-
-def initialize
-  @errors = []
-end
-
-
-def errors?() @errors.size > 0; end
-
-
-
-#########
-## parse - false (default) - tokenize (only)
-##       - true            - tokenize & parse
-def read( path, parse: true )
-  ## note: every (new) read call - resets errors list to empty
-  @errors = []
-  @tree   = []
-
-  outline = QuickMatchOutline.read( path )
-
-  outline.each_para do |lines|
-  
-       ## flatten lines (array of strings) into all-in-one string
-       txt  = lines.reduce( String.new ) do |mem, line|
-         mem << line
-         mem << "\n"
-         mem
-       end
-
-     if parse
-       if debug?
-         puts "lines:"
-         pp txt   
-       end
- 
-       ##   pass along debug flag to parser (& tokenizer)?
-       parser = RaccMatchParser.new( txt )   ## use own parser instance (not shared) - why? why not?
-       tree, errors = parser.parse_with_errors
-
-         if errors.size > 0
-            ## add to "global" error list
-            ##   make a triplet tuple (file / msg / line text)
-            errors.each do |msg|
-                @errors << [path, msg]
-            end
-         end
-
-       if debug?
-         puts "parse tree:"
-         pp tree  
-       end
-
-       @tree += tree   ## add nodes
-
-     else   ## process for tokenize only
-
-        if debug?
-          puts "lines:"
-          pp txt   
-        end
-
-##
-##   add (bakc) a line-by-line tracing (debug) option - why? why not?
-##      now debug output is by section (not line-by-line)
-
-        lexer = Lexer.new( txt )
-        t, errors  =  lexer.tokenize_with_errors
-                            
-         if errors.size > 0
-            ## add to "global" error list
-            ##   make a triplet tuple (file / msg / line text)
-            errors.each do |msg|
-                @errors << [path, msg]
-            end
-         end
-
-         if debug?
-           puts "tokens:"
-           pp t   
-         end
-
-         @tree += t   ## add tokens to "tree"
-      end   # parse? (or tokenize?) 
-   end  # each para (node)
-
-
-   ##
-   ## auto-add error if no tokens/tree nodes
-   ##   note - do NOT report empty file error by default 
-   if @tree.empty? && warn?  ## @tree.size == 0
-      @errors << [path, "empty; no #{parse ? 'parse tree nodes' : 'tokens'}"]
-   end
-
-   @tree   ## return parse tree 
-end  # method read
-end  # class Linter
-
-
-end   # class Parser
-end   # module SportDb