fbtok 0.1.1 → 0.2.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: b4ab88883b8f80ee4824cab352297f037e2afcb72351bf6956a8e48bddd5a164
-  data.tar.gz: 6bd75bea2015124e34b01f506757a54a5537ca0544559a0377b6f0a6a0533c0b
+  metadata.gz: 0db5fb0d9cbda9ad626f4debbf27751f591c90684455e2e3cabc66162a02e3c6
+  data.tar.gz: cc4d1d160356785e54b99fb7033e078af9edf63cd60a469ab8bb2ad685603b7f
 SHA512:
-  metadata.gz: 94b3582db70136fdeadeebe003fefdff304ca923bd76edcc0a40d28731c64e73f372b7c0ecf74e58623358102c32ad51e0a233becee4555d5ccd322552d53d80
-  data.tar.gz: dca47c3d3e134ed9c103ba873e76e0c4793d690bbaf6ced8a34693e4c2fea3fc1ff43af0d45ebc62738022f02bb02a0e0bcc04c153dd752a49c7a50d62ba6bba
+  metadata.gz: d688b40a4c97971eab87594a83ffe7966b245370944ac140950f0e935076d25c4317dfcb526ac0f22470c1027a7bc488812f37f07f6d1d413edb8e7c538c3f2b
+  data.tar.gz: beefaf087fe66d2ab3ecaae23cc786184e0558b3dd7bcc44e4e131cf70fca200a3f723060c42a2921d5cbef1f2b6513dc567dc693a7dd03bc2adcd8b7ad3a6e2

data/CHANGELOG.md

@@ -1,4 +1,4 @@
-### 0.1.1
+### 0.2.0
 ### 0.0.1 / 2025-01-02
 
 * Everything is new. First release.

data/Manifest.txt

@@ -5,6 +5,7 @@ Rakefile
 bin/fbchk
 bin/fbt
 bin/fbtok
+bin/fbtree
 bin/fbx
 lib/fbtok.rb
 lib/fbtok/linter.rb

data/Rakefile

@@ -2,7 +2,7 @@ require 'hoe'
 
 
 Hoe.spec 'fbtok' do
-  self.version = '0.1.1'
+  self.version = '0.2.0'
 
   self.summary = "fbtok - football.txt lint tools incl. tokenizer, parser & more"
   self.description = summary
@@ -19,9 +19,9 @@ Hoe.spec 'fbtok' do
   self.licenses = ['Public Domain']
 
   self.extra_deps = [
-     #  ['sportdb-parser',  '>= 0.2.2'],
      #  ['sportdb-structs', '>= 0.5.0'],
      #  ['logutils',        '>= 0.6.1'],
+     ['sportdb-parser',  '>= 0.5.0'],
      ['sportdb-formats', '>= 2.1.2'],
   ]
 

data/bin/fbtok

@@ -10,7 +10,6 @@ args = ARGV
 
 opts = {
     debug: true,
-    metal: false,
     file:  nil,
 }
 
@@ -27,11 +26,14 @@ parser = OptionParser.new do |parser|
     opts[:debug] = true
   end
 
+=begin
   parser.on( "--metal",
                  "turn off typed parse tree; show to the metal tokens"+
                    " (default: #{opts[:metal]})" ) do |metal|
     opts[:metal] = true
   end
+=end
+
 
   parser.on( "-f FILE", "--file FILE",
                 "read datafiles (pathspecs) via .csv file") do |file|
@@ -85,7 +87,7 @@ specs.each_with_index do |(paths, rec),i|
 
    paths.each_with_index do |path,j|
       puts "==> [#{j+1}/#{paths.size}] reading >#{path}<..."
-      linter.read( path, parse: !opts[:metal] )
+      linter.read( path )
 
       errors += linter.errors    if linter.errors?
    end
@@ -94,10 +96,10 @@ specs.each_with_index do |(paths, rec),i|
       puts
       pp errors
       puts
-      puts "!!   #{errors.size} parse error(s) in #{paths.size} datafiles(s)"
+      puts "!!   #{errors.size} tokenize error(s) in #{paths.size} datafiles(s)"
    else
       puts
-      puts "OK   no parse errors found in #{paths.size} datafile(s)"
+      puts "OK   no tokenize errors found in #{paths.size} datafile(s)"
    end
 
    ## add errors to rec via rec['errors'] to allow

data/bin/fbtree

@@ -0,0 +1,227 @@
+#!/usr/bin/env ruby
+
+## tip: to test run:
+##   ruby -I ./lib bin/fbtree
+
+require 'fbtok'
+
+
+###
+## note - Linter for now nested inside Parser - keep? why? why not?
+class RaccLinter
+
+def self.debug=(value) @@debug = value; end
+def self.debug?() @@debug ||= false; end  ## note: default is FALSE
+def debug?()  self.class.debug?; end
+
+
+
+attr_reader :errors
+
+def initialize
+  @errors = []
+end
+
+
+def errors?() @errors.size > 0; end
+
+
+
+  ## note:  colon (:) MUST be followed by one (or more) spaces
+  ##      make sure mon feb 12 18:10 will not match
+  ##        allow 1. FC Köln etc.
+  ##               Mainz 05:
+  ##           limit to 30 chars max
+  ##          only allow  chars incl. intl buut (NOT ()[]/;)
+  ##
+  ##   Group A:
+  ##   Group B:   - remove colon
+  ##    or lookup first
+
+  ATTRIB_RE = %r{^
+                   [ ]*?     # slurp leading spaces
+                (?<key>[^:|\]\[()\/; -]
+                       [^:|\]\[()\/;]{0,30}
+                 )
+                   [ ]*?     # slurp trailing spaces
+                   :[ ]+
+                (?<value>.+)
+                    [ ]*?   # slurp trailing spaces
+                   $
+                }ix
+
+
+#########
+## parse - false (default) - tokenize (only)
+##       - true            - tokenize & parse
+def read( path )
+  ## note: every (new) read call - resets errors list to empty
+  @errors = []
+
+  nodes = SportDb::OutlineReader.read( path )
+
+  ##  process nodes
+  h1 = nil
+  h2 = nil
+  orphans = 0    ## track paragraphs's with no heading
+
+  attrib_found = false
+
+
+  nodes.each do |node|
+    type = node[0]
+
+    if type == :h1
+        h1 = node[1]  ## get heading text
+        puts "  = Heading 1 >#{node[1]}<"
+    elsif type == :h2
+        if h1.nil?
+          puts "!! WARN - no heading for subheading; skipping parse"
+          next
+        end
+        h2 = node[1]  ## get heading text
+        puts "  == Heading 2 >#{node[1]}<"
+    elsif type == :p
+
+       if h1.nil?
+         orphans += 1    ## only warn once
+         puts "!! WARN - no heading for #{orphans} text paragraph(s); skipping parse"
+         next
+       end
+
+       lines = node[1]
+
+       txt  = []
+       lines.each_with_index do |line,i|
+
+        if debug?
+         puts
+         puts "line >#{line}<"
+        end
+
+
+        ## skip new (experimental attrib syntax)
+        if attrib_found == false &&
+            ATTRIB_RE.match?( line )
+          ## note: check attrib regex AFTER group def e.g.:
+          ##         Group A:
+          ##         Group B:  etc.
+          ##     todo/fix - change Group A: to Group A etc.
+          ##                       Group B: to Group B
+           attrib_found = true
+           ## logger.debug "skipping key/value line - >#{line}<"
+           next
+        end
+
+        if attrib_found
+          ## check if line ends with dot
+          ##  if not slurp up lines to the next do!!!
+          ## logger.debug "skipping key/value line - >#{line}<"
+          attrib_found = false   if line.end_with?( '.' )
+              # logger.debug "skipping key/value line (cont.) - >#{line}<"
+              next
+        end
+
+        txt << line
+        txt << "\n"
+      end
+
+      ## flatten
+      txt = txt.join
+      pp txt   if debug?
+
+      parser = RaccMatchParser.new( txt )   ## use own parser instance (not shared) - why? why not?
+      tree = parser.parse
+      pp tree
+    else
+        pp node
+        raise ArgumentError, "unsupported (node) type >#{type}<"
+    end
+  end  # each node
+end  # read
+end  # class RaccLinter
+
+
+
+
+###############################################
+#  start with code
+
+
+
+args = ARGV
+ 
+
+opts = {
+    debug: true,
+}
+
+parser = OptionParser.new do |parser|
+  parser.banner = "Usage: #{$PROGRAM_NAME} [options] PATH"
+
+
+  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
+end
+parser.parse!( args )
+
+puts "OPTS:"
+p opts
+puts "ARGV:"
+p args
+
+
+## todo/check - use packs or projects or such
+##                instead of specs - why? why not?
+   paths = if args.empty?
+             [
+               '../../../openfootball/euro/2021--europe/euro.txt',
+               '../../../openfootball/euro/2024--germany/euro.txt',
+             ]
+          else
+            ## check for directories
+            ##   and auto-expand
+            SportDb::Parser::Opts.expand_args( args )
+          end
+
+
+
+
+errors = []
+linter = RaccLinter.new
+
+paths.each_with_index do |path,i|
+
+  puts "==> [#{i+1}/#{paths.size}] reading >#{path}<..."
+
+  linter.read( path )
+end
+
+puts "bye"
+
+
+
+
+__END__
+
+   if errors.size > 0
+      puts
+      pp errors
+      puts
+      puts "!!   #{errors.size} parse error(s) in #{paths.size} datafiles(s)"
+   else
+      puts
+      puts "OK   no parse errors found in #{paths.size} datafile(s)"
+   end
+
+   ## add errors to rec via rec['errors'] to allow
+   ##   for further processing/reporting
+   rec['errors'] = errors
+end
+

data/bin/fbx

@@ -88,12 +88,12 @@ paths.each_with_index do |path,i|
     ## pp secs
 
     secs.each_with_index do |sec,j|   ## sec(tion)s
-      season = sec[:season]
+      season = Season.parse( sec[:season] )   ## convert (str) to season obj!!!
       league = sec[:league]
       stage  = sec[:stage]
       lines  = sec[:lines]
 
-      puts "   section #{j+1}/#{secs.size} -   #{league.name} #{season}, #{stage}  -  #{lines.size} line(s)"
+      puts "   section #{j+1}/#{secs.size} -   #{league} #{season}, #{stage}  -  #{lines.size} line(s)"
 
       next if opts[:outline]
 

data/lib/fbtok/linter.rb

@@ -51,7 +51,7 @@ def errors?() @errors.size > 0; end
 #########
 ## parse - false (default) - tokenize (only)
 ##       - true            - tokenize & parse
-def read( path, parse: false )
+def read( path  )
   ## note: every (new) read call - resets errors list to empty
   @errors = []
 
@@ -119,13 +119,8 @@ def read( path, parse: false )
               next
         end
 
-        t, error_messages  =  if parse
-                                  @parser.parse_with_errors( line )
-                              else
-                                  @parser.tokenize_with_errors( line )
-                              end
-
-
+        t, error_messages  =  @parser.tokenize_with_errors( line )
+                            
          if error_messages.size > 0
             ## add to "global" error list
             ##   make a triplet tuple (file / msg / line text)
@@ -137,6 +132,26 @@ def read( path, parse: false )
             end
          end
 
+          ## post-process tokens
+          ##  - check for round, group, etc.
+          t = t.map do |tok|
+            #############
+            ## pass 1
+            ##   replace all texts with keyword matches (e.g. group, round, leg, etc.)
+                if tok[0] == :TEXT
+                   text = tok[1]
+                    if @parser.is_group?( text )
+                           [:GROUP, text]
+                    elsif @parser.is_round?( text ) || @parser.is_leg?( text )
+                           [:ROUND, text]
+                    else
+                        tok  ## pass through as-is (1:1)
+                    end
+                else
+                   tok
+                end
+          end
+ 
          pp t   if debug?
 
          tree << t

data/lib/fbtok/opts.rb

@@ -18,11 +18,12 @@ class Opts
     ## 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{ (?: ^|/ )      # beginning (^) or beginning of path (/)
                        #{SEASON}
-                     /[a-z0-9_.-]+\.txt$  ## txt e.g /1-premierleague.txt
+                     /[a-z0-9][a-z0-9_.-]*\.txt$  ## txt e.g /1-premierleague.txt
                 }x
 
 

metadata

@@ -1,15 +1,29 @@
 --- !ruby/object:Gem::Specification
 name: fbtok
 version: !ruby/object:Gem::Version
-  version: 0.1.1
+  version: 0.2.0
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-01-02 00:00:00.000000000 Z
+date: 2025-01-15 00:00:00.000000000 Z
 dependencies:
+- !ruby/object:Gem::Dependency
+  name: sportdb-parser
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.5.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ">="
+      - !ruby/object:Gem::Version
+        version: 0.5.0
 - !ruby/object:Gem::Dependency
   name: sportdb-formats
   requirement: !ruby/object:Gem::Requirement
@@ -64,6 +78,7 @@ executables:
 - fbchk
 - fbt
 - fbtok
+- fbtree
 - fbx
 extensions: []
 extra_rdoc_files:
@@ -78,6 +93,7 @@ files:
 - bin/fbchk
 - bin/fbt
 - bin/fbtok
+- bin/fbtree
 - bin/fbx
 - lib/fbtok.rb
 - lib/fbtok/linter.rb