sportdb-parser 0.5.4 → 0.5.6

data/lib/sportdb/parser.rb

@@ -1,8 +1,5 @@
 ## pulls in
 require 'cocos'
-require 'season/formats'  # e.g. Season() support machinery
-
-
 
 
 ####
@@ -27,13 +24,9 @@ require_relative 'parser/token-status'
 require_relative 'parser/token'
 require_relative 'parser/tokenizer'
 
-require_relative 'parser/parser'
-
-
-####
-##  todo/check - move outline reader upstream to cocos - why? why not?
-##       use  read_outline(), parse_outline()  - why? why not?
-require_relative 'parser/outline_reader'
+require_relative 'parser/parser'   ## auto-generated by racc (from parser.y)
+require_relative 'parser/racc_parser'
+require_relative 'parser/racc_tree'
 
 
 
@@ -53,9 +46,11 @@ end  # module SportDb
 
 
 module SportDb
-
-
-
+###
+##  todo/fix -   use LangHelper or such
+##   e.g.     class Parser
+##                include LangHelper
+##            end
 class Parser
   ## keep "old" access to checking for group, round & friends
   ##    for now for compatibility
@@ -63,302 +58,9 @@ class Parser
   def is_round?( text )  Lang.is_round?( text ); end
   def is_leg?( text )    Lang.is_leg?( text ); end
 end
+end  # module SportDb
 
 
 
-class Tokenizer  
-    
-   attr_reader :tokens
- 
-   def initialize( txt )
-      parser = Parser.new
- 
-      tree = []
-      
-      lines = txt.split( "\n" )
-      lines.each_with_index do |line,i|
-          next if line.strip.empty? || line.strip.start_with?( '#' )
-          ##   support for inline (end-of-line) comments
-          line = line.sub(  /#.*/, '' ).strip
-
-          puts "line >#{line}<"
-          tokens = parser.tokenize( line )
-          pp tokens
-      
-          tree << tokens
-      end
- 
-
-=begin   
-      ## quick hack
-      ##   turn all  text tokens followed by minute token
-      ##     into player tokens!!!
-      ##
-      ##   also auto-convert text tokens into team tokens - why? why not?
-      tree.each do |tokens|
-         tokens.each_with_index do |t0,idx|
-            t1 = tokens[idx+1]
-            if t1 && t1[0] == :minute && t0[0] == :text
-                 t0[0] = :player 
-            end
-         end
-      end
-=end
-
-=begin
-## auto-add/insert start tokens for known line patterns
-##    START_GOALS for  goals_line
-##    why? why not?
-=end
-
-      ## flatten
-      @tokens = []
-      tree.each do |tokens|
-         @tokens += tokens 
-         @tokens  << [:NEWLINE, "\n"]   ## auto-add newlines
-      end
- 
-
-      ## convert to racc format
-      @tokens = @tokens.map do |tok|
-           if tok.size == 1
-             [tok[0].to_s, tok[0].to_s]
-           elsif tok.size == 2
- #############
- ## pass 1
- ##   replace all texts with keyword matches (e.g. group, round, leg, etc.)
-               if tok[0] == :TEXT
-                  text = tok[1]
-                  tok = 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
-               end
- ## pass 2
-              tok
-       else
-              raise ArgumentError, "tokens of size 1|2 expected; got #{tok.pretty_inspect}"
-           end
-      end
-   end
- 
- 
-
-   def next_token
-      @tokens.shift
-   end
- end  # class Tokenizer
-end # module SportDb
-
-
-
-####
-#   RaccMatchParser support machinery (incl. node classes/abstract syntax tree)
-
-class RaccMatchParser
-
-
-LineupLine = Struct.new( :team, :lineup ) do
-  def pretty_print( printer )
-    printer.text( "<LineupLine " )
-    printer.text( self.team )
-    printer.text( " lineup=" + self.lineup.pretty_inspect )
-    printer.text( ">" )
-  end
-end
-
-Lineup     = Struct.new( :name, :card, :sub ) do
-  def pretty_print( printer )
-    buf = String.new
-    buf <<  self.name 
-    buf << " card=" + self.card.pretty_inspect    if card
-    buf << " sub=" + self.sub.pretty_inspect      if sub
-    printer.text( buf ) 
-  end
-end
-
-
-Card       = Struct.new( :name, :minute ) do
-  def to_s
-    buf = String.new
-    buf << "#{self.name}"
-    buf << " #{self.minute.to_s}"   if self.minute
-    buf
-  end
-
-  def pretty_print( printer )
-    printer.text( to_s )
-  end  
-end
-
-
-Sub        = Struct.new( :minute, :sub )  do
-  def pretty_print( printer )
-    buf = String.new 
-    buf << "(#{self.minute.to_s} " 
-    buf << self.sub.pretty_inspect  
-    buf << ")"
-    printer.text( buf ) 
-  end
-end
-
-
-
-GroupDef   = Struct.new( :name, :teams ) do
-  def pretty_print( printer )
-    printer.text( "<GroupDef " )
-    printer.text( self.name )
-    printer.text( " teams=" + self.teams.pretty_inspect )
-    printer.text( ">" )
-  end
-end
-
-
-RoundDef   = Struct.new( :name, :date, :duration )  do
-  def pretty_print( printer )
-    printer.text( "<RoundDef " )
-    printer.text( self.name )
-    printer.text( " date=" + self.date.pretty_inspect ) if date
-    printer.text( " durattion=" + self.duration.pretty_inspect ) if duration
-    printer.text( ">" )
-  end
-end
-
-DateHeader = Struct.new( :date ) do
-  def pretty_print( printer )
-    printer.text( "<DateHeader " )
-    printer.text( "#{self.date.pretty_inspect}>" )
-  end
-end
-
-GroupHeader = Struct.new( :name ) do
-  def pretty_print( printer )
-    printer.text( "<GroupHeader " )
-    printer.text( "#{self.name}>" )
-  end
-end
-
-RoundHeader = Struct.new( :names ) do
-  def pretty_print( printer )
-    printer.text( "<RoundHeader " )
-    printer.text( "#{self.names.join(', ')}>" )
-  end
-end
-
-MatchLine   = Struct.new( :ord, :date, :time,
-                          :team1, :team2, :score,
-                          :status, 
-                          :geo,
-                          :timezone )  do   ## change to geos - why? why not?
-
-  def pretty_print( printer )
-    printer.text( "<MatchLine " )
-    printer.text( "#{self.team1} v #{self.team2}")
-    printer.breakable
-
-    members.zip(values) do |name, value|
-      next if [:team1, :team2].include?( name )
-      next if value.nil?
-      
-      printer.text( "#{name}=#{value.pretty_inspect}" )
-    end    
-
-    printer.text( ">" )
-  end  
-
-end
-
-GoalLine    = Struct.new( :goals1, :goals2 ) do
-  def pretty_print( printer )
-    printer.text( "<GoalLine " )
-    printer.text( "goals1=" + self.goals1.pretty_inspect + "," )
-    printer.breakable
-    printer.text( "goals2=" + self.goals2.pretty_inspect + ">" )
-  end  
-end
-
-Goal        = Struct.new( :player, :minutes ) do
-  def to_s
-    buf = String.new
-    buf << "#{self.player}"
-    buf << " "
-    buf << minutes.map { |min| min.to_s }.join(' ')
-    buf
-  end
-
-  def pretty_print( printer )
-    printer.text( to_s )
-  end  
-
-end
-
-
-##
-##  fix - move :og, :pen  to Goal if possible - why? why not?
-##  or change to GoalMinute ???
-Minute      = Struct.new( :m, :offset, :og, :pen )  do
-    def to_s
-      buf = String.new
-      buf << "#{self.m}"
-      buf << "+#{self.offset}"  if self.offset 
-      buf << "'"
-      buf << "(og)"   if self.og
-      buf << "(pen)"  if self.pen
-      buf
-    end
- 
-    def pretty_print( printer ) 
-       printer.text( to_s ) 
-    end  
-end
-
-
-
-
-def initialize(input)
-    puts "==> input:"
-    puts input
-    @tokenizer = SportDb::Tokenizer.new(input)
-  end
-  
-
-  def next_token
-    tok = @tokenizer.next_token
-    puts "next_token => #{tok.pretty_inspect}"
-    tok
-  end
-  
-#  on_error do |error_token_id, error_value, value_stack|
-#      puts "Parse error on token: #{error_token_id}, value: #{error_value}"
-#  end  
-
-  def parse    
-     puts "parse:"
-     @tree = [] 
-     do_parse
-     @tree
-  end
-
-
-  def on_error(*args)
-    puts
-    puts "!! on parse error:"
-    puts "args=#{args.pretty_inspect}"
-    exit 1  ##   exit for now  -  get and print more info about context etc.!!
-  end
-
-
-=begin
-on_error do |error_token_id, error_value, value_stack|
-    puts "Parse error on token: #{error_token_id}, value: #{error_value}"
-end
-=end
-
-end 
-
-
 puts SportDb::Module::Parser.banner    # say hello
 

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: sportdb-parser
 version: !ruby/object:Gem::Version
-  version: 0.5.4
+  version: 0.5.6
 platform: ruby
 authors:
 - Gerald Bauer
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2025-01-18 00:00:00.000000000 Z
+date: 2025-01-20 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: cocos
@@ -24,20 +24,6 @@ dependencies:
     - - ">="
       - !ruby/object:Gem::Version
         version: 0.4.0
-- !ruby/object:Gem::Dependency
-  name: season-formats
-  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: racc
   requirement: !ruby/object:Gem::Requirement
@@ -111,8 +97,9 @@ files:
 - config/rounds_pt.txt
 - lib/sportdb/parser.rb
 - lib/sportdb/parser/lang.rb
-- lib/sportdb/parser/outline_reader.rb
 - lib/sportdb/parser/parser.rb
+- lib/sportdb/parser/racc_parser.rb
+- lib/sportdb/parser/racc_tree.rb
 - lib/sportdb/parser/token-date.rb
 - lib/sportdb/parser/token-score.rb
 - lib/sportdb/parser/token-status.rb

data/lib/sportdb/parser/outline_reader.rb

@@ -1,155 +0,0 @@
-
-
-module SportDb
-
-###
-#  add a simple Outline convenience class
-#            for processing OUtlines with OUtlineReader
-
-class QuickMatchOutline
-   def self.read( path )
-       nodes = OutlineReader.read( path )
-       new( nodes )
-   end      
-
-   def initialize( nodes )
-      @nodes = nodes
-   end
-
-   def each_para( &blk )
-     ## note: every (new) read call - resets errors list to empty
-     ### @errors = []
-
-     ##  process nodes
-     h1 = nil
-     h2 = nil
-     orphans = 0    ## track paragraphs's with no heading
-
-     @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 processing"
-             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]
-           blk.call( lines )
-        else
-          pp node
-          raise ArgumentError, "unsupported (node) type >#{type}<"
-        end
-     end  # each node
-   end  # each_para
-   alias_method :each_paragraph, :each_para
-   alias_method :each_p,         :each_para
-end   # class QuickMatchOutline
-
-
-
-
-class OutlineReader
-
-  def self.debug=(value) @@debug = value; end
-  def self.debug?() @@debug ||= false; end
-  def debug?()  self.class.debug?; end
-
-
-
-  def self.read( path )   ## use - rename to read_file or from_file etc. - why? why not?
-    txt = File.open( path, 'r:utf-8' ) {|f| f.read }
-    parse( txt )
-  end
-
-  def self.parse( txt )
-    new( txt ).parse
-  end
-
-  def initialize( txt )
-    @txt = txt
-  end
-
-  ## note: skip "decorative" only heading e.g. ========
-  ##  todo/check:  find a better name e.g. HEADING_EMPTY_RE or HEADING_LINE_RE or ???
-  HEADING_BLANK_RE = %r{\A
-                        ={1,}
-                        \z}x
-
-  ## note: like in wikimedia markup (and markdown) all optional trailing ==== too
-  HEADING_RE = %r{\A
-                  (?<marker>={1,})       ## 1. leading ======
-                    [ ]*
-                  (?<text>[^=]+)         ## 2. text   (note: for now no "inline" = allowed)
-                    [ ]*
-                    =*                   ## 3. (optional) trailing ====
-                  \z}x
-
-  def parse
-    outline=[]   ## outline structure
-    start_para = true      ## start new para(graph) on new text line?
-
-    @txt.each_line do |line|
-        line = line.strip      ## todo/fix: keep leading and trailing spaces - why? why not?
-
-        if line.empty?    ## todo/fix: keep blank line nodes?? and just remove comments and process headings?! - why? why not?
-          start_para = true
-          next
-        end
-
-        break if line == '__END__'
-
-        next if line.start_with?( '#' )   ## skip comments too
-        ## strip inline (until end-of-line) comments too
-        ##  e.g Eupen | KAS Eupen ## [de]
-        ##   => Eupen | KAS Eupen
-        ##  e.g bq   Bonaire,  BOE        # CONCACAF
-        ##   => bq   Bonaire,  BOE
-        line = line.sub( /#.*/, '' ).strip
-        pp line    if debug?
-
-        ## todo/check: also use heading blank as paragraph "breaker" or treat it like a comment ?? - why? why not?
-        next if HEADING_BLANK_RE.match( line )  # skip "decorative" only heading e.g. ========
-
-         ## note: like in wikimedia markup (and markdown) all optional trailing ==== too
-        if m=HEADING_RE.match( line )
-           start_para = true
-
-           heading_marker = m[:marker]
-           heading_level  = heading_marker.length   ## count number of = for heading level
-           heading        = m[:text].strip
-
-           puts "heading #{heading_level} >#{heading}<"   if debug?
-           outline << [:"h#{heading_level}", heading]
-        else    ## assume it's a (plain/regular) text line
-           if start_para
-             outline << [:p, [line]]
-             start_para = false
-           else
-             node = outline[-1]    ## get last entry
-             if node[0] == :p      ##  assert it's a p(aragraph) node!!!
-                node[1] << line    ## add line to p(aragraph)
-             else
-               puts "!! ERROR - invalid outline state / format - expected p(aragraph) node; got:"
-               pp node
-               exit 1
-             end
-           end
-        end
-    end
-    outline
-  end # method read
-end # class OutlineReader
-
-end # module SportDb