sportdb-parser 0.6.3 → 0.6.4

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: d1dd9f29584bb09508c0fb6ad253a84ff299c7f33650d27195abbc679c1e893f
-  data.tar.gz: 6af07bc6f008e1b6c53cd796253f66bd68f2f28299a750896be83176bae7d085
+  metadata.gz: 00ec5bcacfe56b29e9589507c11d3bfe361caed4b45ebdfa3b05901c8229b019
+  data.tar.gz: bd36a9b6c0b84a9a033d721c4adf086ff25703ed31dde3fc8265c421ba6273c1
 SHA512:
-  metadata.gz: f8adae46063426008e23f01ef3d625a3477c01a827b942edadddfad7e774e8689de3545c1c6ae8e32d0fa43b2bfbf5c9dc9962947c97be79ab325a41a7896c48
-  data.tar.gz: e0b843c9b2eee0ebfec203889ed5ea90f94eb22f80ca0df58d8865f05dd08f4d4bc3547b59f0fbe00e323a09df9c13289954ddea34bf437cad6d196c29a2e436
+  metadata.gz: db5568eb30b924f0e963402ed3089edc1b1cdbcf31cfaa4177da6215b82e0646badd428e94a14a42abfbb4688c7543f8b04f698eb1551082e9731b1937a23e19
+  data.tar.gz: d3619f1ea496cf4fdc08d0904afee295679d6f1aa0c35aacd09ccf8e19ad00beb78f832efef3f24189b8793b0dd24064544854e84d7d26653398962d79bda311

data/CHANGELOG.md

@@ -1,4 +1,4 @@
-### 0.6.3
+### 0.6.4
 ### 0.0.1 / 2024-07-12
 
 * Everything is new. First release.

data/lib/sportdb/parser/lexer.rb

@@ -246,9 +246,10 @@ def tokenize_with_errors
              nodes << [:TEAM, buf.next[1]]
              nodes << buf.next
              nodes << [:TEAM, buf.next[1]]
-          elsif buf.match?( :TEXT, :MINUTE )
-             nodes << [:PLAYER, buf.next[1]]
-             nodes << buf.next
+   #   note - now handled (upstream) with GOAL_RE mode!!!
+   #       elsif buf.match?( :TEXT, :MINUTE )
+   #          nodes << [:PLAYER, buf.next[1]]
+   #          nodes << buf.next
           elsif buf.match?( :DATE, :TIME )   ## merge DATE TIME into DATETIME
                date = buf.next[1]
                time = buf.next[1]
@@ -315,11 +316,85 @@ def _tokenize_line( line )
   @re  ||= RE     ## note - switch between RE & INSIDE_RE
 
 
+  if @re == RE  ## top-level
+    ### check for modes once (per line) here to speed-up parsing
+    ###   for now goals only possible for start of line!!
+    ###        fix - remove optional [] - why? why not?  
+  
+    ##  start with prop key (match will switch into prop mode!!!)
+    ##   - fix - remove leading spaces in regex (upstream) - why? why not?
+    m = PROP_KEY_RE.match( line )
+    if m
+      ###  switch into new mode
+      ##  switch context  to PROP_RE
+        @re = PROP_RE
+        puts "  ENTER PROP_RE MODE"   if debug?
+        tokens << [:PROP, m[:key]]
+
+        offsets = [m.begin(0), m.end(0)]
+        pos = offsets[1]    ## update pos
+    end
+
+    m = PLAYER_WITH_SCORE_RE.match( line )
+    if m
+      ##  switch context to GOAL_RE (goalline(s)
+      ##   split token (automagically) into two!! - player AND minute!!!
+      @re = GOAL_RE
+      puts "  ENTER GOAL_RE MODE"   if debug?
+
+      score = {}
+      ## must always have ft for now e.g. 1-1 or such
+      ###  change to (generic) score from ft -
+      ##     might be score a.e.t. or such - why? why not?
+      score[:ft] = [m[:ft1].to_i(10),
+                    m[:ft2].to_i(10)]  
+      ## note - for debugging keep (pass along) "literal" score
+      tokens << [:SCORE, [m[:score], score]]
+
+      ## auto-add player token 
+      tokens << [:PLAYER, m[:name]]
+  
+      offsets = [m.begin(0), m.end(0)]
+      pos = offsets[1]    ## update pos
+    end 
+
+    m = PLAYER_WITH_MINUTE_RE.match( line )
+    if m
+      ##  switch context to GOAL_RE (goalline(s)
+      ##   split token (automagically) into two!! - player AND minute!!!
+      @re = GOAL_RE
+      puts "  ENTER GOAL_RE MODE"   if debug?
+
+      ## check for optional open_bracket
+      tokens << [:'[']     if m[:open_bracket]
+
+      ## check for  -;  (none with separator)
+      ##    todo - find a better way? how possible?
+      tokens << [:NONE, "<|NONE|>"]   if m[:none]
+      
+
+
+      ## auto-add player token first
+      tokens << [:PLAYER, m[:name]]
+      ## minute props
+      minute = {}
+      minute[:m]      = m[:value].to_i(10)
+      minute[:offset] = m[:value2].to_i(10)   if m[:value2]
+      ##  t is minute only
+      tokens << [:MINUTE, [m[:minute], minute]]
+
+      offsets = [m.begin(0), m.end(0)]
+      pos = offsets[1]    ## update pos
+    end
+  end
+
+
+
   while m = @re.match( line, pos )
-#    if debug?
-#      pp m
-#      puts "pos: #{pos}"
-#    end
+    # if debug?
+    #  pp m
+    #  puts "pos: #{pos}"
+    # end
     offsets = [m.begin(0), m.end(0)]
 
     if offsets[0] != pos
@@ -389,16 +464,53 @@ def _tokenize_line( line )
              puts "!!! TOKENIZE ERROR (PROP_RE) - no match found"
              nil 
          end
+      elsif @re == GOAL_RE
+         if m[:space] || m[:spaces]
+              nil    ## skip space(s)
+         elsif m[:prop_name]    ## note - change prop_name to player
+             [:PLAYER, m[:name]] 
+         elsif m[:minute]
+              minute = {}
+              minute[:m]      = m[:value].to_i(10)
+              minute[:offset] = m[:value2].to_i(10)   if m[:value2]
+             ## note - for debugging keep (pass along) "literal" minute
+             [:MINUTE, [m[:minute], minute]]
+         elsif m[:score]
+              score = {}
+              ## must always have ft for now e.g. 1-1 or such
+              ###  change to (generic) score from ft -
+              ##     might be score a.e.t. or such - why? why not?
+              score[:ft] = [m[:ft1].to_i(10),
+                            m[:ft2].to_i(10)]  
+              ## note - for debugging keep (pass along) "literal" score
+              [:SCORE, [m[:score], score]]
+         elsif m[:og]
+             [:OG, m[:og]]    ## for typed drop - string version/variants ??  why? why not?
+         elsif m[:pen]
+             [:PEN, m[:pen]]
+         elsif m[:sym]
+            sym = m[:sym]
+            ## return symbols "inline" as is - why? why not?
+            ## (?<sym>[;,@|\[\]-])
+ 
+            case sym
+            when ',' then [:',']
+            when ';' then [:';']
+            when '[' then [:'[']
+            when ']' then [:']']
+            else
+              nil  ## ignore others (e.g. brackets [])
+            end
+         else
+            ## report error
+            puts "!!! TOKENIZE ERROR (GOAL_RE) - no match found"
+            nil 
+         end
       ###################################################
       ## assume TOP_LEVEL (a.k.a. RE) machinery
       else  
         if m[:space] || m[:spaces]
            nil   ## skip space(s)
-        elsif m[:prop_key]
-           ##  switch context  to PROP_RE
-           @re = PROP_RE
-           puts "  ENTER PROP_RE MODE"  if debug?
-           [:PROP, m[:key]]
         elsif m[:text]
           [:TEXT, m[:text]]   ## keep pos - why? why not?
         elsif m[:status]   ## (match) status e.g. cancelled, awarded, etc.
@@ -436,7 +548,9 @@ def _tokenize_line( line )
             date = {}
  ## map month names
  ## note - allow any/upcase JULY/JUL etc. thus ALWAYS downcase for lookup
-            date[:y] = m[:year].to_i(10)  if m[:year]
+            date[:y]  = m[:year].to_i(10)  if m[:year]
+            ## check - use y too for two-digit year or keep separate - why? why not?
+            date[:yy] = m[:yy].to_i(10)    if m[:yy]    ## two digit year (e.g. 25 or 78 etc.)
             date[:m] = m[:month].to_i(10)  if m[:month]
             date[:m] = MONTH_MAP[ m[:month_name].downcase ]   if m[:month_name]
             date[:d]  = m[:day].to_i(10)   if m[:day]
@@ -480,6 +594,8 @@ def _tokenize_line( line )
         elsif m[:score]
             score = {}
             ## must always have ft for now e.g. 1-1 or such
+            ###  change to (generic) score from ft -
+            ##     might be score a.e.t. or such - why? why not?
             score[:ft] = [m[:ft1].to_i(10),
                           m[:ft2].to_i(10)]  
           ## note - for debugging keep (pass along) "literal" score
@@ -490,10 +606,6 @@ def _tokenize_line( line )
               minute[:offset] = m[:value2].to_i(10)   if m[:value2]
              ## note - for debugging keep (pass along) "literal" minute
              [:MINUTE, [m[:minute], minute]]
-        elsif m[:og]
-           [:OG, m[:og]]    ## for typed drop - string version/variants ??  why? why not?
-        elsif m[:pen]
-           [:PEN, m[:pen]]
         elsif m[:vs]
            [:VS, m[:vs]]
         elsif m[:sym]
@@ -514,8 +626,13 @@ def _tokenize_line( line )
           when '---'  then [:'---']   # level 3
           when '----' then [:'----']  # level 4
           else
+            puts "!!! TOKENIZE ERROR (sym) - ignore sym >#{sym}<"
             nil  ## ignore others (e.g. brackets [])
           end
+        elsif m[:any]
+           ## todo/check log error
+           puts "!!! TOKENIZE ERROR (any) - no match found >#{m[:any]}<"
+           nil   
         else
           ## report error
            puts "!!! TOKENIZE ERROR - no match found"
@@ -560,6 +677,12 @@ def _tokenize_line( line )
         tokens << [:PROP_END, "<|PROP_END|>"]    
      end
    end
+
+
+   if @re == GOAL_RE   ### ALWAYS switch back to top level mode
+     puts "  LEAVE GOAL_RE MODE, BACK TO TOP_LEVEL/RE"  if debug?
+     @re = RE 
+   end
   
   [tokens,errors]
 end