json_mend 0.3.3 → 0.3.5

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: ae2e6bb67d6227598226e35e852e4b4ae7f681e6ce8bcb94d11bd5e61c74a989
-  data.tar.gz: 0fbd0b9cc00caa626a794e2024d34d0f31d9bf407f31e7ce8cbb0e1b78c9773d
+  metadata.gz: 7dcc378c148dc0514753b966693e0afe286bf145c283d069f01c0b9c74f59a75
+  data.tar.gz: 37fa52116d57bc80b168c20feba1fd7b33edab73d1f157759bc71f4a3362802b
 SHA512:
-  metadata.gz: afd9daaec420a7bdb29ed2bf5d6438e5b335d2990eeae24503ab0fc10e6221a299dbcf63f4d0b09e093a7a03a523f5fa90db5ff46fc7cbec40670cac71f4955a
-  data.tar.gz: 730768fa1d1954c626c88c17b8afc3ec7e4726abcd2b4602072453e35d910da0cd6e1b92c5841e571abcd22c6bc7e02df65de7b8dfab45e9cd1210de24ca6f94
+  metadata.gz: c94ff3c9d0c2e3602b6a125e74fdc72896709b311221eeceeecd3fbf96b0aef8dc82edfed136512443b5ccc403d512d7beb3476c59732f1a04280a71bbf7de5b
+  data.tar.gz: f3603aa1d247686329ec0effe5d1819f4feec47cb43adbff520ed7695831cc30649fa99b4e21a51d4e460754af3d825480335cea10ffbb79feea773fabfa1cd9

data/lib/json_mend/parser.rb

@@ -44,6 +44,11 @@ module JsonMend
     # Pre-compile regexes for performance
     NUMBER_REGEX = /[#{Regexp.escape(NUMBER_CHARS.to_a.join)}]+/
     NUMBER_NO_COMMA_REGEX = /[#{Regexp.escape(NUMBER_CHARS.dup.tap { |s| s.delete(',') }.to_a.join)}]+/
+    INVALID_NUMBER_TRAILERS_REGEX = /[#{Regexp.union(*INVALID_NUMBER_TRAILERS)}]+\z/
+    HEX_ESCAPE_REGEXES = {
+      'u' => /[0-9a-fA-F]{4}/,
+      'x' => /[0-9a-fA-F]{2}/
+    }.freeze
 
     def initialize(json_string)
       @scanner = StringScanner.new(json_string)
@@ -72,7 +77,7 @@ module JsonMend
             # Ignore strings that look like closing braces garbage (e.g. "}", " ] ")
             next if new_json.is_a?(String) && new_json.match?(/\A\s*[}\]]+\s*\z/)
 
-            if both_hash?(json.last, new_json)
+            if json.last.is_a?(Hash) && new_json.is_a?(Hash)
               json[-1] = deep_merge_hashes(json.last, new_json)
             else
               json << new_json
@@ -390,20 +395,14 @@ module JsonMend
           # Handle JSON_STOP_TOKEN from parse_json (EOS or consumed terminator)
           if value == JSON_STOP_TOKEN
             # Do nothing, just skipped garbage
-          elsif strictly_empty?(value)
-            # Only consume if we didn't just hit a terminator that parse_json successfully respected
-            @scanner.getch unless value.nil? && TERMINATORS_ARRAY.include?(peek_char)
           elsif value == '...' && @scanner.string.getbyte(@scanner.pos - 1) == 46
             # just skip if the previous byte was a dot (46)
           else
             arr << value
           end
 
+          @scanner.skip(/[\s,]+/)
           char = peek_char
-          while char && char != ']' && (char.match?(/\s/) || char == ',')
-            @scanner.getch
-            char = peek_char
-          end
         end
 
         # Handle a potentially missing closing bracket, a common LLM error.
@@ -922,13 +921,7 @@ module JsonMend
           rstring_delimiter_missing = false
         elsif peek_char(j)
           # Check for an unmatched opening brace in string_parts
-          string_parts.reverse_each do |c|
-            next unless c == '{'
-
-            # Ok then this is part of the string
-            rstring_delimiter_missing = false
-            break
-          end
+          rstring_delimiter_missing = false if string_parts.include?('{')
         end
 
       end
@@ -962,21 +955,12 @@ module JsonMend
           entry_pos = @scanner.pos
           @scanner.getch # consume 'u' or 'x'
 
-          num_chars = (char == 'u' ? 4 : 2)
-          hex_parts = []
-
-          # Use getch in loop to correctly extract chars (handling multibyte)
-          num_chars.times do
-            c = @scanner.getch
-            break unless c
-
-            hex_parts << c
-          end
+          hex_regex = HEX_ESCAPE_REGEXES.fetch(char)
 
           # Validate valid hex digits
-          if hex_parts.length == num_chars && hex_parts.all? { |c| c.match?(/[0-9a-fA-F]/) }
+          if (hex_str = @scanner.scan(hex_regex))
             string_parts.pop
-            hex_val = hex_parts.join.to_i(16)
+            hex_val = hex_str.to_i(16)
 
             if char == 'u' && hex_val.between?(0xD800, 0xDBFF)
               # Handle high surrogate pair
@@ -1111,18 +1095,21 @@ module JsonMend
       # Save the original length so we can safely roll back if it's completely invalid
       original_length = scanned_str.bytesize
 
-      # Handle cases where the number ends with an invalid character.
-      if !scanned_str.empty? && INVALID_NUMBER_TRAILERS.include?(scanned_str[-1])
-        # Do not rewind scanner, simply discard the invalid trailing char (garbage)
-        scanned_str = scanned_str[0...-1]
       # Handle cases where what looked like a number is actually a string.
-      # e.g. "123-abc"
-      elsif peek_char&.match?(/\p{L}/)
+      # e.g. "123-abc" or "-Infinity". We exclude strings ending in a comma
+      # to preserve comma recovery logic (e.g. `105,next_key`).
+      if peek_char&.match?(/\p{L}/) && !scanned_str.end_with?(',')
         # Roll back the entire scan and re-parse as a string.
         @scanner.pos -= original_length
         return parse_string
       end
 
+      # Handle cases where the number ends with one or more invalid characters.
+      if !scanned_str.empty? && scanned_str.match?(INVALID_NUMBER_TRAILERS_REGEX)
+        # Do not rewind scanner, simply discard the invalid trailing chars (garbage)
+        scanned_str.sub!(INVALID_NUMBER_TRAILERS_REGEX, '')
+      end
+
       # Reject non-numbers (e.g., stray periods "." or dashes "-" from LLM conversational text)
       unless scanned_str.match?(/\d/)
         @scanner.pos -= original_length
@@ -1306,20 +1293,6 @@ module JsonMend
       res
     end
 
-    def both_hash?(obj1, obj2)
-      obj1.is_a?(Hash) && obj2.is_a?(Hash)
-    end
-
-    def strictly_empty?(value)
-      # Check if the value is a container AND if it's empty.
-      case value
-      when String, Array, Hash, Set
-        value.empty?
-      else
-        false
-      end
-    end
-
     # Skips whitespaces
     def skip_whitespaces
       @scanner.skip(/\s+/)

data/lib/json_mend/version.rb

@@ -1,5 +1,5 @@
 # frozen_string_literal: true
 
 module JsonMend
-  VERSION = '0.3.3'
+  VERSION = '0.3.5'
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: json_mend
 version: !ruby/object:Gem::Version
-  version: 0.3.3
+  version: 0.3.5
 platform: ruby
 authors:
 - Oleksii Vasyliev