add (optional) diff and text columns to output.

wiki_diff_matcher.py

@@ -16,10 +16,10 @@ import pywikidiff2
 class DiffToOperationMap:
     def __init__(self, diff, tokenizer):
         self.tokenizer = tokenizer
-        self.diff = json.loads(diff)
         self.from_par_move_dict = {}
         self.to_par_move_dict = {}
         self.highlights_without_offset = []
+        self.diff = diff
         # we need to keep track of the bytes of line numbers to recover when wikidiff2 loses offsets.
         self.to_linenumber_bytes_map: SortedDict[int, int] = SortedDict()
         self.from_linenumber_bytes_map: SortedDict[int, int] = SortedDict()
@@ -375,7 +375,7 @@ class WikiDiffMatcher:
             numContextLines=1000000, moved_paragraph_detection_cutoff=200000
         )
         # Pre-compute diffs to reduce traffic overhead.
-        self.diffs = differ.inline_json_diff_sequence(list(texts))
+        self.diffs = [json.loads(diff) for diff in differ.inline_json_diff_sequence(list(texts))]
         self.tokenizer = tokenizer or TOKENIZER
 
     class Processor(DiffEngine.Processor):