collective/mediawiki_dump_tools
18
0
Fork 0
Code Issues Pull Requests Activity

make wikiq memory efficient again via batch processing.

Browse Source
This commit is contained in:
Nathan TeBlunthuis
2025-07-15 15:20:17 -07:00
parent 76d54ae597
commit c40506137b
6 changed files with 1809 additions and 1605 deletions
Download Patch File Download Diff File Expand all files Collapse all files

773
src/wikiq/__init__.py
View File

File diff suppressed because it is too large Load Diff

11
src/wikiq/tables.py
View File

@@ -63,7 +63,7 @@ class RevisionTable:
return pa.schema([c.field for c in self.columns])
def pop(self) -> dict:
data = {}
data = dict()
for column in self.columns:
data[column.field.name] = column.pop()
@@ -174,15 +174,6 @@ class RevisionTextChars(RevisionField[Union[int, None]]):
return None
class RevisionText(RevisionField[str]):
field = pa.field("text", pa.string())
def extract(self, page: mwtypes.Page, revisions: list[mwxml.Revision]) -> str:
revision = revisions[-1]
return revision.text
class RevisionIsMinor(RevisionField[bool]):
field = pa.field("minor", pa.bool_())

62
src/wikiq/wiki_diff_matcher.py
View File

@@ -96,48 +96,6 @@ class DiffToOperationMap:
del self.from_par_move_dict[rkey]
break
# if len(self.from_par_move_dict) > 0 or len(self.to_par_move_dict) > 0:
# print("Couldn't find exact matches for all parmoves!")
# # we couldn't find all the matches via exact match
# # let's try matching based on line number instead
# lkeys_to_remove = []
# for lkey, from_diff in self.from_par_move_dict.items():
# from_linenum = from_diff["moveInfo"]["linkId"].split("_")[2]
# rkey_to_remove = None
# for rkey, to_diff in self.to_par_move_dict.items():
# to_linenum = rkey.split("_")[2]
# if from_linenum == to_linenum:
# print("Matching on line number")
# yield from self.doParMove(from_diff, to_diff)
# rkey_to_remove = rkey
# lkeys_to_remove.append(lkey)
# break
# if rkey_to_remove is not None:
# del self.to_par_move_dict[rkey_to_remove]
# for lkey in lkeys_to_remove:
# del self.from_par_move_dict[lkey]
# if len(self.from_par_move_dict) > 0 or len(self.to_par_move_dict) > 0:
# print("Couldn't find exact matches for all parmoves!")
# # we couldn't find all the matches via exact match or line number
# # let's try matching based on opIndex instead
# lkeys_to_remove = []
# for lkey, from_diff in self.from_par_move_dict.items():
# rkey_to_remove = None
# from_idx = from_diff["moveInfo"]["linkId"].split("_")[1]
# for rkey, to_diff in self.to_par_move_dict.items():
# to_idx = rkey.split("_")[1]
# print(from_idx)
# print(to_idx)
# if from_idx == to_idx:
# yield from self.doParMove(from_diff, to_diff)
# rkey_to_remove = rkey
# lkeys_to_remove.append(lkey)
# if rkey_to_remove is not None:
# del self.to_par_move_dict[rkey_to_remove]
# for lkey in lkeys_to_remove:
# del self.from_par_move_dict[lkey]
# we couldn't find matches. treat type 4 as removal and type 5 as highlight.
for from_diff in self.from_par_move_dict.values():
yield from self.doDelete(from_diff)
@@ -368,22 +326,21 @@ class DiffToOperationMap:
class WikiDiffMatcher:
def __init__(
self,
texts: list[str] = None,
tokenizer: Optional[RegexTokenizer] = None,
):
differ = pywikidiff2.pywikidiff2(
numContextLines=1000000, moved_paragraph_detection_cutoff=200000
)
# Pre-compute diffs to reduce traffic overhead.
self.diffs = [json.loads(diff) for diff in differ.inline_json_diff_sequence(list(texts))]
self.tokenizer = tokenizer or TOKENIZER
class Processor(DiffEngine.Processor):
def __init__(self, texts, tokenizer=None):
self.diffs = iter(texts)
def __init__(self, tokenizer=None):
self.tokenizer = tokenizer or TOKENIZER
self.last_tokens = []
self.previous_text = ""
self.differ = pywikidiff2.pywikidiff2(
numContextLines=1000000, moved_paragraph_detection_cutoff=200000
)
self.last_diff = None
def update(self, last_tokens):
self.last_tokens = last_tokens
@@ -391,7 +348,8 @@ class WikiDiffMatcher:
def process(self, text, token_class=None):
# The diff has already been computed, but we need to incrementally
# retrieve it to recreate the behavior DiffState expects.
diff = next(self.diffs)
diff = json.loads(self.differ.inline_json_diff(self.previous_text, text))
self.last_diff = diff
diffToOperationsMapper = DiffToOperationMap(diff, self.tokenizer)
diffops = list(diffToOperationsMapper.to_operations())
@@ -444,7 +402,7 @@ class WikiDiffMatcher:
return border_ops, self.last_tokens, tokens
def processor(self, *args, **kwargs):
return self.Processor(self.diffs, self.tokenizer)
return self.Processor(self.tokenizer)
def process(self):
# DiffState checks for this method even though it is not called.