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

add some debug lines.

Browse Source
This commit is contained in:
Nathan TeBlunthuis
2026-01-06 19:58:18 -08:00
parent 8590e5f920
commit 4b8288c016
2 changed files with 2 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
src/wikiq/__init__.py
View File

@@ -854,6 +854,7 @@ class WikiqParser:
else: else:
jsonl_basename = "data.jsonl" jsonl_basename = "data.jsonl"
jsonl_path = Path(self.output_file) / jsonl_basename jsonl_path = Path(self.output_file) / jsonl_basename
print(f"Writing to JSONL: {jsonl_path} (append={append_mode})", file=sys.stderr)
writer = JSONLWriter(str(jsonl_path), schema, append=append_mode) writer = JSONLWriter(str(jsonl_path), schema, append=append_mode)
else: else:
writer = JSONLWriter(self.output_file, schema, append=append_mode) writer = JSONLWriter(self.output_file, schema, append=append_mode)

1
src/wikiq/resume.py
View File

@@ -168,6 +168,7 @@ def get_jsonl_resume_point(output_file, input_file=None):
from wikiq import get_output_filename from wikiq import get_output_filename
jsonl_filename = os.path.basename(get_output_filename(input_file, 'jsonl')) jsonl_filename = os.path.basename(get_output_filename(input_file, 'jsonl'))
output_file = os.path.join(output_file, jsonl_filename) output_file = os.path.join(output_file, jsonl_filename)
print(f"Looking for resume point in: {output_file}", file=sys.stderr)
else: else:
return None return None