Globally, setting DEBUG=1 will set the logging level to DEBUG.
You can choose the exact Wiktionary dump to download by using FORCE_SNAPSHOT=YYYYMMDD.
Setting NO_COLORS=1 will remove all markup styling (italic, bold, etc.).
When an error happens in transforming/expanding a definition, the definition will be skipped (not the entire word, just the definition).
Re-run the command prepending KEEP_UNFINISHED=1 to display the raw HTML, and be able to see where the issue comes from.
Lst all words not taken into account with current head sections:
DEBUG_PARSE=1 python -m wikidict LOCALE --parse >out.logLst all unhandled sections:
DEBUG_SECTIONS=1 python -m wikidict LOCALE --render | sort -u >out.logMake words under a given section to fail:
DEBUG_SECTIONS="<section>" python -m wikidict LOCALE --renderExample with the RO dictionary, and the "{{unități}}" section:
DEBUG_SECTIONS='{{unități}}' python -m wikidict ro --renderList all unhandled words:
DEBUG_EMPTY_WORDS=1 python -m wikidict LOCALE --render >out.log 2>&1Useful to debug Lua expansion issues.
For example, to log all words for each process in order to be able to catch problematic words in a second time (mostly to catch infinite loops):
DEBUG_LUA=1 python -m wikidict LOCALE --render > LOG_FILE 2>&1
tail -f LOG_FILE
# (and when the ouput hangs, hit CTRL+C, multiple times if needed)
python log-analyzer.py LOG_FILEIf more details are needed, use DEBUG_LUA=2 to print Lua errors in real time.
This is useful to list all found part of speech (POS). To be used after --render to have the full dictionary ready to be analyzed.