--language

Switch

--language {en,zh}

Description

Language of the corpus, used to select the right tokenizer/parser. Aliased as --lang.

Argument and Default Value

One of en or zh (argparse choices). Default: en (dlac.DEF_LANG).

Details

Any other value is rejected by argparse itself before DLATK even runs (error: argument --language/--lang: invalid choice). Affects tokenization in --add_tokenized/--add_ngrams and any other message-processing step that depends on language-specific rules.

Other Switches

Optional Switches:

--no_lang (unrelated — excludes language features from a prediction run, not a language selector)

Example Commands

dlatkInterface.py -d chinese_corpus -t msgs -c user_id --language zh --add_tokenized