-m

Switch

-m TABLE

Description

Name of the LDA-labeled message table to create or read. Aliased as --lda_msg_tbl.

Argument and Default Value

A table name. Default: "messages_en_lda$msgs_en_tok_a30" (dlac.DEF_LDA_MSG_TABLE) — almost always overridden in practice.

Details

Verified end-to-end with real Mallet, via --estimate_lda_topics with --lda_lexicon_name set: the table actually created (by MessageTransformer.addLDAMessages(), triggered internally rather than via -m directly) followed the pattern <corptable>_lda$<lda_lexicon_name> — e.g. msgs_lda$testtopics — not the literal -m value in that flow. -m/--lda_msg_tbl is instead used to point TopicExtractor at an existing LDA message table for downstream steps that consume one.

Other Switches

Optional Switches:

--estimate_lda_topics, --lda_lexicon_name

Example Commands

Using the packaged dla_tutorial dataset (see DLATK LDA Interface):

dlatkInterface.py -d dla_tutorial -t msgs -c user_id -f 'feat$1gram$msgs$user_id' \
--estimate_lda_topics --num_topics 100 --lda_lexicon_name dla_topics \
--mallet_path /opt/homebrew/bin/mallet --save_lda_files ./lda_out
# creates msgs_lda$dla_topics, which subsequent commands can point at with -m msgs_lda$dla_topics