Getting the document topic with the highest probability

[FabianUlrich96]

Hello,
I am trying to get the topic with the highest probability for each document (article name).

Using the tmplot report with the trained model and a list of the article names I get results that seem logical in the plot under each topic.
tmp.report(model=model, docs=article_names)

article_names structure: ['article name 1', 'article name 2']

However, when I get the documents vs topics probabilities matrix I get a matrix with the right shape but without the article names. Matching the topic with the highest probability from the matrix with the article_name list gives me a different result than the tmp.report one.

topics_docs = titles_model.matrix_docs_topics_
topics_docs_df = pd.DataFrame(topics_docs)

How do I get a dataframe with the following shape?
Article name, topic with the highest probability
article 1, topic 1
article 2, topic 3
article 3, topic 1

I hope my question is clear.
Thanks in advance for the help!

[maximtrp]

Hello! You can try the latest release. It adds labels_ property and get_docs_top_topic method. There is also an example in README.md file.

[FabianUlrich96]

Thanks for the quick reply and fix!
I now used:
btm.get_docs_top_topic(article_names, titles_model.matrix_docs_topics_)
Receiving results matching the plot.