hierarchical topic reduction improvements

• fixed loading bug
• hierarchical topic reduction bug
• added parameter for optimizing hierarchical reduction speed

Topic indexing bugfix

1.0.33

topic indexing bugfix

1.0.32

Indexing bugfix

gpu hdbscan and topic indexing

1. Added gpu hdsbcan
2. Added topic indexing

gpu umap

1. Changed default embedding model to universal-sentence-encoder-multilingual.
2. Added option for GPU umap with gpu_umap parameter.

Adding compute_topics

• Added a method for computing topics.
• Exposed topic deduplication parameter topic_merge_delta.
• Bug fixes.

Sklearn change in API fix

get_feature_names() -> get_feature_names_out()

Phrases and new embedding options

• New pre-trained transformer models available
• Ability to use any embedding model by passing callable to embedding_model
• New embedding_batch_size option
• Document chunking options for long documents
• Phrases in topics by setting ngram_vocab=True

Query documents and topics fix

1.0.26

updating version

Query documents and topics

Added query_documents and query_topics methods which allow for using a sequence of text such as a question, a sentence, a paragraph or a document to query documents or topics.

Added num_topics parameter to get_documents_topics method which allows retrieving multiple topics per document.