Now accepts .zip file as inputs. Moved semantic search option bar. Minor API mode changes.

Changed embedding model to MiniLM-L6 as faster. Compressed embeddings are now int8. General improvements to API mode

General code improvements and refinements.

Set bm25 in functions explicitly. Some API updates. Now can get connection params on startup.

Some package updates and minor changes

Changed all intermediate file outputs to save to output folder

Allowed for custom output folder, returned Dockerfile to work under user account and port 7860

Added Lambda web adapter to Dockerfile. Updated Gradio version

Now checks for output folder before saving. Minor code cleaning

Fixed cleaning for semantic search. Handles text with backslashes in (if cleaned). Updated packages. requirements file for only keyword search added.

Assigned AWS bucket name to environmental variable

Added additional password auth for AWS-based files. Changed 'Clean' default to no

Gradio 4.21. Limitations on file size and creating embeddings. Added AWS integration

Updated to latest Gradio version that hopefully solved root_url issues. Added ssl_verify=False to launch

Set SDK version to 4.20.0 on huggingface. Added GRADIO_SERVER_PORT to Dockerfile

set gradio version to 4.18.0, before root path issue reappears

ECS not allowing me to save files so increasing container privileges in Dockerfile

Now loads in embedding model locally in Dockerfile

Improved code for cleaning and outputting files. Added Dockerfile

Improved xlsx output formatting. Deals better with cleaning data then analysing in same session.

Added highlight search term functionality to keyword search output

Updated to Gradio 4.16.0. Now works correctly with BGE embeddings

Upgraded to Gradio 4.16.0. Added Spacy fuzzy search functionality.

Changed intro text

Cut out semantic search temporarily while issues with Jina gated model resolved. Improved error/progress tracking and messaging. Placeholder for Spacy fuzzy search.

Better error checking. Doesn't load in embeddings file twice now.

Fixed data input for semantic search. Allowed for docs to be loaded in directly for semantic search. 0.2.1

Minor changes to file path for outputs, documentation, location of pyinstaller build dependencies

Many changes to code organisation. More efficient searches from using intermediate outputs. Version 0.1

Now works correctly with npz. Minor formatting improvements

Changed in_query to keyword_query

Added semantic search using Jina

Faster embedding with GPU, fast document split, writes to chromadb file correctly. No longer needs FAISS or langchain

Now outputs correct dataframe for semantic search. Can join on extra details

Added basic semantic search functionality

added nltk punkt load

Added stopwords and wordnet nltk dependencies

Added nltk download for names for HF use

Updates to readme file. Changed app file name to work with HF.