RedisFI VSS is a data loader/microservice to facilitate the VSS part of RedisFI. It is a search engine on top of a collection of SEC 10K/Q documents that were vectorized by Microsoft's Enterprise Data Science team (who has graciously allowed us to share them with this project) using all-mpnet-base-v2.
Requires Python 3.9 (not above) and Poetry
Once Poetry is installed, setting up redisfi-vss is straight forward. Inside the directory run:
poetry install
Once the dependancies are installed, you should be able to run the CLI by typing:
poetry run VSS
Which will print a menu of the different options available. The main two are load - which will load the prepared data into Redis Enterprise - and run, which will run the microservice to make searching the data possible.
The data set is very large and requires a Redis Cluster to fully work. For testing, it's recommended to only load one of the data files.