yujunjun/unstructured
1
0
Fork 0
mirror of https://github.com/Unstructured-IO/unstructured.git synced 2025-10-31 01:54:25 +00:00
Code Issues Packages Projects Releases Wiki Activity
unstructured/examples/pgvector
History
Matt Robinson e052c2a9b2
docs: example of how to use unstructured with pgvector (#571) 
* pgvector requirements

* first pass on pgvector notebook and sql alchemy file

* created code for loading vectors into db

* added query for embedding distance

* updates to pgvector notebook

* update function with time decay

* update pgvector notebook to use example code

* remove old create table script

* add readme for pgvector

* update example to use get_date()
2023-05-12 13:54:38 -04:00
..
pgvector.ipynb
docs: example of how to use unstructured with pgvector (#571)
2023-05-12 13:54:38 -04:00
README.md
docs: example of how to use unstructured with pgvector (#571)
2023-05-12 13:54:38 -04:00
requirements.txt
docs: example of how to use unstructured with pgvector (#571)
2023-05-12 13:54:38 -04:00

README.md

Loading unstructured outputs into Postgres with pgvector

The following example shows how to load unstructured output into Postgres with the pgvector extension installed. Combining the similarity search functionality of pgvector with the traditional RDBMS capabilities of Postgres allow users to performing similarity searches that are conditioned on metadata or biased toward more recent documents. Use cases include document discovery and more sophisticated retrieval augmented generation for LLMs. The langchain docs have more information about retrieval augmented generation.

Running the example

  1. Install Postgres.
  2. Install pgvector
  3. Run pip install -r requirements.txt to install the Python dependencies.
  4. Run `jupyter-notebook to start.
  5. Run the pgvector.ipynb notebook.