Skip to main content

Semantic and hybrid search agent

Project description

Example of the agent using RAG

Depends on the just-agents package

This package also has some features which are specific for Just-Chat integration.

It allows easier index for markdown files.

  1. Indexing papers

For indexing what you have to do is run

poetry shell
index_markdown <path_to_markdown_folder> -i <index_name>

this will index all markdown files in the folder and create a Meilisearch index. It will all be indexed in the location where you gave the input files from. NOTE: this is run here, in the just-semantic-search folder but can have targets outside of this folder, for example in other projects.

  1. Searching indexed papers

First off- searching is mostly done in the project you are working on. Meaning that the primary usecase is for the user to import the libary.

2.1. Just-Agents have you should configure a web_agent in your agent_profiles.yaml. You can either user meilisearch separately from just-chat or you can extend the just-chat docker-compose.yml file with the following meilisearch service.

meilisearch:
    container_name: meilisearch
    image: getmeili/meilisearch:v1.13.0-rc.2 # we have to use it because migration got easier
    environment:
      - http_proxy
      - https_proxy
      - MEILI_MASTER_KEY=fancy_master_key
      - MEILI_NO_ANALYTICS=${MEILI_NO_ANALYTICS:-true}
      - MEILI_ENV=${MEILI_ENV:-development}
      - MEILI_LOG_LEVEL
      - MEILI_DB_PATH=${MEILI_DB_PATH:-/data.ms}
      - MEILI_EXPERIMENTAL_ENABLE_METRICS=true
      - MEILI_EXPERIMENTAL_ENABLE_VECTORS=true
    ports:
      - ${MEILI_PORT:-7700}:7700
    volumes:
      - ./data.ms:/data.ms
    restart: unless-stopped

2.2. in requirements.txt you have to add just-semantic-search-meili 2.3. in agent_profiles.yaml you have to add the following tools for the agent you want to use it

      - package: "just_semantic_search.meili.tools"
        function: "search_documents"

This will allow the agent to use the search_documents function from the just_semantic_search.meili.tools package. Also, given you have indexed the papers in this project, the only part from the libary you will use is the search_documents function.

2.4. run

 docker compose up 

and querry the agent so that it will have to search your indexed papers

NOTE: to check things before you run the agent you can first check port 0.0.0.0:7700 to see if the meilisearch is running. -key is fancy_master_key There you should be able to see whether meilisearch is running and if there are indexes created.

Following text explains more in details how this library works and it is structured


Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

just_semantic_search_agent-0.1.3.tar.gz (7.6 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

just_semantic_search_agent-0.1.3-py3-none-any.whl (7.4 kB view details)

Uploaded Python 3

File details

Details for the file just_semantic_search_agent-0.1.3.tar.gz.

File metadata

  • Download URL: just_semantic_search_agent-0.1.3.tar.gz
  • Upload date:
  • Size: 7.6 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/2.1.0 CPython/3.11.10 Linux/6.8.0-53-lowlatency

File hashes

Hashes for just_semantic_search_agent-0.1.3.tar.gz
Algorithm Hash digest
SHA256 ff318f1ad9d05d0e5fd5f823148cd493355a6ebcfb9fe3e08cc624193df96fea
MD5 a06888309bf3d456f7efa573a21b977a
BLAKE2b-256 d006877b73e15a58692efb8e19f152b0bf0bfabbac63209377b916c19fc5890b

See more details on using hashes here.

File details

Details for the file just_semantic_search_agent-0.1.3-py3-none-any.whl.

File metadata

File hashes

Hashes for just_semantic_search_agent-0.1.3-py3-none-any.whl
Algorithm Hash digest
SHA256 a2f790d4b55d788c60d2182fa14fa9df93d7084b799f863d0973877ff71bfa09
MD5 c41d82c49b627cf9463e3e1689b6d640
BLAKE2b-256 fc40a8f351922313643f88b8b82220a589955ea36014be7595086761b586e6af

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page