vectara UAgents RAG

This RAG is created using vectara API and wrapped using fetchs' uAgents. It is deployed Agentverse. This answers all your queries related to fetch uagents.

Creating Account on Vectara and creating corpus

• Create an account on Vectara.
• Goto the Console and create a new corpus for yourself using the Create Corpus button.
• Decide the type of application later and give name and key for your corupus here. In our case it is fetchai-docs-explorer. The corpus key will be used to make API calls to the corpus.
• Get the API_KEY from Authorisation in menu.

Adding docs to corpus which will be used as RAG knowledge.

• Goto your corpus and click data tab.
• Click on and add data by clicking on button.
• You can upload text, HTML, PDF, Word and other files here.

Other resources:

• Vectara API docs
• OSS tool to help crawl and ingest data from other sources - vectara-ingest

Creating Agentverse Agent

• Goto Agentverse and login using google.
• Create a news blank agent from here.
• copy the code in agent.py and save in agent.py file. If you want to make just a single query to the agent and not chitchat use agents_single_query.py file instead.
• Add a new python script and name it as helper_functions.py. Copy the script from here and save into that script.
• Add secrets to your agent by hitting button and save the API_KEY and CORPUS_KEY obtained from your vectara console.

creating Agentverse Function

• Go to Deploy tab of your agent and click on button.
• Enter the details on functions as shown in below image.
• Rest details are self populated and click button.
• Once your agent is create click on .

Using agent on deltaV

Watch the video on YouTube

• Use the below video for reference on how to use this on deltaV.

If you want to access this agent go to DeltaV and select Public Function Group.

NOTE :

IF YOU MAKE ANY CHANGES TO THE DATA MODELS PLEASE DO REMEMBER TO GO BACK TO FUNCTIONS AND UPDATE THE FUNCTION EVEN IF THERE ARE NO CHANGES.