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.
- 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
andkey
for your corupus here. In our case it isfetchai-docs-explorer
. The corpus key will be used to make API calls to the corpus. - Get the
API_KEY
from Authorisation in menu.
- 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
- 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. - 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
andCORPUS_KEY
obtained from your vectara console.
- 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 .
- 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.
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.