vanna

vanna copied to clipboard

Query security

2

Any interest in integrating a SQL static analysis tool to make queries secure? https://github.com/amoffat/HeimdaLLM

amoffat

[Feature Request] Ability to Evaluate Answers as Correct or Incorrect from Python Modules

4

I'm using Vanna in my Python application, and I've found that there's currently no way to evaluate the correctness of the AI's answers directly from the Python modules. This makes...

ryandward

Max context length exceeded in __rpc_call(method="generate_questions")

1

```{'error': {'code': -32603, 'message': 'BadRequestError(\'Error code: 400 - {\\\'error\\\': {\\\'message\\\': "This model\\\'s maximum context length is 16385 tokens. However, you requested 16583 tokens (16083 in the messages, 500 in the...

bhuvanrawal

Edge cases for ask() when df has length 1 and print_result = False

1

src/vanna/__init__.py Line 1525: `return sql, df, fig, None` When df has length 1 and print_result = False, fig is not yet defined. Thus, it will return the error: Couldn't run...

danendral

One bad query in connect_to_postgres results in future failures

1

If one query has an error, the subsequent queries to postgres will fail. We either need to open a new connection for each query or we need to do a...

zainhoda

Building UI

5

This is really stunning! Just it's a pity that we can see the results only on the Jupyter notebook. Can we build a lightweight clean frontend which consumes those apis...

v0nerd

PGVector support

8

For larger / more diverse query histories would be nice to externalize the Vector Store to something like PostGres w/ the vector extension

dcieslak19973

Since most llm use-cases are conversational, having memory in vanna would be considerably advantageous

aparnakesarkar

Add qdrant as a vectorDB

aparnakesarkar

aparnakesarkar