Pamela Fox
Pamela Fox
Addressing this question: https://github.com/pamelafox/rag-on-postgres/issues/12
The current one we're using likely is bigger than necessary.
Entra token connection strings expire every 24, we should refresh every 6. Example: https://github.com/microsoft/azure-openai-service-proxy/blob/7f9f035c79fefd371b1e6e395b2628363c28b53c/src/proxy/app/db_manager.py#L130
It's one of the top 3 popular Python web frameworks this day and is async only, so it's important for applications that make long network calls. Would love built-in support...
- **Package Version**: See above - **Operating System**: Mac M1, with Debian Dev Container - **Python Version**: Python 3.10 **Describe the bug** When using evaluate(), I get these deprecation warnings:...
## Purpose The container apps PR accidentally changed two roles that didnt need to change, the roles for the search service principal ID. These roles are needed for integrated vectorization...
## Purpose Tests the evaluation workflow ## Does this introduce a breaking change? When developers merge from main and run the server, azd up, or azd deploy, will this produce...
From Emmanuel: ``` GRANT pg_read_all_data TO YOURACCOUNT; GRANT pg_write_all_data TO YOURACCOUNT; enable the Azure AI extension [https://learn.microsoft.com/en-gb/azure/postgresql/flexible-server/generative-ai-azure-overview#enable-the-azure_ai-extension](https://nam06.safelinks.protection.outlook.com/?url=https%3A%2F%2Flearn.microsoft.com%2Fen-gb%2Fazure%2Fpostgresql%2Fflexible-server%2Fgenerative-ai-azure-overview%23enable-the-azure_ai-extension&data=05%7C02%7Cedeletang%40microsoft.com%7C32817772f3724712c73008dcd6753f4d%7C72f988bf86f141af91ab2d7cd011db47%7C1%7C0%7C638621046309783323%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C0%7C%7C%7C&sdata=uSpg0f0yoTCtd9X7u0jOcvf7tawQqng%2FDXHq%2Fvg%2FprQ%3D&reserved=0) CREATE EXTENSION azure_ai; CREATE EXTENSION pgvector; connect to your Open AI enpoint select...
### Description Add a question caching feature. Pseudocode based on other repo- ``` def chat_completion(user_input): # Generate embeddings from the user input user_embeddings = generate_embeddings(openai_client, user_input) # Query the chat...
May need to configure key vault access specially.