Microsoft Hackaton Activity of December 2023
-
BITS AI integration
- we have access to API and can have subsets of data (or custom API calls created)
-
-
Archibus AI integration
- Limited API integration(?)
- Could potentially have access to subsets of data(?)
- Azure OpenAI Chatbot (using cognitive search)
- Azure AI Studio Examples
- Azure OpenAI Accelerators and Demo Assets
- LLM Grounding
- (Video) Quickly build and deploy OpenAI apps on Azure with your own data
- using function calling with Azure OpenAI
Retrival-Augmented Generation (RAG):
python util/test.py
{
"id": "ef8095cf-23b3-4e46-9a2d-bca6177bbe4f",
"model": "gpt-4-32k",
"created": 1701870043,
"object": "extensions.chat.completion",
"choices": [
{
"index": 0,
"finish_reason": "stop",
"message": {
"role": "assistant",
"content": "From the retrieved documents, the following client names can be identified:\n\n1. Emploi et D\u00e9veloppement social Canada[doc1]\n2. Anciens Combattants Canada[doc2][doc3]\n3. Affaires mondiales Canada[doc4]\n4. Services aux Autochtones Canada[doc5]\n\nPlease note that only four unique client names are available in the retrieved documents.",
"end_turn": true,
"context": {
...Drop the payload into the data/ folder and then just run from the root of this project
python util/split _payload.pyResulting files will end up in output/ folder.
python3 -m venv .venv
source .venv/bin/activate
pip install -r util/requirements.txt --upgradeThen you can run the app via:
flask --debug run --port 5002TEST https://geds20api-sage20api.itsso.gc.ca/gapi/v2/doc
PRODUCTION https://gcdirectory-gcannuaire.ssc-spc.gc.ca/gapi/v2/doc
If your application is in the public domain (not GC internal network) you can access production from here:
https://api.geds-sage.gc.ca/gapi/v2
Same applies for the documentation:
https://api.geds-sage.gc.ca/gapi/v2/doc
However, you will need to change the URL for the swagger.json to :