Changes to make search work better

1. Move back to llm qa "stuff" mode as map_reduce is clueless
2. Created more metadata around documents.
* allow create different collection spaces for db.
/v1/admin/db -d "name={collection}"
* within collection allow file names so that answers can be annotated
{content: {answer :"I don't know", files:["a1","a2"]}}
would mean openAI doesn't know about the question even though
embedding returned two files.

Lot more work to do

api/routes/admin.py

@@ -1,13 +1,15 @@
 #This is to init the vector store
 
+from typing import Annotated
+
 from qdrant_client.models import VectorParams, Distance
-from fastapi import APIRouter
+from fastapi import APIRouter, Body
 from db import vector_store
 
 router = APIRouter()
 
 @router.put("/admin/v1/db")
-async def recreate_collection(name: str = "test"):
+async def recreate_collection(name: Annotated[str, Body(embed=True)]):
     print(f"creating collection {name} in db")
     return vector_store.client.recreate_collection(collection_name=name, 
                                             vectors_config=VectorParams(size=1536, distance=Distance.COSINE))

api/routes/search.py

@@ -1,4 +1,6 @@
-from fastapi import APIRouter, UploadFile, File
+from typing import Annotated
+
+from fastapi import APIRouter, UploadFile, File, Body
 from fastapi.responses import JSONResponse
 import openai
 import io
@@ -12,29 +14,33 @@ from langchain.llms import OpenAI
 from db import vector_store
 
 router = APIRouter()
-_db = vector_store.get_instance()
-_chain = load_qa_chain(OpenAI(temperature=0), chain_type="map_reduce")
+_chain = load_qa_chain(OpenAI(temperature=0), chain_type="stuff")
 
 @router.post("/v1/docs")
-async def index_doc(file: UploadFile = File(...)):
-    async for doc in generate_documents(file):
+async def index_doc(name: Annotated[str, Body()], fileName: Annotated[str, Body()], file: UploadFile = File(...)):
+    _db = vector_store.get_instance(name)
+    if not _db:
+        return JSONResponse(status_code=404, content={})
+    async for doc in generate_documents(file, fileName):
+        print(doc)
         _db.add_documents([doc])
     #todo return something sensible
-    return JSONResponse(status_code=200, content={})
+    return JSONResponse(status_code=200, content={"name": name})
 
-@router.get("/v1/docs")
-async def search(query: str):
+@router.get("/v1/answers/{name}")
+async def search(name: str, query: str):
+    _db = vector_store.get_instance(name)
     print(query)
     docs = _db.similarity_search(query=query)
     print(docs)
     answer = _chain.run(input_documents=docs, question=query)
-    return JSONResponse(status_code=200, content={"answer": answer}) 
+    return JSONResponse(status_code=200, content={"answer": answer, "files": [d.metadata["file"] for d in docs]}) 
 
-async def generate_documents(file: UploadFile):
+async def generate_documents(file: UploadFile, fileName: str):
     num=0
     async for txt in convert_documents(file):
         num += 1
-        document = Document(page_content=txt,metadata={"page": num})
+        document = Document(page_content=txt,metadata={"file": fileName, "page": num})
         yield document
  
 async def convert_documents(file: UploadFile):