first cut working version

.gitignore

@@ -0,0 +1,36 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# Logs and databases
+*.log
+*.sqlite3
+*.db
+
+# Testing directories
+/.pytest_cache/
+/nose-timer-master/
+
+# Environment and compiled files
+.env
+.env.local
+.env.*.local
+*.pyc
+*.swp
+*.bak
+
+# IDE files
+.vscode/
+.idea/
+*.iml
+
+# Distribution / packaging
+/dist/
+/build/
+*.egg-info/
+
+# Project specific
+/config.yml
+/instance/
+/static/

Dockerfile

@@ -0,0 +1,22 @@
+# Use an official Python base image
+FROM python:3.10-slim
+
+# Set the working directory
+WORKDIR /app
+
+# Copy the requirements.txt file into the container
+COPY requirements.txt .
+
+# Install any needed packages specified in requirements.txt
+RUN pip install --trusted-host pypi.python.org -r requirements.txt
+
+# Copy the rest of the application code into the container
+COPY api .
+
+# Expose the port the app runs on
+EXPOSE 8080
+
+ENV OPENAI_API_KEY=zzz
+ENV QDRANT_API_KEY=yyy
+# Start the application using Uvicorn
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "8080"]

api/db/vector_store.py

@@ -0,0 +1,11 @@
+import os
+from qdrant_client import QdrantClient
+from langchain.embeddings import OpenAIEmbeddings
+from langchain.vectorstores import Qdrant
+
+embeddings = OpenAIEmbeddings()
+client = QdrantClient(url="https://32f125d3-5ab1-4058-a10a-bd38a1ebd647.us-east-1-0.aws.cloud.qdrant.io",
+                                    api_key=os.getenv("QDRANT_API_KEY"))
+
+def get_instance(collection: str = "test"):
+    return Qdrant(client=client,collection_name=collection,embeddings=embeddings)

api/exception.py

@@ -0,0 +1,9 @@
+from fastapi import FastAPI, UploadFile, File, HTTPException, Request
+from fastapi.responses import JSONResponse
+from typing import Any, Dict
+
+async def generic_exception_handler(request: Request, exc: Exception) -> JSONResponse:
+    return JSONResponse(
+        status_code=500,
+        content={"message": "Internal server error"},
+    )

api/main.py

@@ -0,0 +1,36 @@
+#!/usr/bin/env python3
+from fastapi import FastAPI
+from routes import embeddings, search, admin
+from fastapi.middleware import Middleware
+from fastapi.middleware.cors import CORSMiddleware
+from datetime import datetime
+from uvicorn.logging import DefaultFormatter
+import logging
+import uvicorn
+from exception import generic_exception_handler
+
+logger = logging.getLogger(__name__)
+handler = logging.StreamHandler()
+handler.setFormatter(DefaultFormatter())
+logger.addHandler(handler)
+
+# Create the FastAPI instance
+app = FastAPI()
+app.include_router(embeddings.router)
+app.include_router(search.router)
+app.include_router(admin.router)
+app.exception_handler(generic_exception_handler)
+
+@app.middleware("http")
+async def log_requests(request, call_next):
+    start_time = datetime.utcnow()
+    response = await call_next(request)
+    process_time = (datetime.utcnow() - start_time).total_seconds() * 1000
+    print(
+        f"{request.method} {request.url.path} {response.status_code} {process_time:.2f}ms"
+    )
+    return response
+
+
+if __name__ == "__main__":
+    uvicorn.run(app, host="0.0.0.0", port=8080)

api/routes/admin.py

@@ -0,0 +1,13 @@
+#This is to init the vector store
+
+from qdrant_client.models import VectorParams, Distance
+from fastapi import APIRouter
+from db import vector_store
+
+router = APIRouter()
+
+@router.put("/admin/v1/db")
+async def recreate_collection(name: str = "test"):
+    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/embeddings.py

@@ -0,0 +1,15 @@
+from fastapi import APIRouter, UploadFile, File
+import openai
+import io
+import os
+from pypdf import PdfReader
+
+router = APIRouter()
+
+openai.api_key = os.getenv("OPENAI_API_KEY")
+
+@router.post("/v1/embeddings")
+async def embed_doc(file: UploadFile = File(...)):    
+    #for now just truncate based on length of words
+    content = await file.read()
+    return openai.Embedding.create(input = content.decode("utf-8"), model = "text-embedding-ada-002")

api/routes/search.py

@@ -0,0 +1,50 @@
+from fastapi import APIRouter, UploadFile, File
+from fastapi.responses import JSONResponse
+import openai
+import io
+import os
+from pypdf import PdfReader
+from langchain.embeddings.openai import OpenAIEmbeddings
+from langchain.vectorstores import Qdrant
+from langchain.schema import Document
+from langchain.chains.question_answering import load_qa_chain
+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="stuff")
+
+@router.post("/v1/docs")
+async def index_doc(file: UploadFile = File(...)):
+    async for doc in generate_documents(file):
+        _db.add_documents([doc])
+    #todo return something sensible
+    return JSONResponse(status_code=200, content={})
+
+@router.get("/v1/docs")
+async def search(query: str):
+    print(query)
+    docs = _db.similarity_search(query=query)
+    answer = _chain.run(input_documents=docs, question=query)
+    return JSONResponse(status_code=200, content={"answer": answer}) 
+
+async def generate_documents(file: UploadFile):
+    num=0
+    async for txt in convert_documents(file):
+        num += 1
+        document = Document(page_content=txt,metadata={"page": num})
+        yield document
+ 
+async def convert_documents(file: UploadFile):
+    #parse pdf document
+    if file.content_type == 'application/pdf':
+        content = await file.read()
+        pdf_reader = PdfReader(io.BytesIO(content))
+        try:
+            for page in pdf_reader.pages:
+                yield page.extract_text()
+        except Exception as e:
+            print(f"Exception {e}")
+    else:
+        return

requirements.txt

@@ -0,0 +1,11 @@
+openai
+uvicorn
+fastapi
+python-multipart
+numpy
+pypdf
+langchain
+tiktoken
+faiss-cpu
+qdrant-client
+

setup.py

@@ -0,0 +1,16 @@
+from setuptools import setup, find_packages
+
+with open("requirements.txt") as f:
+    requirements = f.read().splitlines()
+
+setup(
+    name="doc_index",
+    version="0.1",
+    packages=find_packages(),
+    install_requires= requirements,
+    entry_points={
+        "console_scripts": [
+            # add any command-line scripts here
+        ]
+    },
+)