FastAPI Qdrant PDF Embedding Service

Overview

This service provides an API endpoint to upload PDF files, extract and chunk their text, and generate ONNX-accelerated embeddings for each chunk. It is built with FastAPI and leverages PyMuPDF, LangChain, and SentenceTransformers.

---

Setup Instructions

1. Clone the Repository

git clone <your-repo-url>
cd fastapi_qdrant

2. Create and Activate a Python Environment

We recommend using Miniconda or [venv]:

conda create -n fastapi_qdrant python=3.10 -y
conda activate fastapi_qdrant
# OR
python -m venv .venv
source .venv/bin/activate  # On Windows: .venv\Scripts\Activate.ps1

3. Install Dependencies

cd backend
pip install -r requirements.txt

Note: Ensure only PyMuPDF (shows as pymupdf in pip freeze) is installed for PDF processing. Do not install the unrelated fitz package.

---

Running the API Server

From the backend directory, start the FastAPI server:

uvicorn app.main:app --reload

• The API will be available at: http://127.0.0.1:8000
• Interactive docs: http://127.0.0.1:8000/docs

---

Testing the PDF Upload Endpoint

Option 1: Swagger UI (Recommended)

1. Go to http://127.0.0.1:8000/docs
2. Find the /api/upload-pdf endpoint.
3. Click "Try it out".
4. Upload a PDF file.
5. Click "Execute".
6. View the JSON response with chunk/embedding stats and timings.

Option 2: cURL

curl -X POST "http://127.0.0.1:8000/api/upload-pdf" -H  "accept: application/json" -H  "Content-Type: multipart/form-data" -F "file=@sample.pdf"

Option 3: Python Script

import requests
url = "http://127.0.0.1:8000/api/upload-pdf"
with open("sample.pdf", "rb") as f:
    files = {"file": ("sample.pdf", f, "application/pdf")}
    response = requests.post(url, files=files)
print(response.json())

---

Troubleshooting

• ModuleNotFoundError: No module named 'fitz'
  • Ensure PyMuPDF is installed (pip install PyMuPDF).
  • Do not install the unrelated fitz package from PyPI.
• 500 Internal Server Error
  • Check the server logs for details. Common causes:
    • Invalid or corrupted PDF file.
    • Missing ONNX model file (onnx/model_qint8_avx512.onnx).
    • Dependency issues (see requirements.txt).
• ONNX Model Not Found
  • Ensure the required ONNX model file is present in the correct path, or update the path in the code.

---