About project

---

The project is an implementation of a microservice for reading text from images, powered by Tesseract OCR, that can be easily incorporated in any application via a simple-to-use API built with FastAPI. The whole microservice is containerized using Docker, making it easier for anyone to set up a local copy and bend it to their needs.

The microservice also cleans and processes the uploaded images with OpenCV; improving the OCR predictions of the Tesseract model.

Built With

• Python
• FastAPI
• Tesseract OCR
• OCRmyPDF
• Docker

Run locally

---

Clone the repo

git clone https://github.com/mubasharkk/fastapi_ocr.git

Install dependencies

pip install -r requirements.txt

Run Server

cd app
uvicorn pdfapi:app --host 0.0.0.0 --port 8000 --reload

Run on Docker

docker build -t fastapi_ocr .   

Run the docker container

docker run -d --name my_container api_ocr 

Documentation

---

The api contains the following endpoints

• /extract_text - returns text from uploaded file

• /extract_text_from_many_files - return text from all uploades files

• /extract_text_from_url - return text from url with image

Running Tests

cd /var/www/app

export PYTHONPATH=$PWD

pytest -q tests