Setup

1. Clone the repo locally git clone https://github.com/belharethsami/fastapi-extract-text.git

2. Navigate to the repo using cd fastapi-extract-text

3. Deploy to GCP using gcloud run deploy --source .

API documentation

Base URL: https://fastapi-ocr-155731627853.us-east1.run.app

Method: POST

Path: extract-text

Auth: Public

Request format

Request content type: multipart/form-data

Field name: file

File type: JPEG/JPG only

Max size: 10 MB

Response format and possible error codes

Format for successful response (200):

{
    "success": true,
    "time": 55
    "text": "The quick brown fox jumped over the lazy dog."
}

success: boolean

time: processing time in millisenconds (float)

text: full OCR text (string)

Possible error codes:

400 Bad Request

• File format isn't JPEG/JPG

• File is empty

413 Request Entity Too Large

• If file uploaded is larger than 10 MB

422 Unprocessable Entity

• If file field wasn't sent as a part of multipart/form-data

500 Internal Server Error

• Failed to create Vision API client (credential issues, etc.)

502 Bad Gateway

• Call to Vision API failed

Example curl command for testing

curl -X POST -F "file=@/path/to/image.jpg" https://fastapi-ocr-155731627853.us-east1.run.app/extract-text

Implementation explanation

OCR service used: GCP Vision API

Handling file uploads and validation: An HTTP middleware checks to make sure the file size is less than 10 MB if present.

If the file format is not JPEG/JPG, we return a 400 error. If the binary is empty, we return a 400 error.

Deployment strategy: The FastAPI server is hosted on Cloud Run. It can be deployed by cloning the repo locally, and building on GCP using: gcloud run deploy --source .

GCP automatically manages Application Default Credentials for authentication to the Vision API.