erikboesen / ocrspace

A Python API wrapper for the ocr.space Optical Character Recognition API.

ocrspace

A Python wrapper for using the ocr.space API.

Installation

Simply install from pip:

pip install ocrspace

Use

First you'll need to import and instantiate the API wrapper:

import ocrspace
api = ocrspace.API()
# Or if you have a custom API host, API key or desired language, pass those:
api = ocrspace.API(endpoint='https://example.host', api_key='Insert key here', language=ocrspace.Language.Croatian)

To perform recognition on an image hosted at some URL:

api.ocr_url('URL of image goes here')

Or, if you have an image locally upon which to perform recognition:

api.ocr_file('image.jpg')
# or:
api.ocr_file(open('image.jpg', 'rb'))  # or any other file pointer

That's it! Look at example.py for a demonstration.

Authorship

This package was created by Ali Najafi and is maintained by Erik Boesen.

License

MIT

ocrspace's Issues

way to use it for local image

I am trying to extract text from image of local file but i cannot comprehend the example.py . Can you give a example?
Currently i use this

How can i do this using wrapper?

Exception: Timed out waiting for results continually comes up

Hello, I've been using your wrapper for the past 3 days and I must say I am impressed with it, great job. But one issue repeatedly troubles me and that is Exception: Timed out waiting for results. I first thought it might be my own network connection or a big-sized image file but it comes up with small sizes as well and I have a good internet connection.
This is the code and traceback:

def convrt_return_txt(pic_location):
    logger.info(f"extracting text from picture")
    try:
        output = api.ocr_file(open(pic_location, 'rb'))
    except Exception as e:
        logger.error(f"unable to extract text from picture -- {e}", exc_info=True)
        return "SOMETHING WENT WRONG, PLEASE TRY AGAIN"
    else:
        logger.info(f"successfully extracted text from picture, returning result to bot")
        return output

2021-12-28 18:59:56,384 - convrt_return_txt - ERROR - unable to extract text from picture -- Timed out waiting for results
Traceback (most recent call last):
  File "/home/name/vscodeworkspace/apps/app_name/src/app_helper.py", line 29, in convrt_return_txt
    output = api.ocr_file(open(pic_location, 'rb'))
  File "/home/name/vscodeworkspace/apps/app_name/venv/lib/python3.8/site-packages/ocrspace/main.py", line 73, in ocr_file
    return self._parse(r.json())
  File "/home/name/vscodeworkspace/apps/app_name/venv/lib/python3.8/site-packages/ocrspace/main.py", line 57, in _parse
    raise Exception(raw['ErrorMessage'][0])
Exception: Timed out waiting for results

Any help will be greatly appreciated.

Upload to PyPI

It would be great if this were installable through pip.

license?

What is the license of this project?

OCREngine 2 and 3 does not work in docker

I am running on Docker- Python bot for discord. When i use anything else instead of api = ocrspace.API like api = ocrspace.API(OCREngine=3) or api = ocrspace.API(OCREngine=2), it will return error api = ocrspace.API()(OCREngine=2) TypeError: 'API' object is not callable

How to use the api.ocr_file()

I use following code, it can work.
text = api.ocr_url('https://images-na.ssl-images-amazon.com/images/I/71ovNJN1URL._SL1244_.jpg')

but, following code can not work.
text = api.ocr_file('D:/shot.jpg')

How to code the api.ocr_file()?

no base-64 image support

please help

Code runs with no errors but doesnt return any text?

Code runs with no errors but doesnt return any text?

Directly return text

I think it would be ideal to process the API's JSON output through this library and return just the text, or None if there was no text found, or raise errors if API connection wasn't successful. I am happy to implement this myself.

Add engine parameter to modify OCREngine payload property

Latest version not uploaded to pypy

Was trying to get the **kwargs options running on a recent project, it wasn't working.

Looked in site_packages and ocrspace there is missing the latest commits.
was installed with pip.