Downloading weights library in real time

[utkarsh-vijay]

This is a issue while using AWS lambda as it doesn't allow write permissions on /var/lang/lib/python3.11/site-packages/qrdet/.model'. Can we change the directory of download to a custom path? Like i want to change it to /tmp
[Errno 30] Read-only file system: '/var/lang/lib/python3.11/site-packages/qrdet/.model'

[sam-30]

This issue will have to be fixed in qrdet. I believe it's an easy fix. I'll spin up a pull request on that repository.

[sam-30]

Issue in qrdet. Comments welcomed on my proposed solution there.

[Eric-Canas]

Hi @sam-30 !

Thanks a lot! Answered there!

[pibahamondesw]

I'm having the same issue in case you came up with a solution, I figured I could run a short test like this one to make sure that my image downloaded the weights first:

test_qreader.py

import cv2 as cv
from qreader import QReader

def test_qreader():
    qreader = QReader()
    image = "test.png"
    image = cv.imread(image)
    decoded_text = qreader.detect_and_decode(image=image)
    assert decoded_text[0] == "https://www.example.com"

if __name__ == "__main__":
    test_qreader()

Dockerfile

FROM public.ecr.aws/...
#####
# All my code to prepare the lambda function
#####

# Predownload weights for model
ADD test.png /tmp/test.png
ADD test_qreader.py /tmp/test_qreader.py
RUN cd /tmp && python /tmp/test_qreader.py
RUN rm /tmp/test.png /tmp/test_qreader.py

CMD [ "lambda_function.handler" ]

By the way, test.png is just a QRCode I generated with the URL "https://www.example.com".

Let me know if you implement a more robust solution!

[Eric-Canas]

Hi!

Just solved, now there is a new parameter in the constructor called weights_folder. If you have problems with writing on the default folder, you can instantiate QReader as QReader(weights_folder='/tmp') to make sure the model is downloaded and readed from there.

It should work if you pip install --upgrade qreader