TypeError: argument of type 'AskFileResponse' is not iterable

[poojitharamachandra]

I am trying to upload a pdf file , but unable to read it.(the same code runs without any error if i upload a text file) here is the code:

# Wait for the user to upload a file
while files == None:
    files = await cl.AskFileMessage(
        content="Please upload a text file to begin!", accept=["application/pdf"],max_size_mb=100
    ).send()

file = files[0]

TypeError: argument of type 'AskFileResponse' is not iterable

[willydouhard]

What version of chainlit are you using? You can run pip show chainlit to find out

[poojitharamachandra]

chainlit==0.4.101

[willydouhard]

Yes this is an old version. Try run pip install -U chainlit to upgrade

[poojitharamachandra]

even with 0.6.3, the same issue exists.

[willydouhard]

Are you sure you are running the correct version of chainlit ? You might have multiple python version installed. You can try to start chainlit like that to make sure:

python -m chainlit run …

[poojitharamachandra]

is this implemented in 0.6.3? Upload File button in front-end UI, just like OpenAI code interpretor · Issue #198 · Chainlit/chainlit (github.com)

[poojitharamachandra]

I am able to read .txt file in 0.6.3(TypeError: argument of type 'AskFileResponse' is not iterable is not there anymore), but unable to read a pdf file. How do I correctly read a pdf file here?

[willydouhard]

Here is an example with langchain https://github.com/Chainlit/cookbook/blob/c65d989e90614da867fed2a4e06ad9330d95ff06/pdf-qa/app.py#L75 check the process_file function

[poojitharamachandra]

thanks, i am still unable to process the content of pdf, as it contains special characters. none of the decoding techniques such as utf-8, latin-1,unicode etc are working

[willydouhard]

There are multiple python pdf loader packages available. You can probably find one that fit your need :)

[poojitharamachandra]

is there any way to give the loader as input to AskFileMessage? Because, my default, AskFileMessage give data in form of bytes, which is causing problem for my usecase, as I have multiple special characters in the pdf.

[poojitharamachandra]

hi,

is this implemented in 0.6.3? https://github.com/Chainlit/chainlit/issues/198

[willydouhard]

I don't think it is the responsibility of AskFileMessage to parse the file. The developer should be able to use any loader from the received bytes.

[willydouhard]

hi,

is this implemented in 0.6.3? #198

This will be part of the next release.

[poojitharamachandra]

Here is an example with langchain https://github.com/Chainlit/cookbook/blob/c65d989e90614da867fed2a4e06ad9330d95ff06/pdf-qa/app.py#L75 check the process_file function

is it possible to write the contents to a pdf/word file instead of a temp file?

[willydouhard]

To update an existing file or to create it from scratch? If you just want to create the file you can do

with open('output.pdf', 'wb') as f:
    # Write the bytes to the file
    f.write(file.content)

[AbhijitManepatil]

chainlit==0.7.400 version works for me

[willydouhard]

Since chainlit 1.0.0 this api changed slightly. Check this example