arxiv.HTTPError

[LexTran]

Hi, dear author:

Thanks to your great contribution of releasing this repo, it's a great project. However, I recently encountered a strange error.

I have successfully deploy action before, but I found it failed this morning. The error log is shown below. How can I solve it？

[wy9933]

same issue, seems to be caused by the excessive requests to arxiv
looking forward to the solution from author

[LI-MAX72]

我也遇到了，是不是要把爬虫速率调低些

[Dennis-Huangm]

me too

[cccchou]

me too

[nehSgnaiL]

same issue

[wy9933]

same issue, seems to be caused by the excessive requests to arxiv looking forward to the solution from author

I try to reduce the batchsize in main.py L58 and L59 from 50 to 20, it works

but got another error same as issue 120, I run a demo in my PC, which copied from arxiv document using arxiv version 2.2.0 and 2.3.0

import arxiv
client = arxiv.Client()

paper = next(arxiv.Client().results(arxiv.Search(id_list=["1605.08386v1"])))
print(paper)
print(paper.title)
print(paper.pdf_url)

but get this output

http://arxiv.org/abs/1605.08386v1
Heat-bath random walks with Markov bases
None

and I read code in arxiv __init__.py, self.pdf_url = Result._get_pdf_url(links) in constructor of Result cannot get the pdf_url, in

def _get_pdf_url(links: List[Link]) -> str:
        """
        Finds the PDF link among a result's links and returns its URL.

        Should only be called once for a given `Result`, in its constructor.
        After construction, the URL should be available in `Result.pdf_url`.
        """
        pdf_urls = [link.href for link in links if link.title == "pdf"]
        if len(pdf_urls) == 0:
            return None
        elif len(pdf_urls) > 1:
            logger.warning("Result has multiple PDF links; using %s", pdf_urls[0])
        return pdf_urls[0]

the pdf_urls is empty list []

I don't the error comes from arxiv respositoy or if the arxiv website has updated its access rules

[wy9933]

After further testing, it seems that the arXiv website has changed its response format, which makes the arxiv library unable to get pdf_url properly. I’ve reported this issue to the arxiv repository issue 181

[TideDra]

I will take a look at this when I'm free

[TideDra]

Hey guys. I have added a temporary patch fix to the arxiv lib, according to #120 (comment), can you try again?

[cccchou]

it shows another error in render email: AttributeError: 'NoneType' object has no attribute 'replace'

Hey guys. I have added a temporary patch fix to the arxiv lib, according to #120 (comment), can you try again?

[wy9933]

it shows another error in render email: AttributeError: 'NoneType' object has no attribute 'replace'

Hey guys. I have added a temporary patch fix to the arxiv lib, according to #120 (comment), can you try again?

I succeeded, did you use your own forked repository?

[cccchou]

it shows another error in render email: AttributeError: 'NoneType' object has no attribute 'replace'

Hey guys. I have added a temporary patch fix to the arxiv lib, according to #120 (comment), can you try again?

I succeeded, did you use your own forked repository?

yes, but i sync forked

[LI-MAX72]

it shows another error in render email: AttributeError: 'NoneType' object has no attribute 'replace'

Hey guys. I have added a temporary patch fix to the arxiv lib, according to #120 (comment), can you try again?

I succeeded, did you use your own forked repository?

yes, but i sync forked

I sync forked, and it worked. Maybe other bugs?