Hello.

I am uploading tons of pdf files to Commons with the bot account [[User:PencakeBot]]. Those files are scanned copies of ancient books from the National Library of China (read.nlc.cn).

To automate the upload process, I write a script in Python with mwclient. But just at the very first, I encounter frustrating problems. Some files are of 100+MiBs (typically less than 150MiB). For these (not so) big files, chunked-uploading returns a timeout error (typically 504 upstream request timeout) almost all the time for the final chunk. After spending hours of effort to retry it again and again in different network environments and alternative bot frameworks, I still cannot manage to upload them.

I was told by kind people in #wikimedia-tech that there are long-term stability problems with chunked-uploading. I realize uploading by URLs seems an ideal alternative way. nlc.cn does not provide static URLs for all of its files. So I am planning to let Commons fetch files from my server.

Could you please add my domain hoiscript.toolforge.org to the wgCopyUploadsDomains list so that I can upload files directly from my domain/server to Commons?

Thanks!