Commons:Bots/Requests/LibraryBot

LibraryBot (talk · contribs)

Operator: Andrew Gray (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information) (assign permissions)

Bot's tasks for which permission is being sought: Uploading files provided by the British Library (and some BL-related organisations, eg the International Dunhuang Project) as part of the Wikipedian in Residence program. Files and metadata will be provided locally and checked before uploading.

Automatic or manually assisted: Goal is to be automatic and supervised, but intending to run as manually assisted for the first runs.

Edit type (e.g. Continuous, daily, one time run): Periodic manually-triggered runs when files are available.

Maximum edit rate (eg edits per minute): 2-3? Depends on upload speed.

Bot flag requested: (Y/N): Y.

Programming language(s): Standard pywikipedia bot framework.

Andrew Gray (talk) 22:46, 22 October 2012 (UTC)

Discussion

• This is my first attempt at setting up a bot account; please do feel free to correct any errors! I'll try and make some test uploads tomorrow (don't have any BL material to hand at the moment). Andrew Gray (talk) 22:55, 22 October 2012 (UTC)

  Please make a test run. --EugeneZelenko (talk) 14:51, 23 October 2012 (UTC)

  I've made three manually-controlled uploads - eg/ File:Ludwig van Beethoven - String Quartet Op 131 - British Library Add MS 38070 f51r.jpg - and all looks well. Next step, test with a batch! Andrew Gray (talk) 13:28, 24 October 2012 (UTC)

  Please use language templates for Artist/Title/Description fields. --EugeneZelenko (talk) 14:30, 24 October 2012 (UTC)

  Done, as a pre-programmed batch of ten rather than individually triggered, but still with manual approval of each. (It ran some uploads with no author field, but I've fiddled the database to put {{en|1=unknown}} in those in future.) Andrew Gray (talk) 21:47, 25 October 2012 (UTC)

• I've only looked at File:Ancient Khotan BLER4 AKV2 PLIII PHOTB.jpg so far. The catalogue entry link is broken, and took me here. Also, this is not cc-zero. It was not released into the public domain by Aurel Stein, it's in the public domain because of its age. --99of9 (talk) 01:29, 26 October 2012 (UTC)

  CC-zero was as given to me, and I confess I was a little dubious about its precise appropriateness as well, so thanks for putting my mind to rest here; I'll check it's switched in the other records :-)

  I've made a few layout changes at File:Ancient_Khotan_BLER4_AKV2_PLIII_PHOTB.jpg, including linebreaks and section headers to make it easier for humans to edit later. That's also now my suggestion for the layout and copyright info. I realized they were saying that even if UK law is considered to grant them a new copyright as the result of the photograph/scan, they are releasing that too under CC-0, so it is worth recording for super safety. --99of9 (talk) 09:46, 26 October 2012 (UTC)

  Yeah, I've had a chat with the IDP group about this. We've avoided putting the public domain dedication on the other BL-sourced images, so for the moment I'll leave CC-zero off entirely - as it's released by the group who would otherwise claim rights, I think the release of any "accidental" copyright is fairly clear.

  As to human-readability, I'd love to, but it seems that pywikipedia strips out all the linebreaks and spacing! Andrew Gray (talk) 11:06, 26 October 2012 (UTC)

  I use pywikipedia bot too. I've put a method of getting around this on your talk page. --99of9 (talk) 11:18, 26 October 2012 (UTC)

  It's all down to ", it seems. How's this? Andrew Gray (talk) 11:53, 26 October 2012 (UTC)

  As for the catalogue link, hmm; it looks like "Ancient Khotan" on its would work, as the image records aren't volume-separated - it's a finicky system! I've tested it with the standard MS/photo identifiers and those seem to work okay, so it looks like it'll just be the book extracts needing double-checked. I'll go through these today. Andrew Gray (talk) 08:45, 26 October 2012 (UTC)

  I'll leave this to you to figure out! --99of9 (talk) 09:47, 26 October 2012 (UTC)

  I've found a couple more using "Ancient Khotan" or "Serindia", Stein's two major books, and corrected them. I'll run another test shortly. Andrew Gray (talk) 11:06, 26 October 2012 (UTC)

  If you can also provide a deep link directly to the source image itself, that would be helpful, e.g. when I click on "Large Image", I get [1]. --99of9 (talk) 12:06, 26 October 2012 (UTC)

  I don't know if that's possible - my files only have the master filename and shelfmark, and the template goes through a resolver URL to change the shelfmark to the image. It may be possible, but I think this would have to be done via the template (which has the advantage that we can do it retroactively!) Andrew Gray (talk) 12:16, 26 October 2012 (UTC)

  The template seems to have a url field which you are leaving empty. I'm not sure the template could possibly work out the url from the shelfmark. What is the format of the resolver url? Maybe provide that? --99of9 (talk) 12:23, 26 October 2012 (UTC)

URL is an override in case it's not in the catalogue (eg posted on a blog); it's for a complete URL to an external site. The resolver URL is encoded as idp.bl.uk/database/oo_loader.a4d?pm=SHELFMARK. I'll see if we can extend the template (but this won't impact the content in the bot uploads, I think) Andrew Gray (talk) 12:56, 26 October 2012 (UTC)

• Oh ok, in that case it should be reasonably easy to add a follow-through link to the template. So I think I'm ready to  Support this application based on the last few uploads. Is it a deliberate choice to leave off the usual "Summary" heading within the file page? --99of9 (talk) 13:36, 26 October 2012 (UTC)

  Not really - this is just the way I've been tending to do it when I've used the "raw" upload form. I'll update this for the next batch of templates I generate.

  Thanks for the support - is it okay for me to keep trickling these through under manual supervision while we wait for full bot approval? Andrew Gray (talk) 13:41, 26 October 2012 (UTC)

• Date in non-numeric form should be also enclosed in language template. --EugeneZelenko (talk) 14:47, 26 October 2012 (UTC)

  If I enclose the date field in a language template in all cases, will this cause any problems if there's only numeric data there? Andrew Gray (talk) 14:56, 26 October 2012 (UTC)

  I think this may interfere with internationalization of such dates. --EugeneZelenko (talk) 14:31, 27 October 2012 (UTC)

  Fair enough. I believe all of the current batch are non-numeric ("early 8th century", "19 November 2011") so I'll use language-templates as default, and make a note to check this when preparing any later batches. Andrew Gray (talk) 21:44, 27 October 2012 (UTC)

---

If there are no objections, I think bot status should be granted. --EugeneZelenko (talk) 14:45, 1 November 2012 (UTC)