Jump to content

Commons:Bots/Requests

From Wikimedia Commons, the free media repository
This project page in other languages:

Shortcut: COM:BRFA

If you want to run a bot on Commons, you must get permission first. To do so, file a request following the instructions below.

Please read Commons:Bots before making a request for bot permission.

Requests made on this page are automatically transcluded in Commons:Requests and votes for wider comment.

Requests for permission to run a bot

[edit]

Before making a bot request, please read the new version of the Commons:Bots page. Read Commons:Bots#Information on bots and make sure you have added the required details to the bot's page. A good example can be found here.

When complete, pages listed here should be archived to Commons:Bots/Archive.

Any user may comment on the merits of the request to run a bot. Please give reasons, as that makes it easier for the closing bureaucrat. Read Commons:Bots before commenting.

Operator: 7804j (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought: Batch upload of the remaining swisstopo images datasets. The first task was approved in (Commons:Bots/Requests/OKA bot) but focused on the terrestrial dataset of 50k images (in progress: over 10,000 uploaded so far, see Category:Terrestrial photographs by swisstopo). We've estimated the size of the other collections, published by swisstopo under {{Attribution-Swisstopo}} (opendata.swiss BY):

Collection Images Typical size
Aerial images, black and white (1921–2010) 222,194 ~160 MB
Aerial images, colour (1952–2010) 69,335 ~165 MB
Aerial images, infrared (1951–2009) 52,482 ~160 MB
Aerial images, oblique (1927–2004) 2,254 ~160 MB
Technical images (survey points, instruments; glass plates from 1909) 18,888 ~60 MB
Orthophotos (map-corrected versions of about a third of the black-and-white frames) ~75,000 ~80 MB
  1. Sample uploads, one per collection:
    Same approach as for the terrestrial set: metadata from swisstopo (date, camera, focal length, format, flight altitude, position), localized fields via Wikidata IDs, structured data, captions in en/de/fr, and categories only in per-collection subcategories by municipality and year (following community feedback on the first task).
    Resolution: we propose the original scans at full resolution (about 17,000 × 17,000 px for aerial frames), with only the built-in preview layers removed (no re-encoding), so that the full detail of the film stays available for research and measurement. Where swisstopo publishes an orthophoto (a version corrected to fit the map), it is uploaded as a second file, linked from the scan's page and vice versa. Total volume is about 62 TB, so we would upload collection by collection (oblique and technical images first) and run from Toolforge.
    If that's too much volume, two alternatives: uploading every other frame along each flight line (consecutive frames overlap by about 60%, so coverage is kept, about half the volume), or reduced-resolution copies
  2. Depicts statements from the footprint: for files uploaded by the bot (including the terrestrial set), add depicts (P180) for named summits, lakes and glaciers that lie inside the photographed area, using the Wikidata items that OKA bot maintains on Wikidata from swisstopo's swissNAMES3D data (d:Wikidata:Requests for permissions/Bot/OKA bot).

Automatic or manually assisted: Automatic, supervised. Each batch of 1,000 is checked before upload (duplicates by SHA-1 and inventory number, dry-run parse, coordinate plausibility); files with any problem are set aside, and a batch stops for review if more than 5% are set aside.

Edit type (e.g. Continuous, daily, one time run): One-time runs per collection, in batches.

Maximum edit rate (e.g. edits per minute): up to 4 per second, as approved for the first task

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

Programming language(s): Python (Pywikibot); code: https://github.com/Open-Knowledge-Association/oka-commons-bot

@EugeneZelenko: FYI as you reviewed the first request.

7804j (talk) 12:45, 9 October 2026 (UTC)reply

Discussion

Operator: Hogckk (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought: Uploading public domain local documents of Zhejiang Province, China, such as government gazettes and historical documents. The first job is the Gazette of the People's Government of Zhejiang Province (浙江省人民政府公报, called 浙江政报 before 2008), 1982–2026, from the official website. That's about 1200 PDFs. 56 issues are already in Category:Gazette of the People's Government of Zhejiang Province, uploaded by others under {{PD-PRC-exempt}}; I'll use the same license, category and file naming, and skip files already on Commons.

The bot account was created while I was logged out; ownership is confirmed on both user pages.

Automatic or manually assisted: Supervised, run by hand in batches

Edit type (e.g. Continuous, daily, one time run): Batch uploads from time to time

Maximum edit rate (e.g. edits per minute): 6

Bot flag requested: (Y/N): Y

Programming language(s): Python (Pywikibot)

Hogckk (talk) 10:32, 9 October 2026 (UTC)reply

Discussion

Operator: Lin (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought:

READ-ONLY research support for Shandong Normal University (academic). Search Wikimedia Commons (MediaWiki API) and download free-licensed civil aircraft photographs identified by aircraft registration (e.g. File pages / Special:FilePath or upload.wikimedia.org) into an external research dataset for computer-vision work.

The bot will NOT edit wiki pages, NOT upload files, NOT delete anything, and NOT run continuous 24/7. Dataset metadata will keep file page URL, photographer attribution, and license.

Estimated volume: about 20,000–30,000 image files, intermittent sessions.

User-Agent: LinAircraftBot/1.0 (Shandong Normal University academic research; 256043794@qq.com)

Automatic or manually assisted: Manually assisted (script started by operator; operator monitors; no unsupervised daemon)

Edit type (e.g. Continuous, daily, one time run): Intermittent download sessions (no wiki edits). Download requests about 1/second, single-threaded; on HTTP 429 / Retry-After the script pauses.

Maximum edit rate (e.g. edits per minute): N/A — no page edits. Maximum download request rate: ~60 requests/minute (1 request/second).

Bot flag requested: (Y/N): Y — so automated access is identifiable and not confused with human browsing; if bot flag is inappropriate for read-only downloads, please advise what authorization is preferred instead.

Programming language(s): Python 3 (requests)

Linasdfg (talk) 01:27, 8 October 2026 (UTC)reply

Discussion

This should be declined. Bot can download as per normal API limits, which are reasonably high if a proper user-agent is set. Please note that the layout is different: https://www.mediawiki.org/wiki/API:Etiquette --Schlurcher (talk) 20:36, 10 October 2026 (UTC)reply

Operator: Corplox (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought:

I am requesting permission to operate CorpLoxTradeBot to upload approximately 20,000 original SVG visualizations of international merchandise trade statistics to Wikimedia Commons.

I operate CorpLox (https://corplox.com), a global trade analytics platform. The visualizations are created by CorpLox using UN Comtrade statistics.

The planned collection includes country-level, bilateral and global trade charts, covering merchandise exports and imports, product sectors and other trade indicators.

I propose starting with a small test focused on country-level exports and imports by HS2 product sector for 2025. Further uploads and chart types will be introduced in stages, subject to community approval.

Four examples have already been uploaded using my existing workflow:

The custom Python uploader supports:

  • Checking whether a filename already exists on Commons.
  • Checking for duplicate file content.
  • Preparing descriptive English filenames and descriptions.
  • Identifying UN Comtrade as the statistical source and CorpLox as the visualization creator.
  • Applying the CC BY 4.0 license.
  • Adding appropriate Commons categories.
  • Adding verified structured metadata.
  • Recording upload results and errors.

The uploader is integrated into my private software infrastructure. I do not intend to publish the source code, but I can explain its technical operation, API interactions, duplicate detection, rate limiting and error handling.

I am openly disclosing my affiliation with CorpLox. The purpose of the uploads is to contribute reusable statistical visualizations for educational and research use, not to promote the platform.

I understand that the proposed collection is large. I am requesting permission to begin with a small supervised test and will follow the community's instructions regarding the scope and conditions of subsequent uploads.

Automatic or manually assisted: Automatic (supervised). Each batch is initiated manually by the operator. I will review the results and remain available to stop the process or correct errors.

Edit type (e.g. Continuous, daily, one time run): Intermittent, manually initiated batches. The initial test will consist of 5–20 edits, or another number recommended by the reviewing community. The planned collection will be uploaded gradually, subject to approval.

Maximum edit rate (e.g. edits per minute): Planned maximum of 4 write operations per minute, including file uploads and metadata edits. The rate limit will be implemented and verified before testing.

Bot flag requested (Y/N): Y

Programming language(s): Python 3, using the MediaWiki Action API.

Corplox (talk) 16:15, 25 September 2026 (UTC)reply

Discussion
Oppose. I can't find any evidence of Corplox being cited in acadamia; and the domain was registered just this year; browsing through the site all the "About X company" blurbs are AI-generated, and no sources are provided each entry. Given that this is a:
  1. Untested dataset, with
  2. Dubious verifibility, and
  3. Questionable licensing
I do not want 20,000 graphs of this on Commons.
Questionable licensing
Quoting from the TOS of COMTRADE, one of your listed data sources: 5. Usage. United Nations COMTRADE contains copyrighted material and/or other proprietary information. The materials contained on or otherwise made available to you on United Nations COMTRADE may be copyrighted by the United Nations and, thus, are protected by copyright laws and regulations worldwide. Except as may otherwise be expressly permitted by applicable law, any copying, automated browsing or downloading, redistribution, publication, or commercial exploitation of any material contained on or otherwise made available to you on United Nations COMTRADE is strictly prohibited without the prior written permission of the United Nations. I don't see evidence of a prior written permission of the United Nations.
OpenCorporates also has attribution rules, which I don't see being followed: The ODbL requires you to attribute your use of our database. This is important as it allows us (and in turn the community) to benefit from reuse of our work, and so allows us to continue to provide this service. Use of any data must be accompanied by a hyperlink reading “from OpenCorporates” and linking to either the OpenCorporates homepage or the page referring to the information in question. The URL for this is returned in most API calls and if you are making use of the API we would expect you to use the latter form. The attribution must be large and clear enough to be clearly associated with the usage, and must in all cases be no smaller than 70% of the largest font relating to the information or 7px , whichever is the larger. If you are making the information available via your own API you need to make sure your users comply with all these conditions, and if the information from OpenCorporates forms the substantial part of the web-page, you should use the rel-canonical tag to indicate to search engines that the OpenCorporates page for the company is the core page for this topic.
"Official business and corporate registries" is also delightfully vague for a source.

msk 16:48, 25 September 2026 (UTC)reply

Thank you for reviewing my request and raising these concerns.
I would like to clarify several points.
1. Data sources and verifiability
The proposed trade charts are based on UN Comtrade statistics. CorpLox is the creator of the visualizations, not the original publisher of the underlying trade statistics.
The proposed upload collection is separate from the company-information section of the CorpLox website.
I understand the concern about verifiability. I can provide the specific UN Comtrade data sources, query parameters and calculation methodology for representative charts so that their results can be independently checked.
2. OpenCorporates
OpenCorporates is not a data source for the proposed international trade charts.
I understand that you have also raised concerns about other sections of the CorpLox website. However, those company-information pages are not included in the proposed Wikimedia Commons uploads.
3. UN Comtrade licensing
UN Comtrade's published policy distinguishes between redistribution of original datasets and substantially transformed data. Its documentation also identifies certain exemptions for publishing small numbers of graphs.
I recognize that these provisions do not automatically establish permission for the proposed large collection. I will not claim that the entire collection qualifies for an exemption without establishing the applicable conditions.
4. Proposed upload volume
The figure of approximately 20,000 files represents the potential size of the collection, not an intention to upload everything immediately.
I understand that such a large contribution requires additional scrutiny, particularly regarding licensing, verifiability and the usefulness of individual charts.
I will pause further bot uploads while these concerns remain unresolved.
I would appreciate feedback on the specific source documentation and methodological information that would be useful for assessing the existing examples.
Thank you for your time. Corplox (talk) 17:06, 25 September 2026 (UTC)reply
  1. I am a human being, and I would like to communicate to you as a human being, not your chatbot.
  2. "I violated copyright, but I'm not uploading those specific files to Commons" is not all that convincing of an argument. Who would've thought?
Could I also get a administrator to procedurally close this BRFA for being AI-generated? Thanks. msk 17:35, 25 September 2026 (UTC)reply
I understand your concerns.
I used AI to help formulate my previous response, but I am personally responsible for this project and the proposed uploads.
I have not stated that I violated copyright, and I would appreciate it if my position were not misrepresented. Corplox (talk) 17:45, 25 September 2026 (UTC)reply

Operator: Armin hpj (talk · contributions · Statistics · Recent activity · block log · User rights log · uploads · Global account information)

Bot's tasks for which permission is being sought: Automatically uploads the additional colors of the BSicons.

Automatic or manually assisted: Automatic

Edit type (e.g. Continuous, daily, one time run): Continuous

Maximum edit rate (e.g. edits per minute):

Bot flag requested: (Y/N): Yes

Programming language(s): Python

Armin hpj (talk) 11:56, 11 September 2026 (UTC)reply

Discussion
  • Please make test run. --EugeneZelenko (talk) 15:04, 11 September 2026 (UTC)reply
  • Please pay very careful attention to categorisation, utilising color name sub-categories where appropriate, and avoid uploading large numbers of BSicons to "other colors" categories. Redlinked categories should be created with the relevant template when new BSicons are added to them. A full description should be included and not just "colored icon for BSicon diagrams" or similar. Creating and uploading the BSicons is only a small part of the task, and the bot should include every aspect in its scope without leaving additional work for other editors. -=# Amos E Wolfe talk #=- 01:51, 12 September 2026 (UTC)reply
  • The first edit is not related to the test --Armin hpj (talk) 23:32, 12 September 2026 (UTC)reply
    Following your recent test uploads, you need to look at some of the settings of the Bot. For example there is no "set ex" — icons with that prefix are simply out of service variants of the default set. Likewise "set uex" are part of "set u" and not their own color set.
    As per my earlier comment, when uploading these files, a number of redlinked categories need to be created by adding {{BS-category}} and saving with an edit summary. This needs to be completed by the Bot as soon as a file is uploaded and added to a category that does not yet exist. Please do not rely on other editors to clean up after your Bot.
    You have added the description Colored variant of the BSicon KRZ3+1o+l icon which is not descriptive of the content of the BSicon itself. In this example the default color BSicon has the description Crossing through over line to corner 3 from corner 1, with connection from left which should be simple enough to copy and paste manually, or scrape from the relevant template in the file's description, and have the Bot add (with variants for set name and e / x / ex status) at the time of upload.
    Having the Bot's output be clean, concise, accurate, and not requiring manual revision by other editors is more important than any speed advantage the Bot may or may not confer. Please consider carefully whether running a Bot is the best use case for the task at hand, which may be better suited to semi-automated tools such as the Python script by User:SomeRailFan running on your local machine, followed by batch-uploading using Pattypan. -=# Amos E Wolfe talk #=- 02:29, 13 September 2026 (UTC)reply
    Thank you for saying this. If this bot has been approved, before I run it, I will edit it's scripts (for example, it automatically adds "<color> version of " before the English description and... ) Armin hpj (talk) 11:19, 13 September 2026 (UTC)reply