User talk:Mateusz Konieczny

Examples for pedestrian roads

Hi, I would like to discuss our edits on the Tag:highway=pedestrian page – particularly this partial restoration of content I had removed. To me, this image depicts a wide, comfortable footway – but that's still a case for highway=footway. Unless there are other arguments than what's visible on the image (e.g. signage), I don't see it as a good example of a pedestrian road. --Tordanik 17:28, 27 April 2018 (UTC)

Google using OpenStreetMap data

Re: "When Google stopped map maker? Maybe it was added not by Google employee but by someone convinced to work for corpo for free?"

I was offered a chance to be approved as a unpaid mapper for Google a while back, when I had been still adding photos and POIs to Google Maps. I believe if you are a trusted volunteer contributor you can still get access to update street names and perhaps even geometries, though I decided not to give a corporation free labor anymore. I suspect that you are right, these changes were probably copied from OpenStreetMap by a volunteer mapper who did not properly understand the copyright issues. --Jeisenbe (talk) 23:03, 6 February 2021 (UTC)

building:name=* has no valid reason for use

What about in cases where the name of a business is different from the name of the building that it is in? --Adamant1 (talk) 07:43, 21 April 2021 (UTC)

@Adamant1: Merging business and building is in general violation of One feature, one OSM element. Especially if both are named and so on. In such case proper fix is to represent business as a node/area in a building, rather than starting using weird prefixes. If building has no name or other conflicting attributes such merge is not very problematic (though for example iD has severe problems with it), but building:name=* is ridiculous workaround used instead of a proper soution Mateusz Konieczny (talk) 07:51, 21 April 2021 (UTC)

I guess that makes sense. I'm pretty sure I've used the tag myself a few times. So Thanks for the answer. I'll have to just do what you suggest the next time. --Adamant1 (talk) 08:16, 21 April 2021 (UTC)

Re:Source of file

Hi Mateusz, the image was generated only by OSM data, it was a screenshot from JOSM. I had initially uploaded the image to my account on a free social network that no longer exists called gnewbook. Here you can see other similar images from the mapping event in 2011. --Ovruni (talk) 08:40, 10 March 2022 (UTC)

Hi Mateusz, the file which I used in Mashhad OSM wiki page is produced by myself via a web page that produces a .GIF file of mapping history of a location over the time. Unfortunately this useful site doesn't works yet, but I know that we could download and use its images whenever and where ever we want.

Wikimedia photos

Thank you for keeping an eye on on copyright violations. I can assure you that all photos I have posted on Wikimedia have been taken by me personally. T99 (talk) 09:00, 6 June 2022 (UTC)

@T99: thanks for info, and great that files are not problematic! Specifying license like at https://wiki.openstreetmap.org/w/index.php?title=File:Bike_path_dip_swale.jpg&curid=188024&diff=2337294&oldid=2253049 for files listed in https://wiki.openstreetmap.org/wiki/User:Mateusz_Konieczny/notify_uploaders/T99 would be even more ideal, if possible (only few files are affected, so I guess that it should be doable) Mateusz Konieczny (talk) 09:12, 6 June 2022 (UTC)

Hey Mateusz, thank you for keeping an eye on images copyrights. Maybe the text you add in user spaces when asking for image license clarification could be made a lot shorter: a small text and the useful {} to be added on the image page, along with a link to the long version. Yvecai (talk) 05:05, 19 June 2022 (UTC)

@Yvecai: Which part you would recommend to move to extras? Mateusz Konieczny (talk) 11:09, 21 June 2022 (UTC)

A place to drop of books does imply a place to pickup books…

That would be an amenity = library as stated in the proposal. If anything your opposing vote might indicate that more tagging options might be needed for self service pickup options that can be on or offsite of the library.

I would however state that 99% of the time the pickup location for books would be the library itself and pre-Covid 99% of all locations that offered drop off options still required going to the library itself to pickup and drop off books. Hence why this has been the focus for of the tagging proposal.

I respect your vote and thank you for your feedback. I just wanted to offer this perspective and see if I’m misunderstanding or understanding your objection so that if I need to revisit this for future votes I know I’m addressing all concerns.

Thanks, JPinAR JPinAR (talk) 11:36, 26 June 2022 (UTC)

@JPinAR: "That would be an amenity = library as stated in the proposal" - where it is stated? "99% of all locations that offered drop off options still required going to the library itself" - I agree, but some places have such cases and leaving such obvious gaps almost always leads to misuse of approved tags Mateusz Konieczny (talk) 11:38, 26 June 2022 (UTC)

Thanks I will consider a follow-up pickup location tag especially since now pickup via kiosk has become a thing. Thank you for the constructive feedback and community contribution. JPinAR (talk) 13:03, 26 June 2022 (UTC)

@JPinAR: That would be really helpful! It would be nice to avoid repeat of say historic=wayside_shrine that is used for all shrines, not only wayside ones. I managed to push man_made=cross forward but it was likely too late to solve historic=wayside_cross from being used for all crosses Mateusz Konieczny (talk) 13:05, 26 June 2022 (UTC)

Getting compound key documentation from the MediaWiki API

Hi, this is sort of tangential to the discussions we're having over in Talk:Wiki, so I'm splitting this off here for convenience. It sounds like you're building an interesting tool in Python. I'm surprised that you're finding it necessary to scrape user-facing pages, even from Python. If something about this wiki led you to that approach, versus something more structured, please let the administrators know so we can look into improvements. In general, scraping should be a last resort, so that in the future we don't end up constrained by what's essentially tagging writing for the renderer scraper.

I was just going back and double-checking one of my suggestions to make sure I wasn't misleading you. The following code is a port of the compound key description stuff in Module:Tag – everything but the language name fallback. It requires the Requests package, but there's also a built-in json module if you need to work with a different HTTP client library. I wouldn't be surprised if this runs faster than scraping the 404 page using something like BeautifulSoup.

#!/usr/bin/env python3

from itertools import product
import requests

key_parts = "construction:turn:lanes:both_ways".split(":")
# Enumerate all possible slices of the key, putting longer slices before their subslices.
key_part_range = range(len(key_parts))
key_part_slices = [key_parts[s:e] for s, e in product(key_part_range, reversed(key_part_range)) if e > s]
# Convert the slices into article titles.
titles = ["Key:{0}".format(":".join(s)) for s in key_part_slices]

# Query Wikibase for the data items' descriptions.
params = {
    "action": "wbgetentities",
    "format": "json",
    "sites": "wiki",
    "titles": "|".join(titles),
    # The OSM key name is always stored as the English label.
    # It needs to be part of the response so we can associate keys with their descriptions.
    "props": "labels|descriptions",
    "languages": "en",
}
request = requests.get("https://wiki.openstreetmap.org/w/api.php", params=params)
response = request.json()

# Annotate the key parts with their descriptions.
remaining_key_parts = key_parts
items = []
for qid, entity in response["entities"].items():
    # Omit missing data items.
    if type(qid) != str or "missing" in entity:
        continue
    label = entity["labels"].get("en") and entity["labels"]["en"]["value"]
    description = entity["descriptions"].get("en") and entity["descriptions"]["en"]["value"]
    key_part_slice = label.split(":")
    # TODO: Actually search remaining_key_parts for a common slice.
    if description and key_part_slice == remaining_key_parts[0:len(key_part_slice)]:
        del remaining_key_parts[0:len(key_part_slice)]
        items.append((qid, label, description))

# Output the list of key parts.
for item in items:
    print("* {1}: {2} ({0})".format(*item))

Output:

* construction: Used together with the higher-level tags like highway/building=construction to describe the type of feature which is currently under construction. (Q172)
* turn:lanes: A diagram of the turn lane indications on a one-way road. Each lane is represented by a direction such as left, through, or right, and the lanes are separated by vertical bars. Use key:turn:lanes:forward and key:turn:lanes:backward on a two-way road. (Q796)

The code is a bit dense, but it comes pretty close to the original Lua veresion, so let me know if you have any questions about how it works. Hope this helps.

 – Minh Nguyễn ^💬 06:18, 25 July 2022 (UTC)

@Minh Nguyen:

Thanks!

"I'm surprised that you're finding it necessary to scrape user-facing pages" - in this case I want to verify existence of user-facing pages for https://github.com/streetcomplete/StreetComplete/discussions/3442 and to lesser extent https://github.com/streetcomplete/StreetComplete/issues/4225 - I want to link specific key pages where documentation exists. So I want to catch cases where all necessary building blocks exist but for some reason user generated compound page is missing, such as https://wiki.openstreetmap.org/w/index.php?title=Key:name:mos where compound page is not displayed

Right now specifically it is a Kotlin code (for https://github.com/streetcomplete/StreetComplete/issues/4225 ), but it would be easy to adapt. Still, this code cares what is shown to user - so checking generated pages is deliberate. After all, even if data items are listed and compound page lister is broken or disabled by design on some pages: it is still not shown to users. Mateusz Konieczny (talk) 06:48, 25 July 2022 (UTC)

OK, the scraping should be workable as long as it isn't coupled so tightly that StreetComplete would be broken by routine formatting changes. Others have been tinkering with the site stylesheets and banner templates lately, so it's only a matter of time before the 404 page gets some interior decoration too. :^) I guess the infobox needs classes/IDs/microformats too. It'll be interesting to see how this feature turns out. iD took a very different approach by hitting the MediaWiki API for data items and displaying that information inline. However, it doesn't have any compound key logic, because arbitrary combinations of key parts tend not to have dedicated fields anyways. Maybe that would change if it ever gains lane-editing functionality like StreetComplete has. – Minh Nguyễn ^💬 07:54, 25 July 2022 (UTC)

@Minh Nguyen:: It is less fragile as the plan is to blindly link wiki pages based on simple rules (do not link values of freeform pages such as width=* or name=*, do not link value pages with cycleway: prefix like cycleway:surface=asphalt, link all building values such as building=office, always link key pages and so on). Right now only name:lang pages are not working as expected (as sometimes compound info is not shown). With script just verifying that linked OSM Wiki pages are actually containing info (useful compound pages or existing pages or redirects leading to an useful documentation page), and not running in app itself. Script above may be useful to test whether compound pages are missing some info and just assume that if this info is present then it will be used. As side effect I reviewed keys used by SC and created for example building=pagoda Mateusz Konieczny (talk) 11:02, 25 July 2022 (UTC)

How about subkeys unrelated to names, like change:lanes:both_ways (Q9383) and seamark:virtual_aton:mmsi (Q20907)? Should the individual parts be listed even though there's a pretty comprehensive description in the infobox? Or should the description be repeated outside of the infobox? – Minh Nguyễn ^💬 00:38, 27 July 2022 (UTC)

@Minh Nguyen: I would also expect listing compound part there, important part is providing links to pages with actual documentation and explanation. Without that I feel that creating OSM Wiki description pages would be substantial improvement and therefore worth doing. I am also considering "dedicated nonempty data item exists for this tag" as reason to create OSM Wiki description page but it is much weaker Mateusz Konieczny (talk) 04:30, 27 July 2022 (UTC)

Doppelt hält besser? ;-)

Special:Diff/2314646/next Gruß --Reneman (talk) 15:31, 1 August 2022 (UTC)

@Reneman: - note that it is NOT a duplicate, this linked tags are different - one with spaces, one with underscores Mateusz Konieczny (talk) 15:33, 1 August 2022 (UTC)

yes the difference in the space I have seen. and what is the point? Why is the infobox needed twice? Shouldn't this difference be pointed out in ONE box? --Reneman (talk) 16:23, 1 August 2022 (UTC)

@Reneman: (1) it is worth mentioning both as if only one would be mentioned then it could be retagged while other would be still present (2) it is bot edit and bot has no special support for this case (if you are interested you can write pull request that would improve handling in such case ) Mateusz Konieczny (talk) 16:29, 1 August 2022 (UTC)

yes it may be worth placing both links. But why in TWO infoboxes? Both links can be placed one after the other in ONE box (avoid redundancies). I have changed the file page as an example for illustration (Proposal Special:Diff/2363763/next). --Reneman (talk) 17:17, 1 August 2022 (UTC)

"But why in TWO infoboxes" it is a bot edit and bot has no special support for this case (if you are interested you can write pull request that would improve handling in such case ). It is not an edit that I have made, it is edit made by fully automated program following very simple instructions. @Reneman: Mateusz Konieczny (talk) 17:33, 1 August 2022 (UTC)

Resolved: Mateusz Konieczny (talk) 06:07, 30 August 2022 (UTC)

Bot account for mass wiki edits

Hey, could you use a separate bot account for mass edits such as special:diff/2367957?

I think it would be better if such edits wouldn't spam Special:RecentChanges (which by default uses the hidebots=1 filter).

--Push-f (talk) 16:48, 7 August 2022 (UTC)

@Push-f: (1) it is actually human reviewed edit (2) at least some editors react to it and change images to a better one, see for example https://wiki.openstreetmap.org/w/index.php?title=IT:Tag:highway%3Dmotorway&curid=33322&action=history . (3) recent large burst likely will not reoccur any time soon

But I will remember this suggestion and if someone else will suggest this I will strongly move to switching this to unreviewed bot edit

Mateusz Konieczny (talk) 17:08, 7 August 2022 (UTC)

@Push-f: I switched some fully automated edits (generation of image listings in my userspace) into bot edits. This edits are actually all human reviewed (and sometimes wrong) so hiding them as bot edits is IMHO a bad idea and as promised I do them less often nowadays Mateusz Konieczny (talk) 19:06, 7 September 2022 (UTC)

Re: Missing file information (barrierswinggate.jpg)

Hello Mateusz,

Thank you for reaching out. My apologies. I was just testing the Wiki upload image feature since my colleagues who just recently joined the OSM Wiki are unable to upload a picture. Thus I tried to see if I got the same issue or not. How can I delete the image?

Best, Yantisa

@Yantisa: - you can use {{Delete}}. I did it for you in https://wiki.openstreetmap.org/w/index.php?title=File:Barrierswinggate.jpg&diff=2382114&oldid=2381205 Note that waiting time for deletion is about two months - but do not worry about it Mateusz Konieczny (talk) 06:07, 30 August 2022 (UTC)

Resolved: Mateusz Konieczny (talk) 19:03, 7 September 2022 (UTC)

Re: Missing file information (dónal image uploads)

Hi Mateusz! I've added license information to the images I was notified about (see below) and removed the "Unknown" template. Do I need to do anything additional?

• File:Fuel-truck-pngrepo-com.png
• File:Telco_Mast_and_hut_-_Co._Cork,_Ireland.jpg

@Dónal: everything looks fine! Mateusz Konieczny (talk) 21:09, 7 September 2022 (UTC)

note to self: consider export to Commons Mateusz Konieczny (talk) 21:09, 7 September 2022 (UTC)

Uploaded image directly from editing

Hi Mateusz, I uploaded an image directly from the article I am writing, but it did not asked me about the license. How can I correct this? I know you are working on this.

This is the image: https://wiki.openstreetmap.org/wiki/File:Notathon-Ecuador-Cayambe.svg

@Angoca: - that is annoying one. For start, have you made this logo or reused existing file? Mateusz Konieczny (talk) 17:31, 8 September 2022 (UTC)

I used a OSM background, but the rest I did from scratch. AngocA (talk) 02:11, 9 September 2022 (UTC)

@Angoca:

Oh, not so bad. Then you need

{{ODbL OpenStreetMap}} and {{CC0-self}} + "I used a OSM background, but the rest I did from scratch" clarification

or {{ODbL OpenStreetMap}} and {{CC-BY-SA-4.0|Angoca}} + "I used a OSM background, but the rest I did from scratch" clarification

use CC0 if you want people to be able to use/modify it freely without worrying about complications such as to where one should put credit on poster or whether CC license materials are OK to post on Facebook and so on

use CC-BY-SA-4.0 if you want to demand that people credit you and you are willing to take legal action to enforce it. Personally for my photos and work like this I use CC0 as unlike for OSM editing and my programming projects I am not planning to take legal action, so making it easy for everyone seem better

Other licences are also possible, but for nearly all uses this two will be preferable for media

Mateusz Konieczny (talk) 10:59, 9 September 2022 (UTC)

access=private on road only ?

I don't see why you removed the documentation about the (low but) usage of access=private of other stuffs than road, for ex a parking https://www.openstreetmap.org/way/357669359 a gate https://www.openstreetmap.org/node/10007918014 a entrance https://www.openstreetmap.org/node/6845731522 Marc marc (talk) 23:03, 9 September 2022 (UTC)

It would be probably better to comment on talk page of article - I have no idea what edit you refer to. Can you link it? If you mean https://wiki.openstreetmap.org/w/index.php?title=Tag:access%3Dprivate&diff=prev&oldid=2389350 then it was at least intended to do exact opposite thing Mateusz Konieczny (talk) 08:10, 10 September 2022 (UTC)

File:Mixed fence.png

https://wiki.openstreetmap.org/wiki/File:Mixed_fence.png Nie pamiętam, skąd wytrzasnąłem to zdjęcie, ale Google Photos nie pamięta, żebym je zrobił, poza tym pewnie byłby to JPEG. Najprawdopodobniej wyciąłem clipping toolem skądś. Pewnie będzie bezpieczniej wywalić je i zastąpić innym.

A strange link to the Ca: category

Many Catalan pages have an automatically generated invalid link to the Ca: category in the footer. Can I ask for a possible correction or forwarding of this information to the appropriate places? I couldn't find which template is causing this. --Lenochod (talk) 08:28, 23 September 2022 (UTC)

@Lenochod: - I would ask on Talk:Wiki It looks like it is added by infobox. Maybe it is caused by missing translations in https://wiki.openstreetmap.org/w/index.php?title=Template:DescriptionCategoriesLang&action=edit ? 08:03, 23 September 2022 (UTC)

I added the translation in https://wiki.openstreetmap.org/w/index.php?title=Template:DescriptionCategoriesLang&action=edit and now it doesn't appear anymore. --Lenochod (talk) 12:00, 23 September 2022 (UTC)