mapsme / subways Goto Github PK

And also lines with repeating stations. To do that, know stop from platforms and expect these to come in packs, not spread out across the route. Example: PL1-ST1-PL2-ST2-PL1-ST3 is unacceptable (as it is now). Ending with PL1 would be also unacceptable. But PL1-ST1-PL2-ST2-ST1-PL1 would be okay: it marks a circular route. The key is in stops: we don't expect two stops for the same station to be an error. Exception is two consecutive identical stops.

Depends on #15.

Track exits and entrances

Any idea why Dalian Metro Line 12 is not recognized in http://osmz.ru/s ubways/dalian.yaml ?
The route master relation is https://www.openstreetmap.org/relation/7957874

It contains the same network=Dalian Metro tag as the other three lines that are recognized correctly.

Only if there is some amount of tracks, obviously.

So we can find unaccounted metro networks.

Subway validator warns about possible wrong track direction for relation 5623319. But the tracks are "backward" relation member so going the other way direction the stop order is correct.

Subway validator is detecting MTR Island Line and South Island Line
https://www.openstreetmap.org/relation/4432666
https://www.openstreetmap.org/relation/6827211

But not other lines, e.g. Tsuen Wan, Tung Chung, Airport Express, Kwun Tong, West Rail Line, East Rail Line
https://www.openstreetmap.org/relation/269669
https://www.openstreetmap.org/relation/5317704
https://www.openstreetmap.org/relation/5317239
https://www.openstreetmap.org/relation/6452935
https://www.openstreetmap.org/relation/6582966
https://www.openstreetmap.org/relation/4248590

Right now you have to set up your own validation to get yaml and geojson files. It would be much simpler to fix metro networks if they are created for each network in the world. The main issue would be generating a file name for each network and coming up with a good argument name.

We have a track and stop positions. Cut out a segment and calculate its length.

agency_id is now set to zero. It should be calculated somehow: for example, using a hash of a city name.

Orange and Red subway routes were built months ago, but are not being picked up by subway validator. See the following node, which has route relations where the two routes intersect.
https://www.openstreetmap.org/node/4580663526

Also, the Circular light rail line is present as a tram route. It looks more like a tram route than light rail as per wikipedia:
https://en.wikipedia.org/wiki/Circular_light_rail

See following node with the route relation for the circular line, as a tram route
https://www.openstreetmap.org/node/4576225719

Instead of Hammersmith→Paddington→Edgware Road→Paddington→Edgware Road it stops on the second Paddington. The reverse direction is even worse: it contains only Edgware Road→Paddington and stops one station before the second Edgware Road.

In multiple cases, e.g. Mexico, Los Angeles, Cleveland, multiple Asian cities, station counts don't match station counts for several reasons:

• If multiple subways or multiple light rail lines stop at the same station location, they may all be associated with a single station node (or associated stop node) or they may be associated with different station nodes with the same station name.

• If a light rail line and a subway line meet, each requires a separate station node

Counting the number of total stops on all of the applicable routes would be more effective

Duplicate stop "Улица Скобелевская" in route - check stop/platform order (relation 301485, "Бутовская линия")

and

Inconsistent platform-stop "Измайловская" in route (relation 302149, "3")

The Rapid KL have a lot of interchange station which is formerly seperated and the platform are not in a same level.

This is the list, I will update later and I hope somebody to update this table too thank youhttp://osmz.ru/subways/malaysia.html.

KL Sentral (Transport Hub)
Bandar Tasik Selatan
Kajang
Sungai Buloh
Subang Jaya
Titiwangsa
Masjid Jamek
Plaza Rakyat plus Merdeka
Hang Tuah
Maluri
Bukit Bintang

And also can anybody put the KL Monorail route into that Rapid Transit table? In Malaysia, monorail is a type of public transport that almost like subway but only one track.

I want to try, how validation works for a single city, e.g. London. I have tried running
./process_subways.py -c London
but it doesn't work and returns HTTP Error 429: Too Many Requests

However that script uses OverpassAPI, so I guess that I need to download .osm myself. Obviously nobody wants to download entire planet.osm (or planet.pbf) if just a single city is needed. Readme mentions osmconvert tool, but I didn't find anything helpful on https://wiki.openstreetmap.org/wiki/Osmconvert to download data for a city without downloading the planet. It would be great if readme explained that step in more details, maybe providing links that explain data importing in-depth.

http://osmz.ru/subways/usa.html

Stop "190th Street" (pos) is nowhere near the tracks (relation 366774, "NYCS - A Train")

Doesn't make sense because node is on tracks. Also, should report the node rather than node's coordinates.

For now, tracks are only checked for holes. We need to keep the longest stretch of tracks (in km, not in way count). Before and after the stretch, missing parts should be interpolated from stations or stop positions.

Tehran currently has 9-11 interchange stations operational: https://en.wikipedia.org/wiki/Tehran_Metro#Interchange_stations
Please update your reference data at https://docs.google.com/spreadsheets/d/1-UHDzfBwHdeyFxgC5cE_MaNQotF3-Y0r1nW9IwpIEj8/edit#gid=2085714915 to include this figure.

So we can have transfers from one network to another (e.g. in Tokyo).

Tehran metro line 8 (route=subway) doesn't seem to be picked up properly: https://www.openstreetmap.org/relation/6151834
All stations along the way keep getting flagged as unused.

Does the whole route with all its way segments and nodes need to be completely within the bounding box?
I can't find any other reason - particularly tagging - which would prevent it.

Also fill these from route_master.

Actually Kuala Lumpur has only one MRT (subway) and 3 LRT (light rail) Lines (and another one monorail). But the project website-http://osmz.ru/subways/index.html display 4 subway and 0 light rail line. Hope somebody can fix it.

So that we can find ABDCEF errors.

Greetings. I have updated the information of the Lima metro nodes. I have divided into whereabouts, the platform is no longer a building, I added track platforms, and so on. It would be good to update this information as soon as possible.
https://wiki.openstreetmap.org/wiki/Metro_de_Lima

Somehow we lost stop_area_group relations.

For a station, the center should be center of its stops and platforms.

See http://osmz.ru/subways/uk.html — London Docklands
Count only interchanges that have at least two stations belonging to network routes.

There is a placeholder in mapsme_subways, it needs to be implemented.

Station node 1248716467 is listed in two stop_areas, first one: (node 493290713, "Каширская")
https://www.openstreetmap.org/node/1248716467
https://www.openstreetmap.org/node/493290713

Every good city should be cached. When loading from a cache (in case a network turned bad), all osm ids should be checked and entities with obsolete ids should be removed. Only if all stations remain and have at least one entrance and exit, the cached network should be used.

When a station has several stop_area, it is now counted as many stations.

http://osmz.ru/subways/china.html
e.g.
"Stop area has multiple stations (relation 2207110, "郭公庄")"

This entry with identical wording shows up six times. It would be more helpful to list the relation, with a list of the nodes for different stations in the stop_area.

http://osmz.ru/subways/mexico.html

For Monterrey, reports "Tracks seem to go in the opposite direction to stops (relation 7890222, "Línea 1") "

Not possible for track to be in the opposite direction because the route consists of a single way

http://osmz.ru/subways/philippines.html

All subway lines and all light rail lines are being counted in both "Manila" and "Manila - Light Rail"

A→B→C→D→E→F is okay

A→B→C→B→C→D→E→F is not.

Each route_master should have at least 2 routes (except circular), plus each route has to have a barward copy (warning).

red → #ff0000 etc.

Currently we only compare route relations regardless of route_masters.

Parsing a log is stupid. It should be a JSON with all the source data, so the to_html script doesn't need parsing the source spreadsheet.

Store entrances (that station share) separately, and check that platforms and stop positions are in only one stop_area. After that maybe support multiple stop_areas with a single station, and stop_area_groups on these.

A station is too general: you don't know where to go to board a specific train. Thus, a route station should be a namedtuple with a station object, a platform element and a stop position as a coordinate. The latter should be filled either from a stop_position node, or a platform center, or a station node, better projected onto a railway line.

There is "S-Bahn Mitteldeutschland" in the Leipzig-Halle region in Germany with 10 lines and 141 stations. Including a tunnel section under the city center: https://en.wikipedia.org/wiki/S-Bahn_Mitteldeutschland
None of the infrastructure is subway, but everything light rail. Are you intending to support these networks as well?
If yes: How would I request it to be added to the spreadsheet?

What is meant by the "unused station" warning? For example in Germany every city gets a lot of these warnings. All the stations I checked were in a stop area relation and used by some route.

Seoul Neotrans has one subway line, but it is not being recognized by subway validator
Here is the routemaster:
https://www.openstreetmap.org/relation/7728256

http://osmz.ru/subways/malaysia.html

Warning states:
"31 subway entrances are not in stop_area relations" but no nodes listed

Include them in counts, but not in actual routes. Flag them like .active = False.

http://osmz.ru/subways/brazil.html

For Rio de Janeiro
6 unused stations: n4317159358, n4317159359, n4317159360, n4317159361, n5302148205, n5362542035

Yet these stations are members of two Linha 4 relations, one in each direction

This is needed for easy searching of missing or extra stations. The format should be either plain text or yaml.

If there is a stop_area, uid should be its relation id, otherwise — station's id.