Modify

Opened 13 days ago

Closed 13 days ago

Last modified 10 hours ago

#14655 closed defect (fixed)

JOSM / ELI synchronization

Reported by: Don-vip Owned by: Klumbumbus
Priority: normal Milestone: 17.04
Component: Core imagery Version:
Keywords: Cc: stoecker, Klumbumbus

Description

The test is failing for several days now:

Error: at least a red line has been found, see https://josm.openstreetmap.de/wiki/ImageryCompare for details
Displaying only red entries. The test fails if at least one is found
*** URLs found in ELI but not in JOSM (36): ***
-  [BR] Alagoas Litoral 2006 - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Ortofotos%202006&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Arapiraca AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Arapiraca&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Branquinha AL - http://geoserver.dados.al.gov.br:8080/geoserver/Alagoas/ows?service=WMS&version=1.1.0&request=GetMap&layers=Branquinha&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Cajueiro AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Cajueiro&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Campo Alegre AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Campo%20Alegre&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Capela AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Capela&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] DER-ES - http://portal.der.es.gov.br/geoserver/wms?SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.1&SERVICE=WMS&REQUEST=GetMap&LAYERS=deres-ext:TRECHOINFO&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Delmiro Gouveia AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Delmiro%20Gouveia&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Estrela de Alagoas - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Estrela%20de%20Alagoas&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Girau do Ponciano - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Girau%20do%20Ponciano&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Major Isidoro AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Major%20Isidoro&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Mar Vermelho AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Mar%20Vermelho&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Minador do Negrão - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Minador%20do%20Negrao&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Olho d'Agua do Casado AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Olho%20Dagua%20do%20Casado&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Palestina AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Palestina&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Palmeira dos Indios AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Palmeira%20dos%20Indios&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Piranhas AL - wms:http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Piranhas&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Poço das Trincheiras AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Poco%20das%20Trincheiras&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Pão de Açucar AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Pao%20de%20Acucar&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Santana do Ipanema AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Santana%20do%20Ipanema&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] São Miguel dos Campos AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Sao_miguel_dos_campos&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Teotonio Vilela AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Teotonio%20Vilela&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] União dos Palmares AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Uniao%20dos%20Palmares&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
-  [BR] Viçosa AL - http://geo.seplande.al.gov.br/teogc/terraogcwms.cgi?LAYERS=Vicosa&SERVICE=WMS&FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.0&SERVICE=WMS&REQUEST=GetMap&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}

Attachments (1)

wow.PNG (3.1 KB) - added by Don-vip 9 days ago.

Download all attachments as: .zip

Change History (35)

comment:1 Changed 13 days ago by Klumbumbus

Owner: changed from team to Klumbumbus
Status: newassigned

comment:2 Changed 13 days ago by Klumbumbus

Resolution: fixed
Status: assignedclosed

comment:3 Changed 12 days ago by Don-vip

new errors:

Displaying only red entries. The test fails if at least one is found
*** Mismatching shapes: ***
+++ JOSM shape 1 unclosed: [DE] Deutsche Bahn VzG lines Nov 2013 - https://wms.michreichert.de/vzg-strecken?FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.1&SERVICE=WMS&REQUEST=GetMap&LAYERS=vzg_strecken,station_codes,level_crossings&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}
*** Miscellaneous checks: ***
* Bounds do not match shape (is 47.2363,5.53806,55.2817,15.3711, calculated <bounds min-lat='47.391238' min-lon='6.003688' max-lat='54.990006' max-lon='15.039257'>): [DE] Deutsche Bahn VzG lines Nov 2013 - https://wms.michreichert.de/vzg-strecken?FORMAT=image/png&TRANSPARENT=TRUE&VERSION=1.1.1&SERVICE=WMS&REQUEST=GetMap&LAYERS=vzg_strecken,station_codes,level_crossings&STYLES=&SRS={proj}&WIDTH={width}&HEIGHT={height}&BBOX={bbox}

comment:4 in reply to:  3 Changed 12 days ago by stoecker

Replying to Don-vip:

new errors:

Well, these messages already contained the way to fix them in the message itself.

  • Replace the slightly wrong bounds with the calculated ones.
  • Copy the last point to close the polygon

comment:5 Changed 12 days ago by Don-vip

Probably, but I'm not interested in ELI synchronization. I added the test because you asked to :)

comment:6 in reply to:  5 Changed 11 days ago by stoecker

Replying to Don-vip:

Probably, but I'm not interested in ELI synchronization. I added the test because you asked to :)

Hah, I got you. That message doesn't come from the syncing, but from the additional sanity checks implemented in the same script. And you are very well into cleanup and warning fixing. You did this mostly at least the last 2 years...

;-)

comment:7 Changed 11 days ago by Don-vip

:)

comment:8 Changed 9 days ago by bastiK

Commits to ELI come at least weekly, almost daily, many of which require JOSM wiki update. Who feels responsible for ELI synchronization? If no one does, we should disable this test.

Tests should not fail for a long period of time. The person responsible should take note and make an effort to fix it, so the broken tests do not pile up like last month.
https://josm.openstreetmap.de/jenkins/job/JOSM/jdk=JDK8/test/trend

comment:9 Changed 9 days ago by stoecker

While that comment in general is right, here is the wrong target. The ELI issues have been synced mostly by Klumbumbus and sometimes me usually very fast. Only the bunch of 24 Brazilian imagery entries took some days longer.

comment:10 Changed 9 days ago by bastiK

Okay, fair enough.

comment:11 Changed 9 days ago by stoecker

BTW: It would be much easier if the quality of the entries would be good enough to simply copy new entries...

comment:12 Changed 9 days ago by Klumbumbus

I usually try to fix the sync issues quickly. I was a bit demotivated to add the 24 brazilian layers as adding the two czech entries before took me very long (one or two hours), so I let the brazilian layers wait some days :)

Is there a RSS feed or email notification for single Jenkins tests? Using the whole Jenkins RSS feed (https://josm.openstreetmap.de/jenkins/job/JOSM/rssFailed) doesn't really help me as most of the time Jenkins fails anyway as shown in the graphic above.

comment:13 in reply to:  12 Changed 9 days ago by bastiK

Replying to Klumbumbus:

Is there a RSS feed or email notification for single Jenkins tests? Using the whole Jenkins RSS feed (https://josm.openstreetmap.de/jenkins/job/JOSM/rssFailed) doesn't really help me as most of the time Jenkins fails anyway as shown in the graphic above.

Possibly by making it a separate project in Jenkins?

comment:14 Changed 9 days ago by Don-vip

I was thinking about it. It would make sens to extract the integration tests from our main job. The tests that fail very often because of external factor are:

  • plugins integration test
  • IANA TLD integration test
  • ELI synchronization test
  • presets integration test
  • map paint styles integration test
  • taginfo popular tags test

Changed 9 days ago by Don-vip

Attachment: wow.PNG added

comment:15 Changed 9 days ago by Don-vip

@Klumbumbus: Don't despair, you can fix all of them!


comment:16 Changed 9 days ago by stoecker

Hey. Don't spread debug info ... BTW the first number were blue entries (no reds ATM) and the second the data age of the last test :-)

@Klumbumbus: Well, now you should get such an message when there are red entries. Data is updated once a day or when the compare page is called (which is done by the test I think).

@Don-vip, bastiK: As TRAC_ADMINs get all permissions you also get that note if I find no way to check individual permission settings beside the TRAC_ADMIN.

That information could be expanded for other status stuff as well. Ideas?

comment:17 Changed 9 days ago by Klumbumbus

Nice :)

As I'm also interested when some of the other jenkins tests fail, I think I found a solution to monitor individual jenkins test results via https://addons.mozilla.org/en-us/firefox/addon/alertbox/

So if this works as expected (it will take some time to test) the trac message could be removed again if it bothers other team members.

comment:18 Changed 9 days ago by Don-vip

OK done:

  • JOSM job now only runs real unit tests. By real I mean tests that should not fail unless we have changed some Java code.
  • JOSM-Integration job runs integration tests that fail frequently depending on the external changes in OSM ecosystem. It will run every 6 hours and should be fast.

comment:19 in reply to:  18 Changed 8 days ago by bastiK

Replying to Don-vip:

OK done:

  • JOSM job now only runs real unit tests. By real I mean tests that should not fail unless we have changed some Java code.
  • JOSM-Integration job runs integration tests that fail frequently depending on the external changes in OSM ecosystem. It will run every 6 hours and should be fast.

Thanks, this is a great improvement!

comment:20 Changed 8 days ago by stoecker

Aaargh. Again 11 differences and again many errors in ELI - Actually the argument that ELI has peer review is void. There is NO peer review. Broken PRs are simply accepted.

comment:21 Changed 8 days ago by Klumbumbus

@stoecker: regarding

Unjustified large zoom
ELI 	* Maxzoom differs (20 != 19): [CH] Kanton Aargau 25cm (AGIS 2014) - http://mapproxy.osm.ch:8080/tiles/AGIS2014/EPSG900913/{zoom}/{x}/{y}.png?origin=nw
ELI 	* Maxzoom differs (20 != 19): [CH] Kanton Aargau 25cm (AGIS 2016) - http://mapproxy.osm.ch:8080/tiles/AGIS2016/EPSG900913/{zoom}/{x}/{y}.png?origin=nw

according to https://github.com/osmlab/editor-layer-index/pull/308#issuecomment-295737054 the base zoom level is z20

comment:22 in reply to:  21 Changed 7 days ago by stoecker

according to https://github.com/osmlab/editor-layer-index/pull/308#issuecomment-295737054 the base zoom level is z20

Then maybe he made an error in server setup? All places I tried zoom 20 is a simple upscale of 19.

comment:23 Changed 7 days ago by stoecker

@Klumbumbus:

I've seen that the double point I added as ignore was in our data and you fixed it. I though I got rid of them already, that's why I was sure it is an ignore :-)

In r11964 I added another sanity check for the double points and also a first step to get proper formatted XML directly from the GeoJSON, so that an "ant imageryindexdownload" in SVN produces a file directly usable for Copy&Paste. Currently contains only name, id and the shape/bounds, which usually need most handwork. Will add the other parts at the weekend.

comment:24 Changed 7 days ago by Klumbumbus

👍

comment:25 Changed 6 days ago by stoecker

In 11975/josm:

see #14655 - implement nearly all features in ELI sync XML output

comment:26 Changed 6 days ago by stoecker

As I added code to read most properties now anyway - Should I add any additional ELI<->JOSM checks for not yet checked properties? URLs, projections, ...?

comment:27 Changed 6 days ago by Don-vip

In 11977/josm:

see #14655 - fix javadoc warnings

comment:28 Changed 6 days ago by Klumbumbus

projections and description would be nice. All the license/terms_of_use/attribution URLs would be important but I think thats not really possible atm as josm and ELI has different properties.

comment:29 Changed 6 days ago by stoecker

In 11981/josm:

see #14655 - add more ELI - JOSM checks, new results comment by SKIP text (does not trigger error notes)

comment:30 Changed 6 days ago by stoecker

Well, a lot of new output, all marked as SKIP to silence the checks.

Permission-ref and terms-of-use both map to license_url. I'm not sure how the links to generic CC-BY-SA and alike should be handled. Plain text in license_url sounds like "Ignore" for me.

comment:31 in reply to:  23 Changed 3 days ago by Klumbumbus

Replying to stoecker:

proper formatted XML directly from the GeoJSON, so that an "ant imageryindexdownload" in SVN produces a file directly usable for Copy&Paste.

Is this file already somewhere available or do I need to create it myself?

comment:32 Changed 3 days ago by stoecker

You need to call "ant imageryindexdownload" or "int imageryindex" (no new download) in SVN. File is not automatically created somewhere. I don't plan to release it: Neither the ELI not the JOSM output is fully complete.

P.S. When copying the descriptions from ELI I think they should be cleaned. E.g. the imagico entries have ugly information in there, which is not really for human consumption.

comment:33 in reply to:  8 Changed 2 days ago by Klumbumbus

Replying to bastiK:

https://josm.openstreetmap.de/jenkins/job/JOSM/jdk=JDK8/test/trend

This graph is gone at Jenkins :(

Last edited 2 days ago by Klumbumbus (previous) (diff)

comment:34 Changed 10 hours ago by stoecker

@Klumbumbus:

I did some long planned changes this week:

  • The preview now handles also WMS servers not supporting EPSG:4326 and EPSG:3857.
  • There is also a new hidden feature: In the preview you can add "&epsg=" and either 1...x (for the first, second, ... entry) or the epsg number to choose a specific ESPG code to the URL. This allows testing if all EPSGs work on the server.

Modify Ticket

Change Properties
Set your email in Preferences
Action
as closed The owner will remain Klumbumbus.
as The resolution will be set. Next status will be 'closed'.
The resolution will be deleted. Next status will be 'reopened'.

Add Comment


E-mail address and name can be saved in the Preferences.

 
Note: See TracTickets for help on using tickets.