get_sequences finished but doesn't save to file?

[thibaultmol]

script seems to finish but not save to file. seq_dataset.csv is empty (only hasorg_id,seq_id

Operating System: Garuda Linux
KDE Plasma Version: 5.27.2
KDE Frameworks Version: 5.103.0
Qt Version: 5.15.8
Kernel Version: 6.2.2-zen1-1-zen (64-bit)
Graphics Platform: X11
Processors: 16 × AMD Ryzen 7 3700X 8-Core Processor
Memory: 31,3 GiB of RAM
Graphics Processor: NVIDIA GeForce GTX 1080/PCIe/SSE2

[gchoumos]

Hello @thibaultmol !

Thanks for using the scripts and reporting this issue! I just tried this for a small bounding box and it worked for my particular settings searching for sequences contributed to a specific organization.

Are you trying to search/filter against a specific organization id? Or a user id?

In case you doing the latter, it seems that the user id is not provided by the Mapillary API v4, so attempts to use it for searching will lead to no results unfortunately. I will have to see if there is any workaround to adapt the scripts accordingly. A "hacky" solution has been mentioned here:
https://forum.mapillary.com/t/bugreport-image-author-info-unavailable-in-the-new-api/4994/3

Thanks,
George

[thibaultmol]

Hi, yeah I was indeed trying to download from a specific user id

[thibaultmol]

So yeah George i already use the method from Trekview to turn the username string into the user ID. if that's what you're referring to you. But that seems to be yieling no results.

User id 107745295202137 (laps360) is the user i'm trying to download

[thibaultmol]

Actually, it seems to be broken for org's as well?

WVI brugge has the org id: 1111421462997936
but that doesn't actually save the sequences either.

https://www.mapillary.com/app/org/wvibrugge

[gchoumos]

Hello @thibaultmol !

Apologies, haven't yet managed to look at the user id issue and apply a workaround, but I hope I will be able to do it soon.

With regards to the organization ID, I actually gave it a try and got some results!

Have you verified you apply the correct settings for the org_id you are searching against? Also, have you checked the latitude and longitude of your search are correct?

As an example I considered this bounding for which the organization you mentioned has imagery:

So in settings.py, if I set:

• ll_lat to 50.837
• ll_lon to 2.732
• ur_lat to 50.854
• ur_lon to 2.762
• org_id to 1111421462997936

I do get the following result (also attached file):

seq_dataset.csv

Alternatively, you can create a redacted.py file on the directory of the script, and use the following code (for the example) which will override the corresponding settings:

""" settings overrides """
from settings import SETTINGS

SETTINGS['AOI_bbox']['ll_lat'] = 50.837
SETTINGS['AOI_bbox']['ll_lon'] =  2.732
SETTINGS['AOI_bbox']['ur_lat'] = 50.854
SETTINGS['AOI_bbox']['ur_lon'] =  2.762
SETTINGS['org_id'] = 1111421462997936

[thibaultmol]

oh sorry, I'll try that, I forgot to change the bbox size. Is that a requirement actually? It would be nice if I could just download all of them regardless of the bbox ALSO is there a max size on the bbox?

[gchoumos]

Hey Thibault!

By the way, I really appreciate that you use the scripts and provide all this useful feedback. I will definitely take it into consideration and add a few improvements on my backlog, both in terms of how the code works, as well as in the documentation!

Now regarding the bbox, for the moment it is indeed a requirement. I don't remember the exact reason why I initially did it this way. It probably was because without a bbox and a specific organization, the amount of requests and data was massive. Or maybe because there was limited documentation available on Mapillary side when I started writing these! However, I will create a way to call this without the bbox filter (if the Mapillary API v4 allows for this!).

Not sure if there is a bbox size limitation, haven't tried it yet. If there is one, it should be on the Mapillary side, either as a sanity check, or as a kind of protection against very large requests. But maybe to include the whole map you could try something like:

""" settings overrides """
from settings import SETTINGS

SETTINGS['AOI_bbox']['ll_lat'] = -90
SETTINGS['AOI_bbox']['ll_lon'] = -180
SETTINGS['AOI_bbox']['ur_lat'] =  90
SETTINGS['AOI_bbox']['ur_lon'] = 180
SETTINGS['org_id'] = 1111421462997936

Or something less on the edge:

""" settings overrides """
from settings import SETTINGS

SETTINGS['AOI_bbox']['ll_lat'] = -75
SETTINGS['AOI_bbox']['ll_lon'] = -160
SETTINGS['AOI_bbox']['ur_lat'] =  75
SETTINGS['AOI_bbox']['ur_lon'] = 160
SETTINGS['org_id'] = 1111421462997936

Haven't tested this yet though!