My AOI always fails in "spectral_indices" notebook

[gussabina]

Topic

Notebook

Ask away!

Hello:
I'm successfully running the example for "spectral_indices" notebook, but when I try to adapt it to my case, it always fails to pass the s2.filter step (in run.monitor()).

I'm defining the area of interest as a Polygon:

geom = shpg.Polygon([[-119.736327, 46.288830], [-119.732203, 46.288777], [-119.736303, 46.287675], [-119.732268,46.287648]])
time_range = (datetime(2024, 5, 1), datetime(2024, 8, 11))

And I can verify the area as:

print(geom.wkt)
POLYGON ((-119.736327 46.28883, -119.732203 46.288777, -119.736303 46.287675, -119.732268 46.287648, -119.736327 46.28883))

My questions are:

• is there something wrong with the way I'm defining the Polygon?
• Is there any minimum area to consider in this case?

The example uses a Point but I'm not sure how it determines the AOI in this case, which seems to also be a rectangle. Since I would like to have more control on the AOI, I prefer to use the Polygon to define it:

geom = shpg.Point(-119.21896203939313, 46.44578909859286).buffer(0.05, cap_style=3)

I would appreciate any help.

Thanks
Gus

[rafaspadilha]

Hi, @gussabina. Could you provide the output of run.reason when the workflow fails?
Is it failing for other time ranges as well?

[gussabina]

Hello @rafaspadilha
While trying to run it again to get the run.reason, it now gets stuck as 'queued' and it does not move forward:
🌎 FarmVibes.AI 🌍 Custom: 'indices' 🌏
Run name: Spectral indices
Run id: 55fc5e27-bde8-474d-82ff-96e358bd7490
Run status: queued
Run duration: 00:10:12
┏━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ Task Name ┃ Status ┃ Start Time ┃ End Time ┃ Duration ┃ Progress ┃
┡━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━┩
│ lswi.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ ndwi.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ reci.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ ndre.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ ndmi.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ evi.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ ndvi.compute_index │ pending │ N/A │ N/A │ 00:00:00 │ │
│ s2.merge │ pending │ N/A │ N/A │ 00:00:00 │ │
│ s2.group │ pending │ N/A │ N/A │ 00:00:00 │ │
│ s2.download │ pending │ N/A │ N/A │ 00:00:00 │ │
│ s2.filter │ queued │ N/A │ N/A │ 00:00:00 │ ━━━━━━━━━━━━━━━━━━ 0/1 │
│ s2.list │ done │ 2024/08/19 18:40:45 │ 2024/08/19 18:40:45 │ 00:00:00 │ ━━━━━━━━━━━━━━━━━━ 1/1 │
└────────────────────┴─────────┴─────────────────────┴─────────────────────┴──────────┴───────────────────────────┘
Last update: 2024/08/19 18:50:58 UTC
Do you know the reason for this to get stuck? Also, how to move forward from this point?

Thanks
Gus

[rafaspadilha]

Does restarting the cluster get your workflow out of the queued state?
We also updated the rabbitqm and redis images on dev recently. Could you retry with the latest commit on dev?