Pre-submission Checklist
Component / Module
Layer Generation (/computing)
Severity
Low (minor inconvenience, workaround exists)
Bug Description
While reviewing change_deforestation_afforestation(), I noticed that the temporal smoothing step explicitly operates on raw class 3.
Based on the current class mapping, raw class 3 corresponds to a water category, while raw class 6 corresponds to tree/forest.
Since this logic exists inside the deforestation/afforestation workflow, I would like clarification regarding the intended behavior.
Steps to Reproduce
cond1 = (
before.eq(3)
.And(middle.neq(3))
.And(after.eq(3))
)
cond2 = (
before.neq(3)
.And(middle.eq(3))
.And(after.neq(3))
)
middle = middle.where(cond1, 3)
middle = middle.where(cond2, before)
Class Reference
| Raw Class |
Meaning |
| 3 |
Water (Kharif + Rabi) |
| 6 |
Tree / Forest |
Expected Behavior
In change_deforestation_afforestation() the temporal smoothing is expected to be of classes related to forest.
Actual Behavior
The smoothing logic restores or removes occurrences of raw class 3.
Because this occurs before the later remapping and transition analysis steps, I would like to confirm whether the focus on class 3 is intentional.
Environment Details
OS: Ubuntu 24.04 (WSL2)
Python version: 3.10.x (Conda environment)
Django version: 5.2.x
Celery version: 5.6.x
Deployment: Local development environment (WSL2 + Docker services)
Logs / Error Messages
No response
Screenshots / Screen Recordings
No response
Additional Context
This issue is being opened for clarification and validation by the developer/scientific team.
During review of the change detection pipeline, I observed the behavior described above. I am not asserting that this is a confirmed bug. The goal of this issue is to determine whether the current behavior is intentional, documented, and scientifically correct, or whether it represents an implementation issue.
Pre-submission Checklist
Component / Module
Layer Generation (/computing)
Severity
Low (minor inconvenience, workaround exists)
Bug Description
While reviewing change_deforestation_afforestation(), I noticed that the temporal smoothing step explicitly operates on raw class 3.
Based on the current class mapping, raw class 3 corresponds to a water category, while raw class 6 corresponds to tree/forest.
Since this logic exists inside the deforestation/afforestation workflow, I would like clarification regarding the intended behavior.
Steps to Reproduce
cond1 = (
before.eq(3)
.And(middle.neq(3))
.And(after.eq(3))
)
cond2 = (
before.neq(3)
.And(middle.eq(3))
.And(after.neq(3))
)
middle = middle.where(cond1, 3)
middle = middle.where(cond2, before)
Class Reference
Expected Behavior
In change_deforestation_afforestation() the temporal smoothing is expected to be of classes related to forest.
Actual Behavior
The smoothing logic restores or removes occurrences of raw class 3.
Because this occurs before the later remapping and transition analysis steps, I would like to confirm whether the focus on class 3 is intentional.
Environment Details
OS: Ubuntu 24.04 (WSL2)
Python version: 3.10.x (Conda environment)
Django version: 5.2.x
Celery version: 5.6.x
Deployment: Local development environment (WSL2 + Docker services)
Logs / Error Messages
No response
Screenshots / Screen Recordings
No response
Additional Context
This issue is being opened for clarification and validation by the developer/scientific team.
During review of the change detection pipeline, I observed the behavior described above. I am not asserting that this is a confirmed bug. The goal of this issue is to determine whether the current behavior is intentional, documented, and scientifically correct, or whether it represents an implementation issue.