Make max_windows a required argument for RandomWindowGeoDataset

[keves1]

Overview

This PR makes max_windows a required argument for RandomWindowGeoDataset, and removes the option for this argument to be None. The purpose of this is to avoid a runtime error that would occur when iterating through a PyTorch dataloader that uses this dataset if max_windows uses the current default value of None. The PR also makes all arguments other than scene keyword args in order to allow the order of the args to stay the same.

Checklist

• Added unit tests, if applicable
• Updated documentation, if applicable
• Added needs-backport label if the change should be back-ported to the previous release
• PR has a name that won't get you publicly shamed for vagueness

Notes

Updated relevant unit tests and tutorials.

Testing Instructions

See updated unit tests.

Test case (should run with no errors)

from rastervision.core.data import ClassConfig
from rastervision.core.data.utils import make_ss_scene
from rastervision.pytorch_learner import SemanticSegmentationRandomWindowGeoDataset

from torch.utils.data import DataLoader

image_uri = 's3://spacenet-dataset/spacenet/SN7_buildings/train/L15-0331E-1257N_1327_3160_13/images/global_monthly_2018_01_mosaic_L15-0331E-1257N_1327_3160_13.tif'
label_uri = 's3://spacenet-dataset/spacenet/SN7_buildings/train/L15-0331E-1257N_1327_3160_13/labels/global_monthly_2018_01_mosaic_L15-0331E-1257N_1327_3160_13_Buildings.geojson'

class_config = ClassConfig(
    names=['background', 'building'],
    colors=['lightgray', 'darkred'],
    null_class='background')

scene = make_ss_scene(
    class_config=class_config,
    image_uri=image_uri,
    label_vector_uri=label_uri,
    label_vector_default_class_id=class_config.get_class_id('building'),
    image_raster_source_kw=dict(allow_streaming=True))

dataset = SemanticSegmentationRandomWindowGeoDataset(scene, out_size=224, size_lims=[220,224], max_windows=10)

dataloader = DataLoader(dataset, batch_size=5, shuffle=True, num_workers=0)
x, y = next(iter(dataloader))

---

Closes #2222

[AdeelH]

Looks great! Thank you for the PR!

[keves1]

Great! You're welcome, glad to contribute.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 90.34%. Comparing base (eb83d30) to head (dba5ff4).
Report is 2 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #2223      +/-   ##
==========================================
- Coverage   90.34%   90.34%   -0.01%     
==========================================
  Files         199      199              
  Lines        9910     9908       -2     
==========================================
- Hits         8953     8951       -2     
  Misses        957      957              

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.