TypeError in evaluation stage: Box.__init__() takes 5 positional arguments but 6 were given

[NoaMills-USDA]

I'm running a semantic segmentation model using rastervision 0.20 and the docker image with tag: bff97f3. I'm running into the following error at the evaluation stage. Any ideas why this is happening?

Running eval command...
2023-07-31 14:59:54:rastervision.core.rv_pipeline.rv_pipeline: INFO - Skipping scene group "test_scenes". Empty scene group.
2023-07-31 14:59:54:rastervision.core.rv_pipeline.rv_pipeline: INFO - Running SemanticSegmentationEvaluator on scene group "validation_scenes"...
2023-07-31 14:59:58:rastervision.core.evaluation.classification_evaluator: INFO - Computing evaluation for scene benton...
2023-07-31 15:00:07:rastervision.core.data.label_store.semantic_segmentation_label_store: INFO - Loading scores...
Traceback (most recent call last):
  File "/opt/conda/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/opt/conda/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 251, in <module>
    _main()
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 247, in _main
    main()
  File "/opt/conda/lib/python3.10/site-packages/click/core.py", line 1130, in __call__
    return self.main(*args, **kwargs)
  File "/opt/conda/lib/python3.10/site-packages/click/core.py", line 1055, in main
    rv = self.invoke(ctx)
  File "/opt/conda/lib/python3.10/site-packages/click/core.py", line 1657, in invoke
    return _process_result(sub_ctx.command.invoke(sub_ctx))
  File "/opt/conda/lib/python3.10/site-packages/click/core.py", line 1404, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/opt/conda/lib/python3.10/site-packages/click/core.py", line 760, in invoke
    return __callback(*args, **kwargs)
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 236, in run_command
    _run_command(
  File "/opt/src/rastervision_pipeline/rastervision/pipeline/cli.py", line 218, in _run_command
    command_fn()
  File "/opt/src/rastervision_core/rastervision/core/rv_pipeline/rv_pipeline.py", line 230, in eval
    evaluator.process(group_scenes, self.tmp_dir)
  File "/opt/src/rastervision_core/rastervision/core/evaluation/classification_evaluator.py", line 34, in process
    evaluation = self.evaluate_scene(scene)
  File "/opt/src/rastervision_core/rastervision/core/evaluation/semantic_segmentation_evaluator.py", line 23, in evaluate_scene
    predictions = scene.label_store.get_labels()
  File "/opt/src/rastervision_core/rastervision/core/data/label_store/semantic_segmentation_label_store.py", line 140, in get_labels
    return self.get_scores()
  File "/opt/src/rastervision_core/rastervision/core/data/label_store/semantic_segmentation_label_store.py", line 187, in get_scores
    extent=Box(0, 0, *score_arr.shape),
TypeError: Box.__init__() takes 5 positional arguments but 6 were given
make: *** [/opt/data/test_2/Makefile:12: 2] Error 1

Here is my get_config:

# flake8: noqa

from rastervision.core.rv_pipeline import *
from rastervision.core.backend import *
from rastervision.core.data import *
from rastervision.pytorch_backend import *
from rastervision.pytorch_learner import *
import albumentations as A
import sys
print(sys.version)
import pandas as pd

#from rastervision.pytorch_backend.examples.semantic_segmentation.utils import (
#    example_multiband_transform, example_rgb_transform)


def get_config(runner, root_uri, learn_rate, epoch, cyc, multiband: bool = False, smooth = True) -> SemanticSegmentationConfig:
    input_uri = '/opt/data/input_data'
    counties = pd.read_csv('/opt/data/input_data/counties.csv', index_col="index", dtype={"fips":str})
    print("NUMBER OF COUNTIES: " + str(counties.shape[0]))
    

    counties["image_uri"] = input_uri+"/"+counties["county"]+"_raster.tif"
    counties["label_uri"] = input_uri+"/"+counties["county"]+"_vector.geojson"

    validation_counties = counties.loc[(counties["set"] == "val")]
    train_counties = counties.loc[(counties["set"] == "train")]
    
    avg_train_weight = train_counties["Percent_ag"].mean()
    non_ag_weight = avg_train_weight/(float(1)-avg_train_weight)
    weight_3 = float(1)

    if multiband:
        # use all 4 channels
        channel_order = [0, 1, 2, 3]
        channel_display_groups = {'RGB': (0, 1, 2), 'IR': (3, )}
    else:
        # use infrared, red, & green channels only
        channel_order = [3, 0, 1]
        channel_display_groups = None

    transforms = A.Compose([A.Flip(), A.Transpose(), A.RandomRotate90()])
    aug_transform = A.to_dict(transforms)

    class_config = ClassConfig(
        names=['agricultural', 'non-agricultural'], colors=['red', 'black']) #, null_class="non-agricultural"

    def make_scene(scene_id: str, image_uri: str,
                   label_uri: str, aoi: str) -> SceneConfig:
        """
        - The GeoJSON does not have a class_id property for each geom,
          so it is inferred as 0 (ie. building) because the default_class_id
          is set to 0.
        - The labels are in the form of GeoJSON which needs to be rasterized
          to use as label for semantic segmentation, so we use a RasterizedSource.
        - The rasterizer set the background (as opposed to foreground) pixels
          to 1 because background_class_id is set to 1.
        """
        raster_source = RasterioSourceConfig(
            uris=[image_uri], channel_order=channel_order)
        vector_source = GeoJSONVectorSourceConfig(
            uris=[label_uri],
            transformers=[
            ClassInferenceTransformerConfig(
                default_class_id=class_config.get_class_id('agricultural'))
            ],
            ignore_crs_field=True) #removed default_class_id=0 for 0.20
        label_source = SemanticSegmentationLabelSourceConfig(
            raster_source=RasterizedSourceConfig(
                vector_source=vector_source,
                rasterizer_config=RasterizerConfig(background_class_id=1)))
        label_store = SemanticSegmentationLabelStoreConfig(
            smooth_output=smooth, smooth_as_uint8=smooth)
        if aoi == None:
            aoi_uris=None 
        else:
            aoi_uris=[aoi]
        return SceneConfig(
            id=scene_id,
            raster_source=raster_source,
            label_source=label_source,
            label_store=label_store,
            aoi_uris=aoi_uris)

    train_scenes = []
    for i in range(len(train_counties)):
        train_scenes.append(make_scene(train_counties["county"].iloc[i], train_counties["image_uri"].iloc[i], train_counties["label_uri"].iloc[i], aoi=None))
	
    validation_scenes = []
    for i in range(len(validation_counties)):
        validation_scenes.append(make_scene(validation_counties["county"].iloc[i], validation_counties["image_uri"].iloc[i], validation_counties["label_uri"].iloc[i], aoi=None))

    scene_dataset = DatasetConfig(
        class_config=class_config,
        train_scenes=train_scenes,
        validation_scenes=validation_scenes
    )

    # Use the PyTorch backend for the SemanticSegmentation pipeline.
    chip_sz = 200
    class_config.ensure_null_class()
    backend = PyTorchSemanticSegmentationConfig(
        data=SemanticSegmentationGeoDataConfig(
            scene_dataset=scene_dataset,
            plot_options=PlotOptions(channel_display_groups=channel_display_groups),
            img_sz=chip_sz,
            aug_transform = aug_transform,
            window_opts=GeoDataWindowConfig(
                method=GeoDataWindowMethod.random,
                size=chip_sz,
                size_lims=(chip_sz, chip_sz + 1),
                max_windows=10)),
        model=SemanticSegmentationModelConfig(
                external_def=ExternalModuleConfig(
                    uri="/opt/data/src/pytorch-fpn",
                    name="fpn",
                    entrypoint="make_fpn_resnet",
                    entrypoint_kwargs={
                        'name':'resnet50',
                        'fpn_type':'panoptic',
                        'num_classes':len(class_config),
                        'in_channels':len(channel_order),
                        'out_size':(chip_sz, chip_sz)
                    })),
        solver=SolverConfig(lr=learn_rate, num_epochs=epoch, batch_sz=8, one_cycle = cyc, class_loss_weights=[1., non_ag_weight, weight_3]), 
        log_tensorboard = True)

    return SemanticSegmentationConfig(
        root_uri=root_uri,
        dataset=scene_dataset,
        backend=backend,
        train_chip_sz=chip_sz,
        predict_chip_sz=chip_sz)

[AdeelH]

This is a bug. Thanks for catching that!

I'll make a PR tomorrow but if you need a quick fix, you can change lines 186-188 in semantic_segmentation_label_store.py to

        _, h, w = score_arr.shape
        labels = SemanticSegmentationSmoothLabels(
            extent=Box(0, 0, h, w), num_classes=len(self.class_config))

[NoaMills-USDA]

Wonderful, thank you so much!

[AdeelH]

Fixed in #1843