train_test_split fails

[tadolphi]

• detectree version: 0.5.1
• Python version: 3.11.9 64 bit
• Operating system: win 11

Description

i wanted to train a model, but train_test_split fails

What I Did

import detectree as dtr
random_state = 10

img_dir = "training_tiles"

ts = dtr.TrainingSelector(img_dir=img_dir)

split_df, evr = ts.train_test_split(
    method="cluster-II", n_components=12, return_evr=True
)
print(split_df)
print(evr)

here is the error message:

    return self._descr_feature_matrix
           ^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'TrainingSelector' object has no attribute '_descr_feature_matrix'. Did you mean: 'descr_feature_matrix'?

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "d:\PythonProjects\detecttrees\tree_training.py", line 8, in <module>
    split_df, evr = ts.train_test_split(
                    ^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xxx\AppData\Local\Programs\Python\Python311\Lib\site-packages\detectree\train_test_split.py", line 197, in train_test_split
    X = self.descr_feature_matrix
        ^^^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xxx\AppData\Local\Programs\Python\Python311\Lib\site-packages\detectree\train_test_split.py", line 140, in descr_feature_matrix
    self._descr_feature_matrix = np.vstack(feature_rows)
                                 ^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\xxx\AppData\Local\Programs\Python\Python311\Lib\site-packages\numpy\_core\shape_base.py", line 291, in vstack
    return _nx.concatenate(arrs, 0, dtype=dtype, casting=casting)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
ValueError: need at least one array to concatenate

[martibosch]

Hello! this issue is most likely because there are no ".tif" images in the provided img_dir. You can either provide the list of images explicitly, e.g., ts = dtr.TrainingSelector(img_filepaths=<your list-like of images here>) or provide img_filename_pattern if the images have another extension, e.g., for jpeg images this would be ts = dtr.TrainingSelector(img_dir=img_dir, img_filename_pattern="*.jpeg").

Feel free to reopen if this is not the issue.