[model:bugtype] ValueError: y should be a 1d array, got an array of shape (198635, 5) instead. (hashtag3956)

[vikramsubramanian]

I'm wondering if the error encountered during the training of the bugtype model is something we should investigate.

Traceback (most recent call last):
  File "/home/promisefru/anaconda3/envs/bugbug/lib/python3.10/runpy.py", line 196, in _run_module_as_main
    return _run_code(code, main_globals, None,
  File "/home/promisefru/anaconda3/envs/bugbug/lib/python3.10/runpy.py", line 86, in _run_code
    exec(code, run_globals)
  File "/home/promisefru/mozilla/bugbug/scripts/trainer.py", line 145, in <module>
    main()
  File "/home/promisefru/mozilla/bugbug/scripts/trainer.py", line 141, in main
    retriever.go(args)
  File "/home/promisefru/mozilla/bugbug/scripts/trainer.py", line 41, in go
    metrics = model_obj.train(limit=args.limit)
  File "/home/promisefru/mozilla/bugbug/bugbug/model.py", line 377, in train
    self.le.fit(y)
  File "/home/promisefru/anaconda3/envs/bugbug/lib/python3.10/site-packages/sklearn/preprocessing/_label.py", line 98, in fit
    y = column_or_1d(y, warn=True)
  File "/home/promisefru/anaconda3/envs/bugbug/lib/python3.10/site-packages/sklearn/utils/validation.py", line 1156, in column_or_1d
    raise ValueError(
ValueError: y should be a 1d array, got an array of shape (198635, 5) instead.

_Originally posted by in

See also: )

[mayil-ai[bot]]

Possible Solution

• Locate the train method in the bugbug/model.py file where the self.le.fit(y) call is made.
• Ensure that y is being transformed into a 1-dimensional array before being passed to self.le.fit(y).
• If y is a multi-dimensional array with shape (198635, 5), it likely means that it contains multiple labels or features per instance. You need to select the correct label column or flatten the array.
• If y should only contain one label per instance, modify the code to select the correct column from y. For example, if y is a 2D numpy array and the labels are in the first column, you would use y = y[:, 0] before fitting the label encoder.
• If y is a list of lists or a 2D array that needs to be flattened, you can use y = np.ravel(y) to flatten it into a 1D array.
• After modifying y to be a 1D array, the self.le.fit(y) call should work without raising a ValueError.

Code snippets to check

bugbug → models → bugtype.py

This snippet is from the 'bugbug/models/bugtype.py' file, which is likely related to the 'bugtype' model mentioned in the issue. The error suggests a problem with the shape of the array being passed to the label encoder, which could be related to the data processing or training logic in this file. https://github.com/Mayil-AI/bugbug-21dec23/blob/0acd00da46afbb37bd047c0bce06ce7cfad21568/bugbug/models/bugtype.py#L1-L92