add time series tutorial

[dkrako]

Description

• there was a tutorial for 1d shapes requested
• related issue New tutorial: for how to use Quantus with 1D shapes #83

Implemented changes

• added tutorial notebook

Minimum acceptance criteria

• I don't know

@annahedstroem unfortunately I currently cannot install the quantus package anymore on my system (python 3.10).
I'm somehow getting dependency issues which cannot be resolved, and even when I ignore this, the numpy versions just don't match? I'm very confused why this happens.

[codecov-commenter]

Codecov Report

Merging #177 (16bb2a0) into main (351f264) will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##             main     #177   +/-   ##
=======================================
  Coverage   92.86%   92.86%           
=======================================
  Files          54       54           
  Lines        2649     2649           
=======================================
  Hits         2460     2460           
  Misses        189      189           

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[annahedstroem]

@dkrako made some smaller updates to the file. but it throws an error:

--------------------------------------------------------------------------

ValueError                                Traceback (most recent call last)

/var/folders/w2/h9z278fs179f393h2jbj3y_80000gn/T/ipykernel_50650/2093353479.py in <module>
      1 # Return ModelParameterRandomisation scores for Integrated Gradients.
----> 2 scores_intgrad = metric_init(
      3     model=net,
      4     x_batch=test_features,
      5     y_batch=test_labels,

~/anaconda3/envs/quantus/lib/python3.9/site-packages/quantus/metrics/randomisation/model_parameter_randomisation.py in __call__(self, model, x_batch, y_batch, a_batch, s_batch, custom_batch, channel_first, explain_func, explain_func_kwargs, model_predict_kwargs, softmax, device, **kwargs)
    146             custom_batch,
    147             custom_preprocess_batch,
--> 148         ) = self.general_preprocess(
    149             model=model,
    150             x_batch=x_batch,

~/anaconda3/envs/quantus/lib/python3.9/site-packages/quantus/metrics/base.py in general_preprocess(self, model, x_batch, y_batch, a_batch, s_batch, custom_batch, channel_first, explain_func, explain_func_kwargs, model_predict_kwargs, softmax, device)
    298         # Reshape input batch to channel first order:
    299         if not isinstance(channel_first, bool):  # None is not a boolean instance.
--> 300             channel_first = utils.infer_channel_first(x_batch)
    301         x_batch = utils.make_channel_first(x_batch, channel_first)
    302 

~/anaconda3/envs/quantus/lib/python3.9/site-packages/quantus/helpers/utils.py in infer_channel_first(x)
    178 
    179     else:
--> 180         raise ValueError(
    181             "Only batched 1d and 2d multi-channel input dimensions supported."
    182         )

ValueError: Only batched 1d and 2d multi-channel input dimensions supported.

when trying to evaluate. the size of the test_features are (1000, 40)

[annahedstroem]

Fixed the error by removing the reduce_axes argument and reshaping the features as follows:

test_features = test_features.reshape(1000, 1, 40)