Update requirements.txt

[MerlinK75]

Just updated the requirements from pyriemann>=0.2.7 -> pyriemann=0.3

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 82.86%. Comparing base (16e9f7e) to head (08247b5).
Report is 1 commits behind head on master.

Additional details and impacted files

@@           Coverage Diff           @@
##           master      #85   +/-   ##
=======================================
  Coverage   82.86%   82.86%           
=======================================
  Files          25       25           
  Lines        2813     2813           
=======================================
  Hits         2331     2331           
  Misses        482      482           

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

[nbara]

Hey @MerlinK75 , upon closer inspection I am not finding any trace of _check_est anywhere in meegkit? Could you post the full error here please?

[ohspc89]

Hello, may I chime in? First of all, thank you for preparing this wonderful package. I am just learning how to apply asr to my EEG data read by a mne.io.Raw object.

I set up a virtual environment (Python 3.11.11) and installed meegkit first. Currently it's meegkit==0.1.7 and pyriemann==0.7.

I am just replicating an example. So here's my code:

asr = ASR(method='euclid')
_, sample_mask = asr.fit(raw._data)

raw._data is just a NumPy array of shape (32, 14995); 32 channels, about 1 min recording.

Here's the error message I get.

ImportError  Traceback (most recent call last)
Cell In[9], line 1
----> 1 _, sample_mask = asr.fit(notch._data)

File ~/Documents/EEG_venv/lib/python3.11/site-
packages/meegkit/asr.py:182, in ASR.fit(self, 
X, y, **kwargs)
    172 clean, sample_mask = clean_windows(
    173     X,
    174     sfreq=self.sfreq,
   (...)
    178     min_clean_fraction=self.min_clean_
fraction,
    179     max_dropout_fraction=self.max_drop
out_fraction)
    181 # Perform calibration
--> 182 M, T = asr_calibrate(
    183     clean,
    184     sfreq=self.sfreq,
    185     cutoff=self.cutoff,
    186     blocksize=self.blocksize,
    187     win_len=self.win_len,
    188     win_overlap=self.win_overlap,
    189     max_dropout_fraction=self.max_drop
out_fraction,
    190     min_clean_fraction=self.min_clean_
fraction,
    191     method=self.method,
    192     estimator=self.estimator)
    194 self.state_ = dict(M=M, T=T, R=None)
    195 self._fitted = True

File ~/Documents/EEG_venv/lib/python3.11/site-
packages/meegkit/asr.py:491, in asr_calibrate(
X, sfreq, cutoff, blocksize, win_len, win_over
lap, max_dropout_fraction, min_clean_fraction,
 method, estimator)
    488 # window length for calculating thresh
olds
    489 N = int(np.round(win_len * sfreq))
--> 491 U = block_covariance(X, window=blocksi
ze, overlap=win_overlap,
    492                      estimator=estimat
or)
    493 if method == "euclid":
    494     Uavg = geometric_median(U.reshape(
(-1, nc * nc)))

File ~/Documents/EEG_venv/lib/python3.11/site-
packages/meegkit/utils/covariances.py:36, in b
lock_covariance(data, window, overlap, padding
, estimator)
     18 def block_covariance(data, window=128,
 overlap=0.5, padding=True, estimator="cov"):
     19     """Compute blockwise covariance.
     20 
     21     Parameters
   (...)
     34 
     35     """
---> 36     from pyriemann.utils.covariance im
port _check_est
     38     assert 0 <= overlap < 1, "overlap 
must be < 1"
     39     est = _check_est(estimator)

ImportError: cannot import name '_check_est' f
rom 'pyriemann.utils.covariance' (/Users/joh/D
ocuments/EEG_venv/lib/python3.11/site-packages
/pyriemann/utils/covariance.py)

Let me know if I need to clarify more. Sorry for my verbosity... just another noob here.

[nbara]

@ohspc89 Thanks for flagging this. Please could you install the latest version from github and try again?

I think this should ought to do the trick:

pip install git+https://github.com/nbara/python-meegkit.git

If this works that means the problem is solved on main, and I just need to release a new version on PyPI

[ohspc89]

@nbara Will check this out tonight and share the result. Thank you!

[ohspc89]

@nbara Yes my code works without much problem. Just one question - right now your requirements.txt lists pyriemann>=0.2.7 to be installed. I initially tested my code with pyriemann==0.4 and meegkit==0.1.8. With that I get a different error:

Traceback (most recent call last):
  File "/home/jinseok/Documents/EEG_notebooks/EEG_Self_Study/testasr.py", line 40, in <module>
    out = asr_preserve_samples(origdata, freq_to_use)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jinseok/Documents/EEG_notebooks/EEG_Self_Study/testasr.py", line 28, in asr_preserve_samples
    _, sample_mask = asr.fit(data)
                     ^^^^^^^^^^^^^
  File "/home/jinseok/Documents/EEG/lib/python3.12/site-packages/meegkit/asr.py", line 182, in fit
    M, T = asr_calibrate(
           ^^^^^^^^^^^^^^
  File "/home/jinseok/Documents/EEG/lib/python3.12/site-packages/meegkit/asr.py", line 491, in asr_calibrate
    U = block_covariance(X, window=blocksize, overlap=win_overlap,
        ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/jinseok/Documents/EEG/lib/python3.12/site-packages/meegkit/utils/covariances.py", line 36, in block_covariance
    from pyriemann.utils.covariance import check_function, cov_est_functions
ImportError: cannot import name 'check_function' from 'pyriemann.utils.covariance' (/home/jinseok/Documents/EEG/lib/python3.12/site-packages/pyriemann/utils/covariance.py)

After updating pyriemann to the recent version (0.7), I did not see the error message. If so, I think it's a safer bet to set the requirement to be pyriemann>=0.7. Just a thought tho ;)

[nbara]

Hi @ohspc89 indeed, I will force pyriemann>=0.7 with the next meegkit release

[nbara]

Hi @ohspc89 @MerlinK75 the problem should be solved as part of release 0.1.9