separate `blech_process.py` workflows for throw_out_noise=True vs False

[abuzarmahmood]

blech_process.py workflow is confusing...processing for throw_out_noise=True vs False is being handled jointly...will likely be cleaner to have distinct (if somewhat redundant) workflows

This was created because of wanting to plot all waveforms regardless of throw_out_noise variable

**Data Processing Flow:**

When throw_out_noise = False:
    Raw Waveforms (slices_dejittered)
           |
           v
    Extract Features
           |
           v
    Train GMM on all features
           |
           v
    Get cluster labels for all waveforms
           |
           v
    Remove outliers & save results

When throw_out_noise = True:
    Raw Waveforms (slices_og)
           |
           +---> Classify (neuRecommend)
           |           |
           |           +---> Noise waveforms (discarded from GMM fitting)
           |           |
           |           +---> Spike waveforms (slices_dejittered)
           |                       |
           |                       v
           |                Extract Features (spike features only)
           |                       |
           |                       v
           |                Train GMM on spike features only
           |                       |
           v                       v
    Extract Features (all)    Get GMM model
           |                       |
           +----------+------------+
                      |
                      v
           Apply GMM to all waveforms
                      |
                      v
           Get cluster labels for all waveforms
                      |
                      v
           Remove outliers & save results