Utilities¶

Utility modules provide supporting functionality for the spike sorting pipeline.

blech_utils¶

Core utility functions used throughout the codebase.

Key Classes and Functions¶

Tee¶

Redirect stdout to both console and file for logging.

from utils.blech_utils import Tee
import sys

# Redirect output to file and console
sys.stdout = Tee('/path/to/data/dir', name='logfile.txt')
print("This goes to both console and file")

path_handler¶

Handle file paths and directory operations.

from utils.blech_utils import path_handler

# Instantiate path handler
ph = path_handler()

# Access blech_clust directory
blech_dir = ph.blech_clust_dir

# Access home directory
home = ph.home_dir

imp_metadata¶

Import and manage experimental metadata.

from utils.blech_utils import imp_metadata
import sys

# Load metadata (requires sys.argv or list with directory path)
metadata = imp_metadata([sys.argv, '/path/to/data'])

# Access metadata attributes
hdf5_name = metadata.hdf5_name
info_dict = metadata.info_dict
layout_df = metadata.layout

Clustering Utilities¶

Located in utils/clustering/, these modules provide clustering algorithms and tools.

Key Functions¶

Spike clustering algorithms
Feature extraction
Cluster validation
Merge/split operations

Data Management¶

ephys_data Module¶

Comprehensive data handling for electrophysiology recordings.

See Ephys Data for detailed documentation.

Quality Assurance¶

qa_utils Module¶

Tools for dataset quality assessment and validation.

See QA Tools for detailed documentation.

Helper Scripts¶

infer_rnn_rates.py¶

Infer firing rates from spike trains using RNN.

python utils/infer_rnn_rates.py <data_dir> [options]

Options:

--train_steps: Number of training steps
--hidden_size: RNN hidden layer size
--bin_size: Spike binning size
--retrain: Force model retraining

blech_data_summary.py¶

Generate comprehensive dataset summary.

python utils/blech_data_summary.py

grade_dataset.py¶

Grade dataset quality based on metrics.

python utils/grade_dataset.py

Configuration Management¶

Parameter Files¶

Utilities for loading and managing parameter files:

JSON parameter loading
Parameter validation
Default value handling

Example¶

import json

# Load parameters
with open('params/sorting_params.json', 'r') as f:
    params = json.load(f)

# Access parameters
max_clusters = params['max_clusters']
min_cluster_size = params['min_cluster_size']

File I/O¶

HDF5 Operations¶

Functions for reading and writing HDF5 files:

import tables

# Open HDF5 file
with tables.open_file('data.h5', 'r') as hf5:
    # Read spike times
    spike_times = hf5.root.spike_times[:]

    # Read unit information
    units = hf5.root.units[:]

Binary Data¶

Functions for reading Intan binary data:

Amplifier data (.dat files)
Digital input data (DIN files)
Auxiliary input data