unravel.cluster_stats.prism module#

Use cstats_prism (prism) from UNRAVEL to organize data for clusters for plotting in Prism.

Inputs:

*.csv from cstats_org_data / cstats_validation outputs in the working dir

CSV naming conventions:

  • Condition: first word before ‘_’ in the file name (use utils_prepend if needed)

  • Sample: second word in file name

Example unilateral inputs:

  • condition1_sample01_<metric>_data.csv

  • condition1_sample02_<metric>_data.csv

  • condition2_sample03_<metric>_data.csv

  • condition2_sample04_<metric>_data.csv

Example bilateral inputs (if any file has _LH.csv or _RH.csv, the command will attempt to pool data):

  • condition1_sample01_<metric>_data_LH.csv

  • condition1_sample01_<metric>_data_RH.csv

Columns in the input .csv files:

sample, cluster_ID, metric, value, value_type, support, support_type, aggregation_method, cluster_volume, …

Outputs:

  • Outputs saved in ./_prism/

  • Cluster order follows -ids order

  • <metric>_summary.csv

  • [<metric>_summary_for_valid_clusters.csv]

  • [<metric>_summary_across_clusters.csv]

  • [cluster_volume_summary.csv]

Note

  • cstats_table saves valid_clusters_dir/valid_cluster_IDs_sorted_by_anatomy.txt

  • Hemisphere suffix usage must be consistent across files (all _LH/_RH or none).

Usage:#

cstats_prism [-ids 1 2 3] [-p /path/to/csv/files/from/cstats_validation] [-v]

unravel.cluster_stats.prism.parse_args()[source]#
unravel.cluster_stats.prism.sort_samples(sample_names)[source]#
unravel.cluster_stats.prism.pool_sample_rows(dfs, metric_name, value_col, support_col, aggregation_method)[source]#

Pool LH/RH data for one sample if both sides are present; otherwise use available side only.

unravel.cluster_stats.prism.generate_summary_table(csv_files, schema, field='value')[source]#

Generate a Prism summary table for the requested field.

field can be:

  • ‘value’

  • ‘support’

  • ‘cluster_volume’

unravel.cluster_stats.prism.main()[source]#