@@ -1,106 +1,60 @@
 #!/usr/bin/env python3
 """
-Run project-specific two-group tests for mean IF cluster data.
+Run project-specific two-group tests for mean intensities from cluster data.
 
-This is a simpler companion to the Psi/Iso 2x2 ANOVA script for cases where
-only two conditions are present, e.g.:
+Prereqs:
+    - `cstats_mean_IF` to generate mean intensity CSVs for each sample.
 
-    AwS = Awake Saline
-    AnS = Anesthetized Saline
-
-Input convention:
-    - each input CSV is one sample
+Inputs:
+    - CSVs in the working directory (one per sample) or specified with -i
     - the condition is encoded as the first underscore-delimited field in the filename
-      e.g., AwS_sample01_cluster_mean_IF.csv -> condition AwS
-    - each CSV contains one row per cluster
+      e.g., AwS_sample01_cluster_mean_intensity.csv -> condition AwS
+    - each CSV contains one row per cluster (columns: cluster_ID, mean_intensity, etc.)
 
 For each cluster, outputs:
     - group means, SD, SEM, and n for the two conditions
     - difference: second condition minus first condition
       default: AnS - AwS
-    - Welch t-test by default, or equal-variance t-test with --equal_var
-    - Holm-Sidak and Benjamini-Hochberg FDR correction across clusters
+    - Student's t-test (equal-variance)
     - higher-mean group
     - Cohen's d and Hedges' g, using condition order:
       positive = second condition > first condition
 
-Example:
-    mean_IF_2group_ttest.py \
-        -i 'cluster_mean_IF_AnS_v_AwS_vox_p_tstat1_q0.05_rev_cluster_index_LH/*.csv' \
-        -o mean_IF_AnS_vs_AwS_ttest_results.csv
-
-Equivalent explicit conditions:
-    mean_IF_2group_ttest.py \
-        -i '*.csv' \
-        --conditions AwS AnS \
-        -o mean_IF_AnS_minus_AwS_ttest_results.csv
+Usage:
+------
+    ./mean_IF_2group_ttest.py -c AwS AnS [OPTIONS]
 """
 
-import argparse
-from pathlib import Path
-import glob as globlib
-
 import numpy as np
 import pandas as pd
+from pathlib import Path
+from rich import print
+from rich.traceback import install
 from scipy.stats import ttest_ind
-from statsmodels.stats.multitest import multipletests
-
 
-DEFAULT_CONDITIONS = ["AwS", "AnS"]
+from unravel.core.help_formatter import RichArgumentParser, SuppressMetavar, SM
+from unravel.core.config import Configuration
+from unravel.core.utils import log_command, match_files, verbose_end_msg, verbose_start_msg
 
 
 def parse_args():
-    p = argparse.ArgumentParser(
-        description="Project-specific two-group t-tests for mean_IF cluster CSVs."
-    )
-    p.add_argument(
-        "-i", "--input_dir", dest="inputs", nargs="*", default=None,
-        help=(
-            "Input directory, CSV file, or quoted glob pattern. May be repeated as "
-            "multiple values after -i. Default: current directory."
-        ),
-    )
-    p.add_argument(
-        "-g", "--glob", default="*.csv",
-        help="Input CSV glob used when an input is a directory. Default: *.csv",
-    )
-    p.add_argument(
-        "--conditions", nargs=2, default=DEFAULT_CONDITIONS, metavar=("REF", "TEST"),
-        help=(
-            "Two condition prefixes to compare. Difference/effect size is TEST - REF. "
-            "Default: AwS AnS, so diff = AnS - AwS."
-        ),
-    )
-    p.add_argument(
-        "-vcol", "--value_col", default="mean_IF_intensity",
-        help="Dependent-variable column. Default: mean_IF_intensity",
-    )
-    p.add_argument(
-        "-ccol", "--cluster_col", default="cluster_ID",
-        help="Cluster ID column. Default: cluster_ID",
-    )
-    p.add_argument(
-        "-scol", "--sample_col", default="sample",
-        help="Sample column. If absent, filename stem is used. Default: sample",
-    )
-    p.add_argument(
-        "-o", "--out", default="mean_IF_2group_ttest_results.csv",
-        help="Output CSV path. Default: mean_IF_2group_ttest_results.csv",
-    )
-    p.add_argument(
-        "--raw_out", default="mean_IF_2group_ttest_raw_data.csv",
-        help="Raw long-format output CSV path. Default: mean_IF_2group_ttest_raw_data.csv",
-    )
-    p.add_argument(
-        "--equal_var", action="store_true", default=False,
-        help="Use equal-variance t-tests. Default: Welch t-tests.",
-    )
-    p.add_argument(
-        "--min_n_per_group", type=int, default=2,
-        help="Minimum n per condition needed for the t-test. Default: 2",
-    )
-    return p.parse_args()
+    parser = RichArgumentParser(formatter_class=SuppressMetavar, add_help=False, docstring=__doc__)
+
+    reqs = parser.add_argument_group('Required arguments')
+    reqs.add_argument('-c', '--conditions', help='Conditions. Each condition must match first word of CSV filenames (underscore-separated).', required=True, nargs='*', action=SM)
+
+    opts = parser.add_argument_group('Optional arguments')
+    opts.add_argument('-i', '--input', help="Path(s) or glob pattern(s) for input CSV files. Default: '*.csv'", default='*.csv', nargs='*', action=SM)
+    opts.add_argument("-o", "--out", default="out/mean_intensity_2group_ttest_results.csv", help="Output CSV path. Default: out/mean_intensity_2group_ttest_results.csv")
+    opts.add_argument("-r", "--raw_out", default="out/mean_intensity_2group_ttest_raw_data.csv", help="Raw long-format output CSV path. Default: out/mean_intensity_2group_ttest_raw_data.csv")
+    opts.add_argument("-vc", "--value_col", default="mean_intensity", help="Dependent-variable column. Default: mean_intensity")
+    opts.add_argument("-cc", "--cluster_col", default="cluster_ID", help="Cluster ID column. Default: cluster_ID")
+    opts.add_argument("-sc", "--sample_col", default="sample", help="Sample column. If absent, filename stem is used. Default: sample")
 
+    general = parser.add_argument_group('General arguments')
+    general.add_argument('-v', '--verbose', help='Verbose output.', action='store_true')
+
+    return parser.parse_args()
 
 def p_to_sig(p):
     if pd.isna(p):
@@ -128,58 +82,7 @@ def safe_float(x):
 def infer_condition(filename):
     return Path(filename).name.split("_")[0]
 
-
-def expand_input_csvs(inputs, glob_pattern):
-    """Expand directories, individual CSV files, and quoted glob patterns."""
-    if not inputs:
-        inputs = ["."]
-
-    csvs = []
-    missing_inputs = []
-
-    for item in inputs:
-        item = str(item)
-
-        # Allows commands like:
-        #   -i 'cluster_mean_IF_dir/*.csv'
-        if globlib.has_magic(item):
-            matches = [Path(m) for m in globlib.glob(item, recursive=True)]
-            csvs.extend([
-                m for m in matches
-                if m.is_file() and m.name.lower().endswith(".csv")
-            ])
-            continue
-
-        path = Path(item).expanduser()
-
-        if path.is_dir():
-            csvs.extend([
-                m for m in path.glob(glob_pattern)
-                if m.is_file() and m.name.lower().endswith(".csv")
-            ])
-        elif path.is_file():
-            if path.name.lower().endswith(".csv"):
-                csvs.append(path)
-        else:
-            missing_inputs.append(item)
-
-    # De-duplicate while keeping deterministic order.
-    csvs = sorted({str(p): p for p in csvs}.values(), key=lambda p: str(p))
-
-    if not csvs:
-        msg = (
-            "No input CSVs found. "
-            f"inputs={inputs!r}; directory glob={glob_pattern!r}"
-        )
-        if missing_inputs:
-            msg += f"; missing inputs={missing_inputs!r}"
-        raise FileNotFoundError(msg)
-
-    return csvs
-
-
-def load_mean_if_data(inputs, glob_pattern, conditions, value_col, cluster_col, sample_col):
-    csvs = expand_input_csvs(inputs, glob_pattern)
+def load_mean_intensity_data(csvs, conditions, value_col, cluster_col, sample_col):
     conditions = [str(c) for c in conditions]
     condition_set = set(conditions)
 
@@ -259,7 +162,7 @@ def higher_by_mean(mean_ref, mean_test, ref_cond, test_cond):
     return test_cond if mean_test > mean_ref else ref_cond
 
 
-def cohen_d_and_hedges_g(vals_ref, vals_test):
+def hedges_g(vals_ref, vals_test):
     """
     Standardized mean difference using pooled SD.
 
@@ -289,18 +192,17 @@ def cohen_d_and_hedges_g(vals_ref, vals_test):
     # Small-sample correction.
     correction = 1 - (3 / (4 * (n_ref + n_test) - 9)) if (n_ref + n_test) > 2 else np.nan
     g = d * correction if not pd.isna(correction) else np.nan
-    return safe_float(d), safe_float(g)
+    return safe_float(g)
 
 
-def run_ttest(vals_ref, vals_test, equal_var, min_n_per_group):
+def run_ttest(vals_ref, vals_test, min_n_per_group=2):
     if len(vals_ref) < min_n_per_group or len(vals_test) < min_n_per_group:
         return np.nan, np.nan
 
     try:
         res = ttest_ind(
             vals_test,
             vals_ref,
-            equal_var=equal_var,
             alternative="two-sided",
             nan_policy="omit",
         )
@@ -309,38 +211,7 @@ def run_ttest(vals_ref, vals_test, equal_var, min_n_per_group):
         return np.nan, np.nan
 
 
-def add_multiple_testing_columns(results):
-    """Add Holm-Sidak and BH-FDR adjusted p-values across clusters."""
-    results = results.copy()
-
-    results["p_holm_sidak"] = np.nan
-    results["sig_holm_sidak"] = ""
-    results["reject_holm_sidak_0.05"] = False
-
-    results["p_fdr_bh"] = np.nan
-    results["sig_fdr_bh"] = ""
-    results["reject_fdr_bh_0.05"] = False
-
-    pvals = results["p"].to_numpy(dtype=float)
-    valid = ~pd.isna(pvals)
-
-    if valid.sum() > 0:
-        reject_hs, p_hs, _, _ = multipletests(pvals[valid], alpha=0.05, method="holm-sidak")
-        reject_bh, p_bh, _, _ = multipletests(pvals[valid], alpha=0.05, method="fdr_bh")
-
-        valid_idx = results.index[valid]
-        results.loc[valid_idx, "p_holm_sidak"] = p_hs
-        results.loc[valid_idx, "reject_holm_sidak_0.05"] = reject_hs
-        results.loc[valid_idx, "sig_holm_sidak"] = [p_to_sig(p) for p in p_hs]
-
-        results.loc[valid_idx, "p_fdr_bh"] = p_bh
-        results.loc[valid_idx, "reject_fdr_bh_0.05"] = reject_bh
-        results.loc[valid_idx, "sig_fdr_bh"] = [p_to_sig(p) for p in p_bh]
-
-    return results
-
-
-def build_results(data, conditions, value_col, cluster_col, equal_var, min_n_per_group):
+def build_results(data, conditions, value_col, cluster_col, min_n_per_group=2):
     ref_cond, test_cond = [str(c) for c in conditions]
     diff_col = f"{test_cond}_minus_{ref_cond}"
 
@@ -364,11 +235,10 @@ def build_results(data, conditions, value_col, cluster_col, equal_var, min_n_per
         t_stat, p = run_ttest(
             vals_ref=vals_ref,
             vals_test=vals_test,
-            equal_var=equal_var,
             min_n_per_group=min_n_per_group,
         )
 
-        cohen_d, hedges_g = cohen_d_and_hedges_g(vals_ref, vals_test)
+        hedges_g_val = hedges_g(vals_ref, vals_test)
 
         rows.append({
             "cluster_ID": cluster_id,
@@ -385,9 +255,8 @@ def build_results(data, conditions, value_col, cluster_col, equal_var, min_n_per
             "t_stat": t_stat,
             "p": p,
             "sig": p_to_sig(p),
-            "cohen_d": cohen_d,
-            "hedges_g": hedges_g,
-            "test": "Student_ttest" if equal_var else "Welch_ttest",
+            "hedges_g": hedges_g_val,
+            "test": "Student_ttest",
             "effect_direction": (
                 f"{test_cond}>{ref_cond}" if not pd.isna(diff) and diff > 0
                 else f"{test_cond}<{ref_cond}" if not pd.isna(diff) and diff < 0
@@ -397,18 +266,28 @@ def build_results(data, conditions, value_col, cluster_col, equal_var, min_n_per
         })
 
     results = pd.DataFrame(rows)
-    results = add_multiple_testing_columns(results)
 
     sort_cols = ["p", "cluster_ID"]
     return results.sort_values(sort_cols, na_position="last")
 
 
+@log_command
 def main():
+    install()
     args = parse_args()
+    Configuration.verbose = args.verbose
+    verbose_start_msg()
+
+    csvs = match_files(args.input)
 
-    data = load_mean_if_data(
-        inputs=args.inputs,
-        glob_pattern=args.glob,
+    # Filter CSVs to only those that match the requested conditions based on filename prefix
+    csvs = [
+        f for f in csvs
+        if any(f.name.startswith(f'{group}_') for group in args.conditions)
+    ]
+
+    data = load_mean_intensity_data(
+        csvs=csvs,
         conditions=args.conditions,
         value_col=args.value_col,
         cluster_col=args.cluster_col,
@@ -424,8 +303,6 @@ def main():
         conditions=args.conditions,
         value_col=args.value_col,
         cluster_col=args.cluster_col,
-        equal_var=args.equal_var,
-        min_n_per_group=args.min_n_per_group,
     )
 
     out = Path(args.out)
@@ -438,6 +315,8 @@ def main():
     print(f"Comparison: {test_cond} - {ref_cond}")
     print(results.head(20).to_string(index=False))
 
+    verbose_end_msg()
+
 
 if __name__ == "__main__":
     main()