Retarget docs to OpenNeuro ds007262 v1.0.6

Author: LMBooth

README.md

@@ -1,6 +1,6 @@
 # Physiological Workload ML Pipeline
 
-Standalone repository for reproducing the multimodal cognitive-workload pipeline used on the BIDS arithmetic dataset. The codebase covers dataset acquisition, trial-table construction, QC, preprocessing, epoching, unimodal feature extraction, fused-table assembly, split-aware machine learning, confusion analysis, and publication-oriented reporting.
+Analysis repository for the OpenNeuro study [`ds007262` version `1.0.6`](https://openneuro.org/datasets/ds007262/versions/1.0.6), titled `Cognitive Workload 8-level arithmetic` (DOI `10.18112/openneuro.ds007262.v1.0.6`). The codebase covers dataset acquisition, trial-table construction, QC, preprocessing, epoching, unimodal feature extraction, fused-table assembly, split-aware machine learning, confusion analysis, and publication-oriented reporting for this specific study snapshot.
 
 ## Repository layout
 
@@ -11,7 +11,7 @@ Standalone repository for reproducing the multimodal cognitive-workload pipeline
 - `data/`: local BIDS dataset root (ignored).
 - `analysis_pipeline/runs/`: run-specific outputs (ignored).
 
-This repository is intentionally separate from any BIDS descriptor or data-publication repository. Track code, configs, and documentation here; keep raw data and generated outputs local.
+This repository is intentionally separate from the OpenNeuro dataset record itself. Track code, configs, and documentation here; keep raw data and generated outputs local.
 
 ## Quick start
 
@@ -28,11 +28,12 @@ python -m pip install -r requirements.txt
 
 ### 2. Acquire the BIDS dataset
 
-OpenNeuro dataset ID:
+OpenNeuro dataset snapshot for this repository:
 
 ```powershell
 python .\scripts\download_bids.py `
-  --dataset-id dsXXXXXX `
+  --dataset-id ds007262 `
+  --snapshot 1.0.6 `
   --target .\data\bids_arithmetic
 ```
 
@@ -44,10 +45,10 @@ python .\scripts\download_bids.py `
   --target .\data\bids_arithmetic
 ```
 
-One-command download plus pipeline execution:
+One-command download plus pipeline execution for `ds007262` `v1.0.6`:
 
 ```powershell
-.\scripts\run_end_to_end.ps1 -DatasetId dsXXXXXX -ForceDownload
+.\scripts\run_end_to_end.ps1 -DatasetId ds007262 -Snapshot 1.0.6 -ForceDownload
 ```
 
 ### 3. Run a checked-in profile
@@ -66,7 +67,7 @@ python .\analysis_pipeline\run_pipeline.py `
   --config .\analysis_pipeline\config\pipeline_unified_classic_nn_baseline_overlap3s_50pct_preproc.yaml
 ```
 
-The checked-in profiles write under `analysis_pipeline/runs/<profile_name>/`. Both set `outputs.clean_start: true`, so rerunning the same profile replaces that profile's run directory only. If you want to preserve an existing run, copy the YAML and change `outputs.root` or set `clean_start: false`.
+The checked-in profiles assume a local copy of OpenNeuro `ds007262` `v1.0.6` under `data/bids_arithmetic`. They write under `analysis_pipeline/runs/<profile_name>/`. Both set `outputs.clean_start: true`, so rerunning the same profile replaces that profile's run directory only. If you want to preserve an existing run, copy the YAML and change `outputs.root` or set `clean_start: false`.
 
 ### 4. Run stage subsets
 

analysis_pipeline/README.md

@@ -1,6 +1,6 @@
 # Analysis Pipeline Package
 
-This directory contains the executable stages, config profiles, and supporting modules that drive the end-to-end workload analysis workflow.
+This directory contains the executable stages, config profiles, and supporting modules that drive the end-to-end workload analysis workflow for the OpenNeuro `ds007262` `v1.0.6` arithmetic study snapshot.
 
 ## Canonical entry points
 

docs/README.md

@@ -3,7 +3,7 @@
 ## GitHub-facing documentation
 
 - `../README.md`
-  Repository overview, quick start, checked-in profiles, and high-level pipeline summary.
+  Repository overview, quick start, checked-in profiles, and high-level pipeline summary for OpenNeuro `ds007262` `v1.0.6`.
 - `pipeline_reference.md`
   Stage-by-stage reference for the executable pipeline, config structure, and common run patterns.
 - `reproducibility.md`

docs/pipeline_reference.md

@@ -1,6 +1,6 @@
 # Pipeline Reference
 
-This document explains how the checked-in pipeline profiles map to the executable code, runtime behavior, and output structure.
+This document explains how the checked-in pipeline profiles map to the executable code, runtime behavior, and output structure for the OpenNeuro `ds007262` `v1.0.6` dataset snapshot.
 
 ## Canonical entry points
 
@@ -32,7 +32,7 @@ Important runtime behavior:
 
 Common settings across both profiles:
 
-- BIDS root: `./data/bids_arithmetic`
+- BIDS root: `./data/bids_arithmetic` containing OpenNeuro `ds007262` `v1.0.6`
 - Stage 1 strict QC enabled
 - Stage 2 EEG preprocessing uses a Butterworth pipeline with `eeg_l_freq=2.0` and `eeg_h_freq=40.0`
 - Stage 5 builds `features_fused_tutorial_baseline.tsv` and `split_manifest_tutorial_baseline.json`
@@ -246,6 +246,15 @@ python .\analysis_pipeline\run_pipeline.py `
   --config .\analysis_pipeline\config\pipeline_unified_classic_nn_baseline_preproc.yaml
 ```
 
+Download the study snapshot expected by the checked-in profiles:
+
+```powershell
+python .\scripts\download_bids.py `
+  --dataset-id ds007262 `
+  --snapshot 1.0.6 `
+  --target .\data\bids_arithmetic
+```
+
 Dry run:
 
 ```powershell

docs/reproducibility.md

@@ -1,6 +1,6 @@
 # Reproducibility and Artifact Policy
 
-This repository is intended to keep the pipeline implementation, configs, and human-written documentation under version control while keeping raw data and generated outputs local.
+This repository is intended to keep the pipeline implementation, configs, and human-written documentation under version control while keeping raw data and generated outputs local for the OpenNeuro `ds007262` `v1.0.6` study snapshot.
 
 ## What belongs in git
 
@@ -34,11 +34,20 @@ Both configs set `outputs.clean_start: true`. Rerunning either profile replaces
 ## Recommended local workflow
 
 1. Create the Python environment and install `requirements.txt`.
-2. Download or place the BIDS dataset under `data/bids_arithmetic`.
+2. Download or place OpenNeuro `ds007262` `v1.0.6` under `data/bids_arithmetic`.
 3. Run one of the checked-in profiles through `analysis_pipeline/run_pipeline.py`.
 4. If Stage 0 to 5 outputs are already available, rerun only `stage6` or `stage6_confusions` as needed.
 5. Build manuscript-facing assets only after the underlying `ml_results*.json` inputs are present locally.
 
+Recommended download command:
+
+```powershell
+python .\scripts\download_bids.py `
+  --dataset-id ds007262 `
+  --snapshot 1.0.6 `
+  --target .\data\bids_arithmetic
+```
+
 ## Linux to Windows handoff for baseline outputs
 
 If Linux already ran the classic and deep baseline training jobs, treat the Linux outputs as the source of truth. On Windows, prefer copying the Stage 6 result artifacts and rebuilding the downstream confusion and reporting assets from those files instead of rerunning Stage 6.

scripts/download_bids.py

@@ -121,7 +121,7 @@ def _download_from_archive(url: str, target: Path) -> None:
 def main() -> None:
     parser = argparse.ArgumentParser(description="Download a BIDS dataset to a local folder.")
     source = parser.add_mutually_exclusive_group(required=True)
-    source.add_argument("--dataset-id", help="OpenNeuro dataset ID (example: ds003XXX).")
+    source.add_argument("--dataset-id", help="OpenNeuro dataset ID (example: ds007262).")
     source.add_argument("--archive-url", help="Direct URL to a BIDS archive (.zip/.tar/.tar.gz/.tgz).")
     parser.add_argument("--snapshot", default=None, help="Optional OpenNeuro snapshot tag/version.")
     parser.add_argument(
@@ -132,7 +132,7 @@ def main() -> None:
     parser.add_argument(
         "--target",
         default="data/bids_arithmetic",
-        help="Local destination folder for the BIDS dataset (default: data/bids_arithmetic).",
+        help="Local destination folder for the BIDS dataset (default: data/bids_arithmetic for ds007262 v1.0.6).",
     )
     parser.add_argument("--force", action="store_true", help="Replace target folder if it already exists.")
     args = parser.parse_args()

scripts/run_end_to_end.ps1

@@ -1,6 +1,7 @@
 param(
     [string]$DatasetId = "",
     [string]$ArchiveUrl = "",
+    [string]$Snapshot = "1.0.6",
     [string]$Target = "data/bids_arithmetic",
     [string]$Config = "analysis_pipeline/config/pipeline_unified_classic_nn_baseline_preproc.yaml",
     [switch]$ForceDownload
@@ -18,6 +19,9 @@ if ($ForceDownload) {
 }
 if (-not [string]::IsNullOrWhiteSpace($DatasetId)) {
     $downloadArgs += @("--dataset-id", $DatasetId)
+    if (-not [string]::IsNullOrWhiteSpace($Snapshot)) {
+        $downloadArgs += @("--snapshot", $Snapshot)
+    }
 } else {
     $downloadArgs += @("--archive-url", $ArchiveUrl)
 }