Add MIMIC-CXR download script and update README with detailed instructions

README.md

@@ -40,15 +40,36 @@ VLM-SubtleBench is a benchmark designed to evaluate VLMs on **subtle comparative
 
 ## Medical Data (MIMIC-CXR)
 
-The medical domain QA entries (362 attribute comparison pairs from MIMIC-CXR chest X-rays) are included in `qa.json`, but the corresponding images are not included due to [PhysioNet licensing requirements](https://physionet.org/content/mimic-cxr-jpg/2.1.0/).
+The medical domain QA entries (362 attribute comparison pairs from MIMIC-CXR chest X-rays, 664 unique images) are included in `qa.json`, but the corresponding images are not included due to [PhysioNet licensing requirements](https://physionet.org/content/mimic-cxr-jpg/2.1.0/).
 
-To obtain the medical images:
+### Step 1: Obtain PhysioNet Credentialed Access
 
-1. Obtain credentialed access to [MIMIC-CXR-JPG v2.1.0](https://physionet.org/content/mimic-cxr-jpg/2.1.0/) on PhysioNet
-2. Download the required chest X-ray images
-3. Place them under `images/mimic/` following the path structure referenced in `qa.json` (e.g., `images/mimic/p15/p15592981/s55194630/{hash}.jpg`)
+1. Create an account at [PhysioNet](https://physionet.org/)
+2. Complete the required [CITI training course](https://physionet.org/about/citi-course/) for "Data or Specimens Only Research"
+3. Go to [MIMIC-CXR-JPG v2.1.0](https://physionet.org/content/mimic-cxr-jpg/2.1.0/) and sign the data use agreement
+4. Wait for your access to be approved
 
-The image paths preserve the original MIMIC-CXR directory hierarchy, so files can be copied directly from a standard MIMIC-CXR-JPG download.
+### Step 2: Download Images
+
+We provide a script that automatically downloads only the 664 images required by `qa.json` and places them at the expected paths (`images/mimic/...`).
+
+```bash
+python download_mimic.py --user <physionet-username> --password <physionet-password>
+```
+
+The script:
+- Parses `qa.json` to find all required MIMIC-CXR image paths
+- Downloads each image from PhysioNet via `wget`
+- Places them under `images/mimic/` preserving the original directory hierarchy (e.g., `images/mimic/p15/p15592981/s55194630/{hash}.jpg`)
+- Skips images that already exist, so it is safe to re-run
+
+You can also download individual images manually:
+
+```bash
+wget --user <username> --password <password> \
+  https://physionet.org/files/mimic-cxr-jpg/2.1.0/files/p15/p15000170/s54385701/3ea0cd5d-b6ef4a9d-bd053deb-a611067c-284e4144.jpg \
+  -O images/mimic/p15/p15000170/s54385701/3ea0cd5d-b6ef4a9d-bd053deb-a611067c-284e4144.jpg
+```
 
 ## Download and Evaluation
 

download_mimic.py

@@ -0,0 +1,109 @@
+"""
+Download MIMIC-CXR images required by qa.json from PhysioNet.
+
+Prerequisites:
+  1. Create a PhysioNet account at https://physionet.org/
+  2. Complete the required CITI training course
+  3. Sign the data use agreement for MIMIC-CXR-JPG v2.1.0
+     https://physionet.org/content/mimic-cxr-jpg/2.1.0/
+
+Usage:
+  python download_mimic.py --user <username> --password <password>
+"""
+
+import argparse
+import json
+import os
+import subprocess
+import sys
+
+
+BASE_URL = "https://physionet.org/files/mimic-cxr-jpg/2.1.0/files"
+
+
+def get_required_images(qa_path: str) -> list[str]:
+    """Extract unique MIMIC image paths from qa.json."""
+    with open(qa_path) as f:
+        data = json.load(f)
+
+    images = set()
+    for entry in data:
+        for key in ("image_1", "image_2"):
+            path = entry.get(key, "")
+            if path.startswith("images/mimic/"):
+                images.add(path)
+    return sorted(images)
+
+
+def download_images(images: list[str], user: str, password: str, root_dir: str):
+    """Download images from PhysioNet and place them at the expected paths."""
+    total = len(images)
+    print(f"Found {total} MIMIC-CXR images to download.\n")
+
+    skipped = 0
+    downloaded = 0
+    failed = 0
+
+    for i, image_path in enumerate(images, 1):
+        dest = os.path.join(root_dir, image_path)
+
+        if os.path.exists(dest):
+            skipped += 1
+            continue
+
+        # images/mimic/p15/p15592981/s55194630/xxx.jpg -> p15/p15592981/s55194630/xxx.jpg
+        physionet_path = image_path.removeprefix("images/mimic/")
+        url = f"{BASE_URL}/{physionet_path}"
+
+        os.makedirs(os.path.dirname(dest), exist_ok=True)
+
+        print(f"[{i}/{total}] Downloading {physionet_path} ...")
+        result = subprocess.run(
+            [
+                "wget", "-q", "-O", dest,
+                "--user", user,
+                "--password", password,
+                url,
+            ],
+            capture_output=True,
+            text=True,
+        )
+
+        if result.returncode != 0:
+            failed += 1
+            print(f"  FAILED: {result.stderr.strip()}")
+            # Remove empty file on failure
+            if os.path.exists(dest) and os.path.getsize(dest) == 0:
+                os.remove(dest)
+        else:
+            downloaded += 1
+
+    print(f"\nDone: {downloaded} downloaded, {skipped} skipped (already exist), {failed} failed.")
+    if failed > 0:
+        print("Re-run the script to retry failed downloads.")
+
+
+def main():
+    parser = argparse.ArgumentParser(description="Download MIMIC-CXR images for VLM-SubtleBench.")
+    parser.add_argument("--user", required=True, help="PhysioNet username")
+    parser.add_argument("--password", required=True, help="PhysioNet password")
+    parser.add_argument("--qa-path", default=None, help="Path to qa.json (default: qa.json in the same directory as this script)")
+    args = parser.parse_args()
+
+    root_dir = os.path.dirname(os.path.abspath(__file__))
+    qa_path = args.qa_path or os.path.join(root_dir, "qa.json")
+
+    if not os.path.exists(qa_path):
+        print(f"Error: qa.json not found at {qa_path}", file=sys.stderr)
+        sys.exit(1)
+
+    images = get_required_images(qa_path)
+    if not images:
+        print("No MIMIC-CXR images found in qa.json.")
+        sys.exit(0)
+
+    download_images(images, args.user, args.password, root_dir)
+
+
+if __name__ == "__main__":
+    main()