Merge pull request #373 from MICA-MNI/369-fix-histology-fslr32k-url

Add warning for fslr32k histology data resolution mismatch

Author: zihuaihuai

brainstat/_utils.py

@@ -150,7 +150,7 @@ def deprecated_func(*args, **kwargs):
 def _download_file(
     url: str, output_file: Path, overwrite: bool = False, verbose=True
 ) -> None:
-    """Downloads a file.
+    """Downloads a file with retry logic for network failures.
 
     Parameters
     ----------
@@ -163,14 +163,36 @@ def _download_file(
     verbose : bool
         If true, print a download message, defaults to True.
     """
+    import time
+    from http.client import RemoteDisconnected
+    from urllib.error import URLError
 
     if output_file.exists() and not overwrite:
         return
 
     if verbose:
         logger.info("Downloading " + str(output_file) + " from " + url + ".")
-    with urllib.request.urlopen(url) as response, open(output_file, "wb") as out_file:
-        shutil.copyfileobj(response, out_file)
+    
+    # Retry logic for intermittent network failures
+    max_retries = 3
+    retry_delay = 2  # seconds
+    
+    for attempt in range(max_retries):
+        try:
+            with urllib.request.urlopen(url, timeout=30) as response, open(output_file, "wb") as out_file:
+                shutil.copyfileobj(response, out_file)
+            return  # Success, exit function
+        except (RemoteDisconnected, URLError, TimeoutError) as e:
+            if attempt < max_retries - 1:
+                logger.warning(
+                    f"Download attempt {attempt + 1}/{max_retries} failed: {e}. "
+                    f"Retrying in {retry_delay} seconds..."
+                )
+                time.sleep(retry_delay)
+                retry_delay *= 2  # Exponential backoff
+            else:
+                logger.error(f"Download failed after {max_retries} attempts.")
+                raise  # Re-raise the exception after all retries fail
 
 
 

brainstat/context/histology.py

@@ -157,6 +157,15 @@ def read_histology_profile(
 
     with h5py.File(histology_file, "r") as h5_file:
         if template == "fslr32k":
+            # Known issue #369: The fslr32k data file on the server actually contains
+            # fs_LR_64k resolution data (64984 vertices) instead of fs_LR_32k (32492 vertices).
+            # This is a data hosting issue. Users expecting 32k resolution data should be aware
+            # that they will receive 64k resolution data until this is fixed on the server.
+            logger.warning(
+                "Known issue: The fslr32k histology profile data currently contains "
+                "fs_LR_64k resolution (64984 vertices) instead of fs_LR_32k (32492 vertices). "
+                "See https://github.com/MICA-MNI/BrainStat/issues/369 for details."
+            )
             profiles = h5_file.get("fs_LR_64k")[...]
         else:
             profiles = h5_file.get(template)[...]
@@ -196,6 +205,12 @@ def download_histology_profiles(
     data_dir.mkdir(parents=True, exist_ok=True)
     if template == "fslr32k":
         output_file = data_dir / "histology_fslr32k.h5"
+        # Known issue #369: warn users about data resolution mismatch
+        logger.warning(
+            "Note: The fslr32k histology profile data currently contains "
+            "fs_LR_64k resolution (64984 vertices) instead of fs_LR_32k. "
+            "See https://github.com/MICA-MNI/BrainStat/issues/369 for details."
+        )
     else:
         output_file = data_dir / ("histology_" + template + ".h5")
 

brainstat/tests/test_histology.py

@@ -18,9 +18,6 @@ def test_urls(template):
     template : list
         Template names.
     """
-    if template == "fslr32k":
-        pytest.skip("Skipping fslr32k due to known netneurotools 0.3.0 Unicode bug")
-    
     try:
         r = requests.head(json["bigbrain_profiles"][template]["url"], timeout=10)
         assert r.status_code == 200