refine

msr.py

@@ -132,8 +132,21 @@ def download_file(url):
             with open(filepath, "wb") as f:
                 f.write(response.content)
             return True
-        
+
+        except requests.exceptions.HTTPError as e:
+            # 404 means the file doesn't exist in GHArchive - skip without retry
+            if e.response.status_code == 404:
+                if attempt == 0:  # Only log once, not for each retry
+                    print(f"   ○ {filename}: Not available (404) - skipping")
+                return False
+
+            # Other HTTP errors (5xx, etc.) should be retried
+            wait_time = DOWNLOAD_RETRY_DELAY * (2 ** attempt)
+            print(f"   ⚠ {filename}: {e}, retrying in {wait_time}s (attempt {attempt + 1}/{MAX_RETRIES})")
+            time.sleep(wait_time)
+
         except Exception as e:
+            # Network errors, timeouts, etc. should be retried
             wait_time = DOWNLOAD_RETRY_DELAY * (2 ** attempt)
             print(f"   ⚠ {filename}: {e}, retrying in {wait_time}s (attempt {attempt + 1}/{MAX_RETRIES})")
             time.sleep(wait_time)