Update pyPDAF/mcp_output/mcp_plugin/mcp_service.py

pyPDAF/mcp_output/mcp_plugin/mcp_service.py

@@ -584,6 +584,150 @@ def run_enoi_pipeline(
         "localization_radius_km": localization_radius if localization_radius > 0 else "none"
     }
 
+@mcp.tool(name="upload_to_hf_dataset", description="Upload generated file to HuggingFace Dataset for easy download")
+def upload_to_hf_dataset(file_path: str, dataset_repo: str = None, hf_token: str = None) -> dict:
+    """
+    Upload generated file to HuggingFace Dataset repository.
+    Users can then download files from the Dataset page.
+
+    Parameters:
+        file_path (str): Absolute path to the file on HF Space
+        dataset_repo (str): Dataset repo name (e.g., 'username/obspy-outputs').
+                           If None, uses HF_DATASET_REPO environment variable
+        hf_token (str): HuggingFace token. If None, uses HF_TOKEN environment variable
+
+    Returns:
+        dict: Success status and download URL
+    """
+    try:
+        from huggingface_hub import HfApi, create_repo
+        from huggingface_hub.utils import RepositoryNotFoundError
+
+        if dataset_repo is None:
+            dataset_repo = os.environ.get("HF_DATASET_REPO")
+            if not dataset_repo:
+                return {
+                    "success": False,
+                    "error": "Dataset repo not specified. Set HF_DATASET_REPO environment variable or pass dataset_repo parameter"
+                }
+
+        if hf_token is None:
+            hf_token = os.environ.get("HF_TOKEN")
+            if not hf_token:
+                return {
+                    "success": False,
+                    "error": "HF token not found. Set HF_TOKEN environment variable or pass hf_token parameter"
+                }
+
+        # Check if file exists and readable
+        if not os.path.exists(file_path):
+            return {
+                "success": False,
+                "error": f"File not found: {file_path}"
+            }
+
+        if not os.path.isfile(file_path):
+            return {
+                "success": False,
+                "error": f"Path is not a file: {file_path}"
+            }
+
+        if not os.access(file_path, os.R_OK):
+            return {
+                "success": False,
+                "error": f"File not readable (permission denied): {file_path}",
+                "hint": "Check file permissions in Docker container"
+            }
+
+        # Get file info for debugging
+        file_size = os.path.getsize(file_path)
+        file_stat = os.stat(file_path)
+
+        # Initialize HF API
+        api = HfApi()
+
+        # Try to create dataset repo if it doesn't exist
+        try:
+            repo_info = create_repo(
+                repo_id=dataset_repo,
+                repo_type="dataset",
+                token=hf_token,
+                exist_ok=True,
+                private=False
+            )
+        except Exception as e:
+            return {
+                "success": False,
+                "error": f"Failed to create/access dataset repo: {str(e)}",
+                "dataset_repo": dataset_repo,
+                "hint": "Check if HF_TOKEN has write permission and dataset_repo format is correct (username/repo-name)"
+            }
+
+        # Get filename and determine path in dataset
+        filename = os.path.basename(file_path)
+
+        # Determine subdirectory based on file location or file type
+        if "/output/" in file_path:
+            path_in_repo = f"output/{filename}"
+        elif "/plots/" in file_path:
+            path_in_repo = f"plots/{filename}"
+        elif "/wave_data/" in file_path:
+            path_in_repo = f"wave_data/{filename}"
+        elif filename.endswith('.nc'):
+            # NetCDF files from pyPDAF (restart, analysis, ensemble, obs)
+            if "restart" in filename or "analysis" in filename:
+                path_in_repo = f"pypdaf_analysis/{filename}"
+            elif "obs" in filename:
+                path_in_repo = f"pypdaf_observations/{filename}"
+            elif "member" in filename or "ensemble" in file_path:
+                path_in_repo = f"pypdaf_ensemble/{filename}"
+            else:
+                path_in_repo = f"pypdaf_data/{filename}"
+        else:
+            path_in_repo = filename
+
+        try:
+            with open(file_path, 'rb') as f:
+                file_content = f.read()
+        except Exception as e:
+            return {
+                "success": False,
+                "error": f"Failed to read file: {str(e)}",
+                "file_path": file_path,
+                "hint": "File may exist but not readable due to permission issues"
+            }
+
+        # Upload file from memory instead of path
+        import io
+        upload_result = api.upload_file(
+            path_or_fileobj=io.BytesIO(file_content),
+            path_in_repo=path_in_repo,
+            repo_id=dataset_repo,
+            repo_type="dataset",
+            token=hf_token
+        )
+
+        # Construct download URL
+        download_url = f"https://huggingface.co/datasets/{dataset_repo}/resolve/main/{path_in_repo}"
+        viewer_url = f"https://huggingface.co/datasets/{dataset_repo}/viewer/default/train?f%5Bfile%5D%5Bvalue%5D={path_in_repo}"
+
+        return {
+            "success": True,
+            "message": f"File uploaded successfully to {dataset_repo}",
+            "dataset_repo": dataset_repo,
+            "filename": filename,
+            "path_in_repo": path_in_repo,
+            "download_url": download_url,
+            "viewer_url": viewer_url,
+            "usage": f"Download directly from: {download_url}"
+        }
+
+    except Exception as e:
+        return {
+            "success": False,
+            "error": str(e),
+            "hint": "Make sure HF_TOKEN and HF_DATASET_REPO are set in Space secrets"
+        }
 
 # ============================================================================
 # Create MCP Server App