Add ELN collection export

pydatalab/pyproject.toml

@@ -53,6 +53,7 @@ server = [
     "Flask-PyMongo ~= 2.3",
     "Flask-Mail ~= 0.10",
     "Flask-Compress ~= 1.15",
+    "APScheduler ~= 3.10",
     "Werkzeug ~= 3.0",
     "python-dotenv ~= 1.0",
     "pillow ~= 11.0",

pydatalab/src/pydatalab/export_utils.py

@@ -0,0 +1,169 @@
+import json
+import shutil
+import tempfile
+import zipfile
+from datetime import datetime, timezone
+from pathlib import Path
+
+from bson import ObjectId
+
+from pydatalab.config import CONFIG
+from pydatalab.mongo import flask_mongo
+
+
+def _convert_objectids_in_dict(d: dict) -> dict:
+    """Recursively convert ObjectIds and datetimes in a dictionary to strings."""
+    result: dict = {}
+    for key, value in d.items():
+        if isinstance(value, ObjectId):
+            result[key] = str(value)
+        elif isinstance(value, datetime):
+            result[key] = value.isoformat()
+        elif isinstance(value, dict):
+            result[key] = _convert_objectids_in_dict(value)
+        elif isinstance(value, list):
+            result[key] = [
+                str(v)
+                if isinstance(v, ObjectId)
+                else v.isoformat()
+                if isinstance(v, datetime)
+                else _convert_objectids_in_dict(v)
+                if isinstance(v, dict)
+                else v
+                for v in value
+            ]
+        else:
+            result[key] = value
+    return result
+
+
+def generate_ro_crate_metadata(collection_data: dict, child_items: list[dict]) -> dict:
+    """Generate RO-Crate metadata for the .eln file.
+
+    Parameters:
+        collection_data: The collection metadata
+        child_items: List of items in the collection
+
+    Returns:
+        RO-Crate metadata as a dictionary
+    """
+
+    experiments: list[dict] = []
+    for item in child_items:
+        experiments.append(
+            {
+                "@id": f"./{item['item_id']}/",
+            }
+        )
+
+    graph: list[dict] = [
+        {
+            "@id": "ro-crate-metadata.json",
+            "@type": "CreativeWork",
+            "about": {"@id": "./"},
+            "conformsTo": {"@id": "https://w3id.org/ro/crate/1.1"},
+            "dateCreated": datetime.now(tz=timezone.utc).isoformat(),
+            "sdPublisher": {
+                "@id": CONFIG.IDENTIFIER_PREFIX or "https://github.com/datalab-org/datalab"
+            },
+        },
+        {
+            "@id": "./",
+            "@type": "Dataset",
+            "name": collection_data.get("title", collection_data.get("collection_id")),
+            "description": collection_data.get("description", ""),
+            "hasPart": experiments,
+        },
+    ]
+
+    metadata = {
+        "@context": "https://w3id.org/ro/crate/1.1/context",
+        "@graph": graph,
+    }
+
+    for item in child_items:
+        item_metadata = {
+            "@id": f"./{item['item_id']}/",
+            "@type": "Dataset",
+            "name": item.get("name", item["item_id"]),
+            "identifier": item["item_id"],
+            "dateCreated": item.get("date", datetime.now(tz=timezone.utc)).isoformat()
+            if isinstance(item.get("date"), datetime)
+            else item.get("date"),
+        }
+
+        if item.get("file_ObjectIds"):
+            files = []
+            for file_id in item["file_ObjectIds"]:
+                file_data = flask_mongo.db.files.find_one({"_id": ObjectId(file_id)})
+                if file_data:
+                    files.append({"@id": f"./{item['item_id']}/{file_data['name']}"})
+            if files:
+                item_metadata["hasPart"] = files
+
+        graph.append(item_metadata)
+
+    return metadata
+
+
+def create_eln_file(collection_id: str, output_path: str) -> None:
+    """Create a .eln file for a collection.
+
+    Parameters:
+        collection_id: ID of the collection to export
+        output_path: Path where the .eln file should be saved
+    """
+
+    collection_data = flask_mongo.db.collections.find_one({"collection_id": collection_id})
+    if not collection_data:
+        raise ValueError(f"Collection {collection_id} not found")
+
+    collection_immutable_id = collection_data["_id"]
+    child_items = list(
+        flask_mongo.db.items.find(
+            {
+                "relationships": {
+                    "$elemMatch": {"type": "collections", "immutable_id": collection_immutable_id}
+                }
+            }
+        )
+    )
+
+    with tempfile.TemporaryDirectory() as temp_dir:
+        temp_path = Path(temp_dir)
+
+        root_folder_name = collection_id
+        root_folder = temp_path / root_folder_name
+        root_folder.mkdir()
+
+        ro_crate_metadata = generate_ro_crate_metadata(collection_data, child_items)
+        with open(root_folder / "ro-crate-metadata.json", "w", encoding="utf-8") as f:
+            json.dump(ro_crate_metadata, f, indent=2, ensure_ascii=False)
+
+        for item in child_items:
+            item_folder = root_folder / item["item_id"]
+            item_folder.mkdir()
+
+            item_metadata = {k: v for k, v in item.items() if k not in ["_id", "file_ObjectIds"]}
+
+            item_metadata = _convert_objectids_in_dict(item_metadata)
+
+            with open(item_folder / "metadata.json", "w", encoding="utf-8") as f:
+                json.dump(item_metadata, f, indent=2, ensure_ascii=False)
+
+            if item.get("file_ObjectIds"):
+                for file_id in item.get("file_ObjectIds", []):
+                    file_data = flask_mongo.db.files.find_one({"_id": ObjectId(file_id)})
+                    if file_data:
+                        source_path = Path(file_data["location"])
+                        if source_path.exists():
+                            dest_file = item_folder / file_data["name"]
+                            shutil.copy2(source_path, dest_file)
+                        else:
+                            print(f"Warning: File not found on disk: {file_data['location']}")
+
+        with zipfile.ZipFile(output_path, "w", zipfile.ZIP_DEFLATED) as zipf:
+            for file_path in root_folder.rglob("*"):
+                if file_path.is_file():
+                    arcname = file_path.relative_to(temp_path)
+                    zipf.write(file_path, arcname)

pydatalab/src/pydatalab/models/export_task.py

@@ -0,0 +1,34 @@
+from datetime import datetime, timezone
+from enum import Enum
+
+from pydantic import BaseModel, Field
+
+
+class ExportStatus(str, Enum):
+    """Status of an export task."""
+
+    PENDING = "pending"
+    PROCESSING = "processing"
+    READY = "ready"
+    ERROR = "error"
+
+
+class ExportTask(BaseModel):
+    """Model for an export task."""
+
+    task_id: str = Field(..., description="Unique identifier for the export task")
+    collection_id: str = Field(..., description="ID of the collection being exported")
+    status: ExportStatus = Field(
+        default=ExportStatus.PENDING, description="Current status of the task"
+    )
+    creator_id: str = Field(..., description="ID of the user who created the export")
+    created_at: datetime = Field(
+        default_factory=lambda: datetime.now(tz=timezone.utc),
+        description="When the task was created",
+    )
+    completed_at: datetime | None = Field(None, description="When the task was completed")
+    file_path: str | None = Field(None, description="Path to the generated .eln file")
+    error_message: str | None = Field(None, description="Error message if status is ERROR")
+
+    class Config:
+        use_enum_values = True

pydatalab/src/pydatalab/mongo.py

@@ -208,4 +208,15 @@ def create_user_fts():
         db.users.drop_index(user_fts_name)
         ret += create_user_fts()
 
+    ret += db.export_tasks.create_index(
+        "task_id", unique=True, name="unique task ID", background=background
+    )
+    ret += db.export_tasks.create_index(
+        "creator_id", name="export task creator", background=background
+    )
+    ret += db.export_tasks.create_index(
+        "created_at", name="export task created at", background=background
+    )
+    ret += db.export_tasks.create_index("status", name="export task status", background=background)
+
     return ret

pydatalab/src/pydatalab/routes/v0_1/__init__.py

@@ -5,6 +5,7 @@
 from .auth import AUTH, OAUTH, OAUTH_PROXIES
 from .blocks import BLOCKS
 from .collections import COLLECTIONS
+from .export import EXPORT
 from .files import FILES
 from .graphs import GRAPHS
 from .healthcheck import HEALTHCHECK
@@ -25,6 +26,7 @@
     HEALTHCHECK,
     INFO,
     GRAPHS,
+    EXPORT,
 )
 
 __all__ = ("BLUEPRINTS", "OAUTH", "__api_version__", "OAUTH_PROXIES")