Web file browsing and drag-and-drop file upload

gateway/config/settings/base.py

@@ -502,3 +502,16 @@ def __get_random_token(length: int) -> str:
     "SDS_NEW_USERS_APPROVED_ON_CREATION",
     default=False,
 )
+
+# File upload limits
+# ------------------------------------------------------------------------------
+# Maximum number of files that can be uploaded at once
+DATA_UPLOAD_MAX_NUMBER_FILES: int = env.int(
+    "DATA_UPLOAD_MAX_NUMBER_FILES", default=1000
+)
+
+# Maximum memory size for file uploads (default: 2.5MB, increased to 100MB)
+DATA_UPLOAD_MAX_MEMORY_SIZE: int = env.int(
+    "DATA_UPLOAD_MAX_MEMORY_SIZE",
+    default=104857600,  # 100MB
+)

gateway/sds_gateway/api_methods/helpers/file_helpers.py

@@ -0,0 +1,146 @@
+from http import HTTPStatus
+
+from rest_framework import status
+from rest_framework.parsers import MultiPartParser
+from rest_framework.request import Request
+from rest_framework.test import APIRequestFactory
+
+from sds_gateway.api_methods.views.capture_endpoints import CaptureViewSet
+from sds_gateway.api_methods.views.file_endpoints import CheckFileContentsExistView
+from sds_gateway.api_methods.views.file_endpoints import FileViewSet
+
+
+def upload_file_helper_simple(request, file_data):
+    """Upload a single file using FileViewSet.create.
+
+    file_data should contain all required fields: name, directory, file,
+    media_type, etc. Returns ([response], []) for success, ([], [error]) for
+    error, and handles 409 as a warning.
+    """
+    factory = APIRequestFactory()
+    django_request = factory.post(
+        request.path,
+        file_data,
+        format="multipart",
+    )
+    django_request.user = request.user
+    drf_request = Request(django_request, parsers=[MultiPartParser()])
+    drf_request.user = request.user
+    view = FileViewSet()
+    view.request = drf_request
+    view.action = "create"
+    view.format_kwarg = None
+    view.args = ()
+    view.kwargs = {}
+    try:
+        response = view.create(drf_request)
+    except (ValueError, TypeError, AttributeError, KeyError) as e:
+        return [], [f"Data validation error: {e}"]
+    else:
+        responses = []
+        errors = []
+
+        if not hasattr(response, "status_code"):
+            errors.append(getattr(response, "data", str(response)))
+        else:
+            http_status = HTTPStatus(response.status_code)
+            response_data = getattr(response, "data", str(response))
+
+            if http_status.is_success:
+                responses.append(response)
+            elif response.status_code == status.HTTP_409_CONFLICT:
+                # Already exists, treat as warning
+                errors.append(response_data)
+            elif http_status.is_server_error:
+                # Handle 500 and other server errors
+                errors.append("Internal server error")
+            elif http_status.is_client_error:
+                # Handle 4xx client errors
+                errors.append(f"Client error ({response.status_code}): {response_data}")
+            else:
+                # Handle any other status codes
+                errors.append(response_data)
+
+        return responses, errors
+
+
+# TODO: Use this helper method when implementing the file upload mode multiplexer.
+def check_file_contents_exist_helper(request, check_data):
+    """Call the post method of CheckFileContentsExistView with the given data.
+
+    check_data should contain the required fields: directory, name, sum_blake3,
+    etc.
+    """
+    factory = APIRequestFactory()
+    django_request = factory.post(
+        request.path,  # or a specific path for the check endpoint
+        check_data,
+        format="multipart",
+    )
+    django_request.user = request.user
+    drf_request = Request(django_request, parsers=[MultiPartParser()])
+    drf_request.user = request.user
+    view = CheckFileContentsExistView()
+    view.request = drf_request
+    view.action = None
+    view.format_kwarg = None
+    view.args = ()
+    view.kwargs = {}
+    return view.post(drf_request)
+
+
+def create_capture_helper_simple(request, capture_data):
+    """Create a capture using CaptureViewSet.create.
+
+    capture_data should contain all required fields for capture creation:
+    owner, top_level_dir, capture_type, channel, index_name, etc.
+    Returns ([response], []) for success, ([], [error]) for error, and handles
+    409 as a warning.
+    """
+    factory = APIRequestFactory()
+    django_request = factory.post(
+        request.path,
+        capture_data,
+        format="multipart",
+    )
+    django_request.user = request.user
+    drf_request = Request(django_request, parsers=[MultiPartParser()])
+    drf_request.user = request.user
+    view = CaptureViewSet()
+    view.request = drf_request
+    view.action = "create"
+    view.format_kwarg = None
+    view.args = ()
+    view.kwargs = {}
+    # Set the context for the serializer
+    view.get_serializer_context = lambda: {"request_user": request.user}
+    try:
+        response = view.create(drf_request)
+    except (ValueError, TypeError, AttributeError, KeyError) as e:
+        return [], [f"Data validation error: {e}"]
+    else:
+        responses = []
+        errors = []
+
+        if not hasattr(response, "status_code"):
+            errors.append(getattr(response, "data", str(response)))
+        else:
+            http_status = HTTPStatus(response.status_code)
+            response_data = getattr(response, "data", str(response))
+
+            if http_status.is_success:
+                responses.append(response)
+            elif response.status_code == status.HTTP_409_CONFLICT:
+                # Already exists, treat as warning
+                errors.append(response_data)
+            elif http_status.is_server_error:
+                # Handle 500 and other server errors
+                errors.append(f"Server error ({response.status_code}): {response_data}")
+            elif http_status.is_client_error:
+                # Handle 4xx client errors
+                errors.append(f"Client error ({response.status_code}): {response_data}")
+            else:
+                # Handle any other status codes
+                errors.append(response_data)
+
+        return responses, errors

gateway/sds_gateway/api_methods/models.py

@@ -20,6 +20,7 @@
 from django.dispatch import receiver
 from django_cog.models import Pipeline
 
+from .utils.metadata_schemas import infer_index_name
 from .utils.opensearch_client import get_opensearch_client
 
 log = logging.getLogger(__name__)
@@ -285,6 +286,13 @@ def save(self, *args, **kwargs):
         if not self.name and self.top_level_dir:
             # Extract the last part of the path as the default name
             self.name = Path(self.top_level_dir).name or self.top_level_dir.strip("/")
+
+        # Set the index_name if not provided
+        if not self.index_name and self.capture_type:
+            # Convert string to CaptureType enum
+            capture_type_enum = CaptureType(self.capture_type)
+            self.index_name = infer_index_name(capture_type_enum)
+
         super().save(*args, **kwargs)
 
     @property

gateway/sds_gateway/api_methods/serializers/file_serializers.py

@@ -217,7 +217,6 @@ def check_file_contents_exist(
             user=user,
         )
 
-        log.debug(f"Checking file contents for user in directory: {safe_dir}")
         identical_file: File | None = identical_user_owned_file.filter(
             directory=safe_dir,
             name=name,
@@ -240,14 +239,12 @@ def check_file_contents_exist(
                     user_mutable_attributes_differ = True
                     break
 
-        payload = {
+        return {
             "file_exists_in_tree": identical_file is not None,
             "file_contents_exist_for_user": file_contents_exist_for_user,
             "user_mutable_attributes_differ": user_mutable_attributes_differ,
             "asset_id": asset.uuid if asset else None,
         }
-        log.debug(payload)
-        return payload
 
 
 class FileCheckResponseSerializer(serializers.Serializer[File]):

gateway/sds_gateway/api_methods/utils/metadata_schemas.py

@@ -3,9 +3,11 @@
 # the mapping below is used for drf capture metadata parsing in extract_drf_metadata.py
 
 import logging
+from typing import TYPE_CHECKING
 from typing import Any
 
-from sds_gateway.api_methods.models import CaptureType
+if TYPE_CHECKING:
+    from sds_gateway.api_methods.models import CaptureType
 
 log = logging.getLogger(__name__)
 
@@ -357,10 +359,9 @@
     "capture_props",
 ]
 
-capture_index_mapping_by_type: dict[CaptureType, dict[str, dict[str, Any]]] = {
-    CaptureType.DigitalRF: drf_capture_index_mapping,
-    CaptureType.RadioHound: rh_capture_index_mapping,
-}
+# Initialize empty dictionary to avoid circular import issues
+# Will be populated dynamically in get_mapping_for_capture_type()
+capture_index_mapping_by_type: dict[Any, dict[str, dict[str, Any]]] = {}
 
 base_properties = {
     "channel": {"type": "keyword"},
@@ -387,9 +388,20 @@
 
 
 def get_mapping_by_capture_type(
-    capture_type: CaptureType,
+    capture_type: Any,
 ) -> dict[str, str | dict[str, Any]]:
     """Get the mapping for a given capture type."""
+    # Local import to avoid circular dependency
+    from sds_gateway.api_methods.models import CaptureType  # noqa: PLC0415
+
+    # Initialize mapping if not already done
+    if not capture_index_mapping_by_type:
+        capture_index_mapping_by_type.update(
+            {
+                CaptureType.DigitalRF: drf_capture_index_mapping,
+                CaptureType.RadioHound: rh_capture_index_mapping,
+            }
+        )
 
     return {
         "properties": {
@@ -406,14 +418,17 @@ def get_mapping_by_capture_type(
     }
 
 
-def infer_index_name(capture_type: CaptureType) -> str:
+def infer_index_name(capture_type: "CaptureType") -> str:
     """Infer the index name for a given capture."""
-    # Populate index_name based on capture type
+    # Local import to avoid circular dependency
+    from sds_gateway.api_methods.models import CaptureType  # noqa: PLC0415
+
+    # Handle enum inputs (strings match fine against StrEnum)
     match capture_type:
         case CaptureType.DigitalRF:
-            return f"captures-{CaptureType.DigitalRF}"
+            return f"captures-{capture_type.value}"
         case CaptureType.RadioHound:
-            return f"captures-{CaptureType.RadioHound}"
+            return f"captures-{capture_type.value}"
         case _:
             msg = f"Invalid capture type: {capture_type}"
             log.error(msg)