Attempt to fix the relays projection in a generic way.

Author: matteius

opensensor/collection_apis.py

@@ -1,6 +1,17 @@
 import logging
 from datetime import datetime, timedelta, timezone
-from typing import Dict, Generic, List, Optional, Type, TypeVar, get_args, get_origin
+from enum import Enum
+from typing import (
+    Any,
+    Dict,
+    Generic,
+    List,
+    Optional,
+    Type,
+    TypeVar,
+    get_args,
+    get_origin,
+)
 
 from bson import Binary
 from fastapi import APIRouter, Depends, HTTPException, Path, Query, Response, status
@@ -184,10 +195,6 @@ def get_initial_match_clause(
     return match_clause
 
 
-def is_pydantic_model(obj):
-    return isinstance(obj, type) and issubclass(obj, BaseModel)
-
-
 def get_nested_fields(model: Type[BaseModel]) -> Dict[str, Type[BaseModel]]:
     nested_fields = {}
     for field_name, field in model.__fields__.items():
@@ -198,17 +205,14 @@ def get_nested_fields(model: Type[BaseModel]) -> Dict[str, Type[BaseModel]]:
     return nested_fields
 
 
-def is_list_of_models(field: ModelField) -> bool:
-    if field.shape != 2:  # 2 represents a list shape in Pydantic
-        return False
-    args = field.sub_fields[0].type_.__args__
-    return args and is_pydantic_model(args[0])
-
-
 def get_list_item_type(field: ModelField) -> Type[BaseModel]:
     return field.sub_fields[0].type_.__args__[0]
 
 
+def is_enum(field_type):
+    return issubclass(field_type, Enum)
+
+
 def create_nested_pipeline(model: Type[BaseModel], prefix=""):
     logger.debug(f"Creating nested pipeline for model: {model.__name__}, prefix: {prefix}")
     match_conditions = {}
@@ -229,13 +233,21 @@ def create_nested_pipeline(model: Type[BaseModel], prefix=""):
             unit_field_name = f"{prefix}{mongo_field}_unit"
             pipeline["unit"] = f"${unit_field_name}"
             match_conditions[unit_field_name] = {"$exists": True}
-        elif is_list_of_models(field):
-            item_model = get_list_item_type(field)
+        elif is_pydantic_model(field.type_):
+            nested_pipeline, nested_match = create_nested_pipeline(field.type_, f"{field_name}.")
+            pipeline[field_name] = nested_pipeline
+            match_conditions.update({f"{field_name}.{k}": v for k, v in nested_match.items()})
+        elif get_origin(field.type_) is List and is_pydantic_model(get_args(field.type_)[0]):
+            item_model = get_args(field.type_)[0]
             nested_pipeline, nested_match = create_nested_pipeline(item_model, "")
             pipeline[field_name] = {
                 "$map": {"input": f"${full_mongo_field_name}", "as": "item", "in": nested_pipeline}
             }
             match_conditions[full_mongo_field_name] = {"$exists": True, "$ne": []}
+        elif is_enum(field.type_):
+            # Handle enum fields as simple fields
+            pipeline[field_name] = f"${full_mongo_field_name}"
+            match_conditions[full_mongo_field_name] = {"$exists": True}
         else:
             # Handle simple field
             pipeline[field_name] = f"${full_mongo_field_name}"
@@ -248,6 +260,10 @@ def create_nested_pipeline(model: Type[BaseModel], prefix=""):
     return pipeline, match_conditions
 
 
+def is_pydantic_model(obj: Any) -> bool:
+    return isinstance(obj, type) and issubclass(obj, BaseModel)
+
+
 def create_model_instance(model: Type[BaseModel], data: dict, target_unit: Optional[str] = None):
     nested_fields = get_nested_fields(model)