feat!: revert to_schema overloads & optimize upsert_many (#224)

- Reverts the previous `to_schema` implementation in the service layer.  This will need to be rethought as it requires too many overloads to any customized service function.  
- Ensures that the filter values for `upsert_many` is a unique list.  This is useful when you are merging a large amount of objects on a foreign key.  The lookup will only contain a single unique entry for each key instead of one for each row.
- Adds a `schema_to_dict` method to convert incoming Pydantic or Msgspec models to dictionaries
- Adds additional type guard helpers to check for a field in a dictionary, Struct, or BaseModel

advanced_alchemy/repository/_async.py

@@ -1584,10 +1584,7 @@ async def upsert_many(
                 matched_values = [
                     field_data for datum in data if (field_data := getattr(datum, field_name)) is not None
                 ]
-                if self._prefer_any:
-                    match_filter.append(any_(matched_values) == field)  # type: ignore[arg-type]
-                else:
-                    match_filter.append(field.in_(matched_values))
+                match_filter.append(any_(matched_values) == field if self._prefer_any else field.in_(matched_values))  # type: ignore[arg-type]
 
         with wrap_sqlalchemy_exception():
             existing_objs = await self.list(
@@ -1598,11 +1595,10 @@ async def upsert_many(
             )
             for field_name in match_fields:
                 field = get_instrumented_attr(self.model_type, field_name)
-                matched_values = [getattr(datum, field_name) for datum in existing_objs if datum]
-                if self._prefer_any:
-                    match_filter.append(any_(matched_values) == field)  # type: ignore[arg-type]
-                else:
-                    match_filter.append(field.in_(matched_values))
+                matched_values = list(
+                    {getattr(datum, field_name) for datum in existing_objs if datum},  # ensure the list is unique
+                )
+                match_filter.append(any_(matched_values) == field if self._prefer_any else field.in_(matched_values))  # type: ignore[arg-type]
             existing_ids = self._get_object_ids(existing_objs=existing_objs)
             data = self._merge_on_match_fields(data, existing_objs, match_fields)
             for datum in data:

advanced_alchemy/repository/_sync.py

@@ -1585,10 +1585,7 @@ def upsert_many(
                 matched_values = [
                     field_data for datum in data if (field_data := getattr(datum, field_name)) is not None
                 ]
-                if self._prefer_any:
-                    match_filter.append(any_(matched_values) == field)  # type: ignore[arg-type]
-                else:
-                    match_filter.append(field.in_(matched_values))
+                match_filter.append(any_(matched_values) == field if self._prefer_any else field.in_(matched_values))  # type: ignore[arg-type]
 
         with wrap_sqlalchemy_exception():
             existing_objs = self.list(
@@ -1599,11 +1596,10 @@ def upsert_many(
             )
             for field_name in match_fields:
                 field = get_instrumented_attr(self.model_type, field_name)
-                matched_values = [getattr(datum, field_name) for datum in existing_objs if datum]
-                if self._prefer_any:
-                    match_filter.append(any_(matched_values) == field)  # type: ignore[arg-type]
-                else:
-                    match_filter.append(field.in_(matched_values))
+                matched_values = list(
+                    {getattr(datum, field_name) for datum in existing_objs if datum},  # ensure the list is unique
+                )
+                match_filter.append(any_(matched_values) == field if self._prefer_any else field.in_(matched_values))  # type: ignore[arg-type]
             existing_ids = self._get_object_ids(existing_objs=existing_objs)
             data = self._merge_on_match_fields(data, existing_objs, match_fields)
             for datum in data: