Add support for collection items in copy/move actions

[GoldenAnpu]

Description

Adds collection as a supported items type for copy / move actions: an entities collection is resolved into a flat list of images and transferred into the destination dataset without preserving source structure.

Changes

• resolve_collection_items() — resolves each collection item via api.entities_collection.get_info_by_id() + get_items() into a flat list of image items. Supports both default and aiSearch collection types. Falls back to the collection's own project_id when source.project.id is not set in the state.
• copy_collection_items_to_dataset() / move_collection_items_to_dataset() — dedicated processing path: all images land in the destination dataset without preserving source structure. Move keeps the existing safety net: source items are not deleted unless every item was transferred.
• disambiguate_collection_names() — a collection can contain same-named images from different datasets; every image involved in a name conflict gets _<src dataset ID> appended before the extension (e.g. img.png from datasets 100 and 200 → img_100.png / img_200.png), so the origin of each copy stays visible. Residual collisions are resolved via generate_free_name. Without this, skip/replace modes silently dropped duplicates (and cancelled source deletion on move), while rename mode could fail the whole upload batch.
• rename_filtered_collection() — auto-created filter collections (named like Filtered entities 2026-07-03T14-31-57-501Z) are renamed to Data Commander (Task <ID>) when the task processes them, so it is clear which task handled the collection. Collections with custom names are left untouched; a rename failure does not fail the task.
• Conflicts with pre-existing destination images are still handled by options.conflictResolutionMode.
• Images listed both explicitly and via a collection are transferred once.

State example

{
  "state": {
    "action": "copy",
    "options": {"conflictResolutionMode": "rename", "cloneAnnotations": true, "preserveSrcDate": false},
    "items": [{"id": 2863, "type": "collection"}],
    "source": {"team": {"id": 10}, "workspace": {"id": 9}, "project": {"id": 7548}},
    "destination": {"team": {"id": 10}, "workspace": {"id": 9}, "project": {"id": 7549}, "dataset": {"id": 45188}}
  }
}

Notes

• Name disambiguation logic covered by standalone checks: two datasets, triple duplicate, name without extension, suffix collision with an existing name.