Capability flags (generated)¶

Generated from custom_components/eufy_vacuum/core/capabilities.py (KNOWN_CAPABILITY_HINTS, detect_capabilities), custom_components/eufy_vacuum/adapters/config_schema.py (ADAPTER_CONFIG_SCHEMA['capabilities']) and custom_components/eufy_vacuum/adapters/registry.py (_validate_adapter).

Three namespaces, one vocabulary¶

Namespace	Enumerated by	Size	Validated at registration	Consumed by
`capabilities` config block	`ADAPTER_CONFIG_SCHEMA['capabilities']['fields']`	21	yes - the schema walker checks types and rejects unknown keys	`custom_components/eufy_vacuum/adapters/registry.py`, `custom_components/eufy_vacuum/core/manager.py`, `custom_components/eufy_vacuum/dispatch/manager.py`, `custom_components/eufy_vacuum/jobs/active_job.py`, `custom_components/eufy_vacuum/room_entities.py`
`capability_hints` config block	`KNOWN_CAPABILITY_HINTS` (a different file)	13	yes - `_validate_adapter` rejects any key outside that frozenset	`detect_capabilities()`
`detect_capabilities()` return	the return dict literal	42 keys, 22 of them `supports_*`	n/a (output, not input)	the per-vacuum capability payload

The capabilities block is fully enumerated in the schema and is never read by detect_capabilities(). The capability_hints block is the opposite: the schema declares it as a bare dict with no fields, so the schema walker cannot see inside it; _validate_adapter checks it instead, against a frozenset that lives in custom_components/eufy_vacuum/core/capabilities.py. The two share key names and are different dictionaries with different consumers.

Config-declarable vs detected-only¶

Flag	`capabilities` block	hintable	in detection output	hint rule
`has_attribute_rooms`	no	yes	no	permissive
`honors_clean_order`	yes	no	no	-
`position_lock_reliable`	yes	no	no	-
`rooms_unique_per_job`	yes	no	no	-
`supports_active_cleaning_target`	no	no	yes	-
`supports_active_map`	no	no	yes	-
`supports_cleaning_stats`	no	no	yes	-
`supports_custom_room_config`	no	yes	yes	authoritative
`supports_dock_status`	no	no	yes	-
`supports_edge_mopping`	yes	yes	yes	authoritative
`supports_empty_dust`	yes	yes	yes	permissive
`supports_goto`	yes	yes	yes	authoritative
`supports_mop_dry`	yes	yes	yes	permissive
`supports_mop_features`	yes	yes	yes	permissive
`supports_mop_wash`	yes	yes	yes	permissive
`supports_passes`	no	yes	yes	authoritative
`supports_path_control`	yes	yes	yes	permissive
`supports_robot_position`	yes	no	yes	-
`supports_room_clean`	no	yes	yes	authoritative
`supports_room_profiles`	yes	no	no	-
`supports_rooms`	no	no	yes	-
`supports_segments`	no	no	yes	-
`supports_station_water`	yes	no	yes	-
`supports_task_status`	no	no	yes	-
`supports_water_control`	yes	yes	yes	authoritative
`supports_work_mode`	no	no	yes	-
`supports_zone_clean`	yes	yes	yes	authoritative
`supports_zone_repeat`	yes	no	no	-
`zone_max`	yes	no	no	-
`zone_max_area_m2`	yes	no	no	-
`zone_max_side_m`	yes	no	no	-
`zone_min_area_m2`	yes	no	no	-
`zone_min_side_m`	yes	no	no	-

Detected-only (8) - emitted by detect_capabilities(), declarable in neither config namespace: supports_active_cleaning_target, supports_active_map, supports_cleaning_stats, supports_dock_status, supports_rooms, supports_segments, supports_task_status, supports_work_mode.
Config-block-only (10) - declarable in the capabilities block, never emitted by detection and not a valid hint: honors_clean_order, position_lock_reliable, rooms_unique_per_job, supports_room_profiles, supports_zone_repeat, zone_max, zone_max_area_m2, zone_max_side_m, zone_min_area_m2, zone_min_side_m.
In both config namespaces (9): supports_edge_mopping, supports_empty_dust, supports_goto, supports_mop_dry, supports_mop_features, supports_mop_wash, supports_path_control, supports_water_control, supports_zone_clean.
Hint-only (4) - a valid capability_hints key with no capabilities block counterpart: has_attribute_rooms, supports_custom_room_config, supports_passes, supports_room_clean.

How each detected flag is derived¶

permissive = hint OR entity presence (True from either source is enough). authoritative = _hint_wins: an explicit hint overrides the derived default, so a brand can declare a categorical False. - = no hint is consulted.

Flag	Hint rule	Derivation
`supports_active_cleaning_target`	-	`bool(active_cleaning_target_registered or active_cleaning_target_entity)`
`supports_active_map`	-	`bool(active_map_registered or active_map_entity)`
`supports_cleaning_stats`	-	`bool(cleaning_area_registered or cleaning_time_registered or cleaning_area_entity or cleaning_time_entity)`
`supports_custom_room_config`	authoritative	`hint if declared, else True`
`supports_dock_status`	-	`bool(dock_status_registered or dock_status_entity)`
`supports_edge_mopping`	authoritative	`hint if declared, else True`
`supports_empty_dust`	permissive	`bool(_hints.get('supports_empty_dust')) or _empty_dust_entity_present`
`supports_goto`	authoritative	`hint if declared, else False`
`supports_mop_dry`	permissive	`bool(_hints.get('supports_mop_dry')) or _dry_mop_entity_present`
`supports_mop_features`	permissive	`bool(_hints.get('supports_mop_features')) or bool(water_level_registered or water_level_entity)`
`supports_mop_wash`	permissive	`bool(_hints.get('supports_mop_wash')) or _wash_mop_entity_present`
`supports_passes`	authoritative	`hint if declared, else True`
`supports_path_control`	permissive	`bool(_hints.get('supports_path_control')) or _cleaning_intensity_present`
`supports_robot_position`	-	`bool(robot_position_x_entity_id and robot_position_y_entity_id)`
`supports_room_clean`	authoritative	`hint if declared, else True`
`supports_rooms`	-	`bool(active_map_registered or active_map_entity or _has_attribute_rooms)`
`supports_segments`	-	`supports_rooms`
`supports_station_water`	-	`bool(water_level_registered or water_level_entity)`
`supports_task_status`	-	`bool(task_status_registered or task_status_entity)`
`supports_water_control`	authoritative	`hint if declared, else supports_mop_features`
`supports_work_mode`	-	`bool(work_mode_registered or work_mode_entity)`
`supports_zone_clean`	authoritative	`hint if declared, else True`

10 of the 22 supports_* flags consult no hint at all: supports_active_cleaning_target, supports_active_map, supports_cleaning_stats, supports_dock_status, supports_robot_position, supports_rooms, supports_segments, supports_station_water, supports_task_status, supports_work_mode.

`supports_` vs `_available`¶

Detect and return a capability map for one vacuum.

When adapter inputs are provided, entity probing uses the supplied candidate lists and model-based hints. When absent (no registered adapter), returns a minimal capability set derived from the vacuum entity alone.

"support" flags indicate the entity exists in the registry or the model supports the feature; "available" flags indicate the entity is currently in the state machine and usable now.

Flag	Value
`active_cleaning_target_available`	`bool(active_cleaning_target_entity)`
`active_map_available`	`bool(active_map_entity)`
`cleaning_stats_available`	`bool(cleaning_area_entity or cleaning_time_entity)`
`dock_status_available`	`bool(dock_status_entity)`
`robot_position_available`	`robot_position_available`
`station_water_available`	`bool(water_level_entity)`
`task_status_available`	`bool(task_status_entity)`
`work_mode_available`	`bool(work_mode_entity)`

What the two shipped adapters declare¶

Flag	eufy `capabilities`	roborock `capabilities`	dreame `capabilities`	eufy `capability_hints`	roborock `capability_hints`	dreame `capability_hints`
`has_attribute_rooms`	-	-	-	`has_attribute_rooms`	-	-
`honors_clean_order`	-	`False`	`True`	-	-	-
`position_lock_reliable`	`False`	`False`	-	-	-	-
`rooms_unique_per_job`	`True`	`False`	-	-	-	-
`supports_custom_room_config`	-	-	-	-	-	-
`supports_edge_mopping`	`caps.get('supports_edge_mopping', True)`	`False`	`False`	-	-	`False`
`supports_empty_dust`	`caps.get('supports_empty_dust', False)`	`False`	`caps.get('supports_empty_dust', station_collectable)`	`model_family in {'x10', 'x8', 'l60', 'l50'}`	-	`station_collectable`
`supports_goto`	-	-	`caps.get('supports_goto', False)`	-	-	`profile.get('has_path_control', False)`
`supports_mop_dry`	`caps.get('supports_mop_dry', False)`	`False`	`caps.get('supports_mop_dry', station_dryable)`	`model_family in {'x10', 'x8'}`	-	`station_dryable`
`supports_mop_features`	`caps.get('supports_mop_features', False)`	`caps.get('supports_mop_features', profile['has_mop'])`	`caps.get('supports_mop_features', profile['has_mop'])`	`model_family in {'x10', 'x8', 'l60', 'l50'}`	-	`profile['has_mop']`
`supports_mop_wash`	`caps.get('supports_mop_wash', False)`	`False`	`caps.get('supports_mop_wash', station_washable)`	`model_family in {'x10', 'x8'}`	-	`station_washable`
`supports_passes`	-	-	-	-	-	-
`supports_path_control`	`caps.get('supports_path_control', False)`	`profile.get('has_path_control', False)`	`profile.get('has_path_control', True)`	`model_family in {'x10', 'x8'}`	-	`profile.get('has_path_control', False)`
`supports_robot_position`	`caps.get('supports_robot_position', False)`	`caps.get('supports_robot_position', False)`	-	-	-	-
`supports_room_clean`	-	-	-	-	-	-
`supports_room_profiles`	-	`mop_settable`	-	-	-	-
`supports_station_water`	`caps.get('supports_station_water', False)`	`False`	-	-	-	-
`supports_water_control`	`caps.get('supports_water_control', False)`	`mop_settable`	`mop_settable`	-	-	-
`supports_zone_clean`	`caps.get('supports_zone_clean', True)`	`caps.get('supports_zone_clean', True)`	`caps.get('supports_zone_clean', False)`	-	-	`profile.get('supports_zone_clean', False)`
`supports_zone_repeat`	`False`	-	-	-	-	-
`zone_max`	`10`	`5`	`10`	-	-	-
`zone_max_area_m2`	-	`3.05`	-	-	-	-
`zone_max_side_m`	`10.0`	-	-	-	-	-
`zone_min_area_m2`	-	`0.0929`	-	-	-	-
`zone_min_side_m`	`0.5`	-	-	-	-	-

The hint contract¶

Every flag capability_hints may carry — the ONLY keys this module reads. A hint is a silent no-op if the key is not in this set: _hints.get(name) simply misses and the derived/default value stands. That makes a typo indistinguishable from an undeclared capability, which is the same silent-wrong-answer shape as the hardcoded True that _hint_wins exists to prevent — a brand declaring supports_zone_cleen: False would be ignored exactly as thoroughly as one declaring nothing. registry._validate_adapter checks adapter capability_hints against this set so the typo is caught at registration instead of at "why is this control showing?". Keep in sync with the _hints reads below; the contract test asserts it matches.

Valid hint keys: has_attribute_rooms, supports_custom_room_config, supports_edge_mopping, supports_empty_dust, supports_goto, supports_mop_dry, supports_mop_features, supports_mop_wash, supports_passes, supports_path_control, supports_room_clean, supports_water_control, supports_zone_clean.