fix(media_player): parse real OWNSoundEvent dimensions and handle malformed RDS frames

[GreenGrassBlueOcean]

Fixes #498.

Problem Summary

In #427 comment 5849429368, @manfredgittmaier-afk reported that on a live installation (MH200N + F500N tuner, OWNd 2.0.0b8), tuner command transmission worked (station select and next/previous stepped the tuner), but incoming reports (frequency, station, source, and media_title) never updated the entity:

• When stepping to next track (*16*6001*101##), the bus replied with dimension frames 6, 7, and 8 (*#16*101*6*0*103500##, *#16*101*7*0*4##, *#16*101*8*65*78*84*69*78*78*69*32##), but the entity remained on stale state.
• Some stations rotate RDS PS text every ~5s (Radio OÖ: *RADIO* / **OOE** ), and transient hardware glitches occasionally emit malformed dimension 8 frames with 12 ASCII codes instead of 8 (e.g. *#16*101*8*42*42*79*79*69*42*79*79*69*42*42*32##).

Root Cause Analysis

1. OWNSoundEvent dimension values attribute mismatch:
   In OWNd.message.OWNSoundEvent, dimension payload parameters are stored in the instance attribute _dimension_value (e.g. ['0', '103500']), while OWNd does not expose a public dimension_value property.
   In custom_components/myhome/sound_source.py, handle_event evaluated getattr(message, "dimension_value", None) or [], which always returned [] on real OWNSoundEvent instances. Because values was empty, all dimension branches (6, 7, and 8) were skipped. In unit tests, mocks had manually set dimension_value=..., masking the mismatch.
2. Malformed 12-code RDS frames:
   Transient hardware buffer glitches occasionally emitted dimension 8 frames containing 12 ASCII codes instead of the standard 8 codes.
3. State hygiene on frequency transition:
   Tuning away to a new frequency should clear any stale media_title from the previous station until new RDS arrives.

Changes

1. custom_components/myhome/sound_source.py:
   • Extract dimension value list falling back across dimension_value, _dimension_value, and dimension_values.
   • Safely convert dimension to integer.
   • For Dimension 8 (RDS), strictly enforce len(values) == 8; log and ignore malformed frames with length != 8 to prevent title corruption or blanking.
   • Clear _attr_media_title = None when frequency changes in _set_frequency_from_bus and async_set_frequency.
2. Tests & Authentic Fixtures:
   • Committed the full authentic MH200N + F500N live capture myhome_trace_MH200N_all_2026-09-26T20-00-28.json contributed by @manfredgittmaier-afk under tests/fixtures/traces/f500_tuner/.
   • Updated existing replay tests in tests/test_sound_source_trace_replay.py to instantiate real OWNSoundEvent objects directly from raw bus frames.
   • Added test_f500n_station_advance_and_rds_rotation_replay replaying the full trace and verifying rotating RDS, 12-code glitch rejection, and station advance across presets 1–4.
   • Added unit tests in tests/test_component_sound_source.py for real OWNSoundEvent handling, malformed RDS rejection, and source following after next/previous.

Verification

• 2,395 passing tests (100% pass rate).
• 100.0% coverage maintained across all 43 modules (8,295 / 8,295 statements covered, 0 missing lines).
• ruff check and scripts/verify_ha_standards.py clean with 0 violations.

[codecov-commenter]

⚠️ Please install the to ensure uploads and comments are reliably processed by Codecov.

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[manfredgittmaier-afk]

Confirmed working on the MH200N + F500N – frequency, station, source and media_title now follow next/previous and preset changes, and the rotating RDS text (Radio OÖ) updates cleanly. Thanks!

[manfredgittmaier-afk]

@GreenGrassBlueOcean – one more from the MH200N + F441M: volume down on the audio zones still doesn't work.

Bus monitor, amplifier 41:

Button	TX	Reply
Volume down (pressed twice)	*16*1000*41##	none – ignored by the amplifier
Volume up	*16*1001*41##	*#16*41*1*7## + *#22*3#4#1*1*7##

I know this is fixed in OWNd#49 (volume down → *16*1101*<zone>##), but MyHOME still pins OWNd 2.0.0b8, so it doesn't reach users yet. Is an OWNd release planned soon – or could the zone's async_volume_down build *16*1101*<zone>## directly in the meantime, like the tuner entity does for its commands?

[GreenGrassBlueOcean]

Yes planning on a new release. @fedem95 did a lot of great reviewing already. Waiting on his review and than we make a new ownd release