Add Bodytone DU30-B1D8 Fitness Bike

[netsoft-ruidias]

This pull request improves the ability of the FTMS device discovery process to correctly identify machine types, especially for devices that do not advertise their type in service data. The main enhancement is a fallback mechanism that inspects GATT characteristics if the machine type cannot be determined from advertisement data. Additionally, the new utility function for GATT-based type detection has been added and exported for broader use.

Enhancements to FTMS device discovery:

• Added a fallback mechanism in discover_ftms_devices to determine machine type by connecting to the device and inspecting its GATT characteristics when service data is insufficient. This increases compatibility with devices like the Bodytone DU30 that do not include machine type in their advertisements.

New utility for machine type detection:

• Implemented get_machine_type_from_gatt, an async function that determines the fitness machine type by checking for known FTMS data characteristics on a connected device.
• Exported get_machine_type_from_gatt in both src/pyftms/client/properties/__init__.py and src/pyftms/client/__init__.py for use elsewhere in the codebase. [1] [2]

Dependency and import updates:

• Updated imports in several modules to include BleakClient, BleakError, and the new utility function, ensuring all necessary components are available for the new fallback logic. [1] [2]

[Copilot]

Pull request overview

This PR adds a GATT-based fallback to FTMS discovery so devices that advertise the FTMS service but omit machine type service data can still be identified.

Changes:

• Adds get_machine_type_from_gatt to infer machine type from FTMS data characteristics.
• Uses the new helper as a fallback during discover_ftms_devices.
• Exports the helper from the properties package and imports it into the client package.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 5 comments.

File	Description
src/pyftms/client/properties/machine_type.py	Adds GATT characteristic-based machine type detection.
src/pyftms/client/properties/__init__.py	Re-exports the new GATT detection helper.
src/pyftms/client/__init__.py	Adds discovery fallback logic that connects to devices lacking machine type service data.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[Copilot]

Pull request overview

Copilot reviewed 3 out of 3 changed files in this pull request and generated 6 comments.

[netsoft-ruidias]

Discovery result:

(.venv) homeserver:~/python-pyftms$ python -m pyftms
Scanning for available FTMS devices...
1. INDOOR_BIKE: name: DU30-B1D8, address: 24:00:0C:A0:3B:98

Connection to INDOOR_BIKE: name: DU30-B1D8, address: 24:00:0C:A0:3B:98
 1. Device Info: {}
 2. Supported settings: ['target_resistance', 'target_power']
 3. Supported ranges: {'target_resistance': SettingRange(min_value=0, max_value=32.0, step=1.0), 'target_power': SettingRange(min_value=0, max_value=999, step=1)}
 4. Supported properties: ['speed_instant', 'speed_average', 'cadence_instant', 'cadence_average', 'distance_total', 'resistance_level', 'power_instant', 'power_average', 'energy_total', 'energy_per_hour', 'energy_per_minute', 'heart_rate', 'time_elapsed', 'training_status']
 5. Available properties: ['speed_instant', 'speed_average', 'cadence_instant', 'cadence_average', 'distance_total', 'resistance_level', 'power_instant', 'power_average', 'energy_total', 'energy_per_hour', 'energy_per_minute', 'heart_rate', 'metabolic_equivalent', 'time_elapsed', 'time_remaining', 'training_status']

Done.

[michaelw]

Hi @netsoft-ruidias, thank you for the Bodytone DU30-B1D8 work and for sharing the successful discovery output. I ported the GATT data-characteristic machine type detection idea from #67 into my pyftms fork on top of the UUID-only advertisement fallback and post-connect type correction work I already had there. See #70. There is also a drop-in HACS repo for the hassio-ftms fork, using the pyftms fork.

The fork now exposes a reusable get_machine_type_from_gatt(...) helper, uses it during the existing post-connect probe, and keeps discovery from opening GATT connections inside the advertisement loop. That also addresses the Copilot review notes about per-device discovery blocking/retries and keeps the scanner UUID filter behavior intact. The _cli disconnect guard and Python 3.14 support were already present in the fork, so those pieces are included there as well.

Would you mind trying the latest fork build with your Bodytone bike when you have a chance? The tag is v0.4.15+mw.6, and I’d really appreciate confirmation that it still discovers and connects cleanly to your DU30.