Merge pull request #591 from zxzxwu/auracast_scan

Improve Broadcast Scanning

apps/auracast.py

@@ -60,7 +60,7 @@
 class BroadcastScanner(pyee.EventEmitter):
     @dataclasses.dataclass
     class Broadcast(pyee.EventEmitter):
-        name: str
+        name: str | None
         sync: bumble.device.PeriodicAdvertisingSync
         rssi: int = 0
         public_broadcast_announcement: Optional[
@@ -135,7 +135,8 @@ def print(self) -> None:
                 self.sync.advertiser_address,
                 color(self.sync.state.name, 'green'),
             )
-            print(f'  {color("Name", "cyan")}:         {self.name}')
+            if self.name is not None:
+                print(f'  {color("Name", "cyan")}:         {self.name}')
             if self.appearance:
                 print(f'  {color("Appearance", "cyan")}:   {str(self.appearance)}')
             print(f'  {color("RSSI", "cyan")}:         {self.rssi}')
@@ -174,7 +175,7 @@ def print(self) -> None:
                     print(color('      Codec ID:', 'yellow'))
                     print(
                         color('        Coding Format:           ', 'green'),
-                        subgroup.codec_id.coding_format.name,
+                        subgroup.codec_id.codec_id.name,
                     )
                     print(
                         color('        Company ID:              ', 'green'),
@@ -274,13 +275,24 @@ async def stop(self) -> None:
         await self.device.stop_scanning()
 
     def on_advertisement(self, advertisement: bumble.device.Advertisement) -> None:
-        if (
-            broadcast_name := advertisement.data.get(
-                bumble.core.AdvertisingData.BROADCAST_NAME
+        if not (
+            ads := advertisement.data.get_all(
+                bumble.core.AdvertisingData.SERVICE_DATA_16_BIT_UUID
             )
-        ) is None:
+        ) or not (
+            any(
+                ad
+                for ad in ads
+                if isinstance(ad, tuple)
+                and ad[0] == bumble.gatt.GATT_BROADCAST_AUDIO_ANNOUNCEMENT_SERVICE
+            )
+        ):
             return
-        assert isinstance(broadcast_name, str)
+
+        broadcast_name = advertisement.data.get(
+            bumble.core.AdvertisingData.BROADCAST_NAME
+        )
+        assert isinstance(broadcast_name, str) or broadcast_name is None
 
         if broadcast := self.broadcasts.get(advertisement.address):
             broadcast.update(advertisement)
@@ -291,18 +303,15 @@ def on_advertisement(self, advertisement: bumble.device.Advertisement) -> None:
         )
 
     async def on_new_broadcast(
-        self, name: str, advertisement: bumble.device.Advertisement
+        self, name: str | None, advertisement: bumble.device.Advertisement
     ) -> None:
         periodic_advertising_sync = await self.device.create_periodic_advertising_sync(
             advertiser_address=advertisement.address,
             sid=advertisement.sid,
             sync_timeout=self.sync_timeout,
             filter_duplicates=self.filter_duplicates,
         )
-        broadcast = self.Broadcast(
-            name,
-            periodic_advertising_sync,
-        )
+        broadcast = self.Broadcast(name, periodic_advertising_sync)
         broadcast.update(advertisement)
         self.broadcasts[advertisement.address] = broadcast
         periodic_advertising_sync.on('loss', lambda: self.on_broadcast_loss(broadcast))

apps/lea_unicast/app.py

@@ -486,7 +486,12 @@ async def run(self) -> None:
 
             def on_pdu(pdu: HCI_IsoDataPacket, ase: ascs.AseStateMachine):
                 codec_config = ase.codec_specific_configuration
-                assert isinstance(codec_config, bap.CodecSpecificConfiguration)
+                if (
+                    not isinstance(codec_config, bap.CodecSpecificConfiguration)
+                    or codec_config.frame_duration is None
+                    or codec_config.audio_channel_allocation is None
+                ):
+                    return
                 pcm = decode(
                     codec_config.frame_duration.us,
                     codec_config.audio_channel_allocation.channel_count,
@@ -495,11 +500,17 @@ def on_pdu(pdu: HCI_IsoDataPacket, ase: ascs.AseStateMachine):
                 self.device.abort_on('disconnection', self.ui_server.send_audio(pcm))
 
             def on_ase_state_change(ase: ascs.AseStateMachine) -> None:
+                codec_config = ase.codec_specific_configuration
                 if ase.state == ascs.AseStateMachine.State.STREAMING:
-                    codec_config = ase.codec_specific_configuration
-                    assert isinstance(codec_config, bap.CodecSpecificConfiguration)
-                    assert ase.cis_link
                     if ase.role == ascs.AudioRole.SOURCE:
+                        if (
+                            not isinstance(codec_config, bap.CodecSpecificConfiguration)
+                            or ase.cis_link is None
+                            or codec_config.octets_per_codec_frame is None
+                            or codec_config.frame_duration is None
+                            or codec_config.codec_frames_per_sdu is None
+                        ):
+                            return
                         ase.cis_link.abort_on(
                             'disconnection',
                             lc3_source_task(
@@ -514,10 +525,17 @@ def on_ase_state_change(ase: ascs.AseStateMachine) -> None:
                             ),
                         )
                     else:
+                        if not ase.cis_link:
+                            return
                         ase.cis_link.sink = functools.partial(on_pdu, ase=ase)
                 elif ase.state == ascs.AseStateMachine.State.CODEC_CONFIGURED:
-                    codec_config = ase.codec_specific_configuration
-                    assert isinstance(codec_config, bap.CodecSpecificConfiguration)
+                    if (
+                        not isinstance(codec_config, bap.CodecSpecificConfiguration)
+                        or codec_config.sampling_frequency is None
+                        or codec_config.frame_duration is None
+                        or codec_config.audio_channel_allocation is None
+                    ):
+                        return
                     if ase.role == ascs.AudioRole.SOURCE:
                         setup_encoders(
                             codec_config.sampling_frequency.hz,

bumble/core.py

@@ -1624,6 +1624,9 @@ def __bytes__(self):
             [bytes([len(x[1]) + 1, x[0]]) + x[1] for x in self.ad_structures]
         )
 
+    def to_bytes(self) -> bytes:
+        return bytes(self)
+
     def to_string(self, separator=', '):
         return separator.join(
             [AdvertisingData.ad_data_to_string(x[0], x[1]) for x in self.ad_structures]

bumble/hci.py

@@ -1482,7 +1482,7 @@ class CodingFormat:
     vendor_specific_codec_id: int = 0
 
     @classmethod
-    def parse_from_bytes(cls, data: bytes, offset: int):
+    def parse_from_bytes(cls, data: bytes, offset: int) -> tuple[int, CodingFormat]:
         (codec_id, company_id, vendor_specific_codec_id) = struct.unpack_from(
             '<BHH', data, offset
         )
@@ -1492,6 +1492,10 @@ def parse_from_bytes(cls, data: bytes, offset: int):
             vendor_specific_codec_id=vendor_specific_codec_id,
         )
 
+    @classmethod
+    def from_bytes(cls, data: bytes) -> CodingFormat:
+        return cls.parse_from_bytes(data, 0)[1]
+
     def to_bytes(self) -> bytes:
         return struct.pack(
             '<BHH', self.codec_id, self.company_id, self.vendor_specific_codec_id

bumble/profiles/bap.py

@@ -398,18 +398,21 @@ class Type(utils.OpenIntEnum):
         OCTETS_PER_FRAME         = 0x04
         CODEC_FRAMES_PER_SDU     = 0x05
 
-    sampling_frequency: SamplingFrequency
-    frame_duration: FrameDuration
-    audio_channel_allocation: AudioLocation
-    octets_per_codec_frame: int
-    codec_frames_per_sdu: int
+    sampling_frequency: SamplingFrequency | None = None
+    frame_duration: FrameDuration | None = None
+    audio_channel_allocation: AudioLocation | None = None
+    octets_per_codec_frame: int | None = None
+    codec_frames_per_sdu: int | None = None
 
     @classmethod
     def from_bytes(cls, data: bytes) -> CodecSpecificConfiguration:
         offset = 0
-        # Allowed default values.
-        audio_channel_allocation = AudioLocation.NOT_ALLOWED
-        codec_frames_per_sdu = 1
+        sampling_frequency: SamplingFrequency | None = None
+        frame_duration: FrameDuration | None = None
+        audio_channel_allocation: AudioLocation | None = None
+        octets_per_codec_frame: int | None = None
+        codec_frames_per_sdu: int | None = None
+
         while offset < len(data):
             length, type = struct.unpack_from('BB', data, offset)
             offset += 2
@@ -427,8 +430,6 @@ def from_bytes(cls, data: bytes) -> CodecSpecificConfiguration:
             elif type == CodecSpecificConfiguration.Type.CODEC_FRAMES_PER_SDU:
                 codec_frames_per_sdu = value
 
-        # It is expected here that if some fields are missing, an error should be raised.
-        # pylint: disable=possibly-used-before-assignment,used-before-assignment
         return CodecSpecificConfiguration(
             sampling_frequency=sampling_frequency,
             frame_duration=frame_duration,
@@ -438,23 +439,43 @@ def from_bytes(cls, data: bytes) -> CodecSpecificConfiguration:
         )
 
     def __bytes__(self) -> bytes:
-        return struct.pack(
-            '<BBBBBBBBIBBHBBB',
-            2,
-            CodecSpecificConfiguration.Type.SAMPLING_FREQUENCY,
-            self.sampling_frequency,
-            2,
-            CodecSpecificConfiguration.Type.FRAME_DURATION,
-            self.frame_duration,
-            5,
-            CodecSpecificConfiguration.Type.AUDIO_CHANNEL_ALLOCATION,
-            self.audio_channel_allocation,
-            3,
-            CodecSpecificConfiguration.Type.OCTETS_PER_FRAME,
-            self.octets_per_codec_frame,
-            2,
-            CodecSpecificConfiguration.Type.CODEC_FRAMES_PER_SDU,
-            self.codec_frames_per_sdu,
+        return b''.join(
+            [
+                struct.pack(fmt, length, tag, value)
+                for fmt, length, tag, value in [
+                    (
+                        '<BBB',
+                        2,
+                        CodecSpecificConfiguration.Type.SAMPLING_FREQUENCY,
+                        self.sampling_frequency,
+                    ),
+                    (
+                        '<BBB',
+                        2,
+                        CodecSpecificConfiguration.Type.FRAME_DURATION,
+                        self.frame_duration,
+                    ),
+                    (
+                        '<BBI',
+                        5,
+                        CodecSpecificConfiguration.Type.AUDIO_CHANNEL_ALLOCATION,
+                        self.audio_channel_allocation,
+                    ),
+                    (
+                        '<BBH',
+                        3,
+                        CodecSpecificConfiguration.Type.OCTETS_PER_FRAME,
+                        self.octets_per_codec_frame,
+                    ),
+                    (
+                        '<BBB',
+                        2,
+                        CodecSpecificConfiguration.Type.CODEC_FRAMES_PER_SDU,
+                        self.codec_frames_per_sdu,
+                    ),
+                ]
+                if value is not None
+            ]
         )
 
 
@@ -466,6 +487,25 @@ class BroadcastAudioAnnouncement:
     def from_bytes(cls, data: bytes) -> Self:
         return cls(int.from_bytes(data[:3], 'little'))
 
+    def to_bytes(self) -> bytes:
+        return self.broadcast_id.to_bytes(3, 'little')
+
+    def __bytes__(self) -> bytes:
+        return self.to_bytes()
+
+    def get_advertising_data(self) -> bytes:
+        return core.AdvertisingData(
+            [
+                (
+                    core.AdvertisingData.SERVICE_DATA_16_BIT_UUID,
+                    (
+                        gatt.GATT_BROADCAST_AUDIO_ANNOUNCEMENT_SERVICE.to_bytes()
+                        + self.to_bytes()
+                    ),
+                )
+            ]
+        ).to_bytes()
+
 
 @dataclasses.dataclass
 class BasicAudioAnnouncement:
@@ -474,26 +514,43 @@ class BIS:
         index: int
         codec_specific_configuration: CodecSpecificConfiguration
 
-    @dataclasses.dataclass
-    class CodecInfo:
-        coding_format: hci.CodecID
-        company_id: int
-        vendor_specific_codec_id: int
-
-        @classmethod
-        def from_bytes(cls, data: bytes) -> Self:
-            coding_format = hci.CodecID(data[0])
-            company_id = int.from_bytes(data[1:3], 'little')
-            vendor_specific_codec_id = int.from_bytes(data[3:5], 'little')
-            return cls(coding_format, company_id, vendor_specific_codec_id)
+        def to_bytes(self) -> bytes:
+            codec_specific_configuration_bytes = bytes(
+                self.codec_specific_configuration
+            )
+            return (
+                bytes([self.index, len(codec_specific_configuration_bytes)])
+                + codec_specific_configuration_bytes
+            )
+
+        def __bytes__(self) -> bytes:
+            return self.to_bytes()
 
     @dataclasses.dataclass
     class Subgroup:
-        codec_id: BasicAudioAnnouncement.CodecInfo
+        codec_id: hci.CodingFormat
         codec_specific_configuration: CodecSpecificConfiguration
         metadata: le_audio.Metadata
         bis: List[BasicAudioAnnouncement.BIS]
 
+        def to_bytes(self) -> bytes:
+            metadata_bytes = bytes(self.metadata)
+            codec_specific_configuration_bytes = bytes(
+                self.codec_specific_configuration
+            )
+            return (
+                bytes([len(self.bis)])
+                + self.codec_id.to_bytes()
+                + bytes([len(codec_specific_configuration_bytes)])
+                + codec_specific_configuration_bytes
+                + bytes([len(metadata_bytes)])
+                + metadata_bytes
+                + b''.join(map(bytes, self.bis))
+            )
+
+        def __bytes__(self) -> bytes:
+            return self.to_bytes()
+
     presentation_delay: int
     subgroups: List[BasicAudioAnnouncement.Subgroup]
 
@@ -505,7 +562,7 @@ def from_bytes(cls, data: bytes) -> Self:
         for _ in range(data[3]):
             num_bis = data[offset]
             offset += 1
-            codec_id = cls.CodecInfo.from_bytes(data[offset : offset + 5])
+            codec_id = hci.CodingFormat.from_bytes(data[offset : offset + 5])
             offset += 5
             codec_specific_configuration_length = data[offset]
             offset += 1
@@ -549,3 +606,26 @@ def from_bytes(cls, data: bytes) -> Self:
             )
 
         return cls(presentation_delay, subgroups)
+
+    def to_bytes(self) -> bytes:
+        return (
+            self.presentation_delay.to_bytes(3, 'little')
+            + bytes([len(self.subgroups)])
+            + b''.join(map(bytes, self.subgroups))
+        )
+
+    def __bytes__(self) -> bytes:
+        return self.to_bytes()
+
+    def get_advertising_data(self) -> bytes:
+        return core.AdvertisingData(
+            [
+                (
+                    core.AdvertisingData.SERVICE_DATA_16_BIT_UUID,
+                    (
+                        gatt.GATT_BASIC_AUDIO_ANNOUNCEMENT_SERVICE.to_bytes()
+                        + self.to_bytes()
+                    ),
+                )
+            ]
+        ).to_bytes()

examples/run_unicast_server.py

@@ -161,7 +161,13 @@ def on_ase_state_change(
             else:
                 file_output = open(f'{datetime.datetime.now().isoformat()}.lc3', 'wb')
                 codec_configuration = ase.codec_specific_configuration
-                assert isinstance(codec_configuration, CodecSpecificConfiguration)
+                if (
+                    not isinstance(codec_configuration, CodecSpecificConfiguration)
+                    or codec_configuration.sampling_frequency is None
+                    or codec_configuration.audio_channel_allocation is None
+                    or codec_configuration.frame_duration is None
+                ):
+                    return
                 # Write a LC3 header.
                 file_output.write(
                     bytes([0x1C, 0xCC])  # Header.