Added transcription packet parser. (#34768)

* Added Transcription packet parser.

* fixed lint errors.

* updating and fixing the module import error

* Fixing module not found issue

* fixing the tests

---------

Co-authored-by: Vinothini Dharmaraj <[email protected]>

Author: v-durgeshs

sdk/communication/azure-communication-callautomation/azure/communication/callautomation/__init__.py

@@ -8,6 +8,7 @@
 from ._version import VERSION
 from ._call_automation_client import CallAutomationClient
 from ._call_connection_client import CallConnectionClient
+from .streaming.streaming_data_parser import StreamingDataParser
 from ._models import (
     CallConnectionProperties,
     FileSource,
@@ -52,11 +53,19 @@
     RecordingState,
     VoiceKind
 )
+from .streaming.models import (
+    TranscriptionMetadata,
+    TranscriptionData
+)
+
 __all__ = [
     # clients
     "CallAutomationClient",
     "CallConnectionClient",
 
+    # parser
+    "StreamingDataParser",
+
     # models for input
     "FileSource",
     "TextSource",
@@ -87,6 +96,10 @@
     "CommunicationCloudEnvironment",
     "UnknownIdentifier",
 
+    # streaming models
+    "TranscriptionMetadata",
+    "TranscriptionData",
+
     # enums
     "CallRejectReason",
     "RecordingContent",

sdk/communication/azure-communication-callautomation/azure/communication/callautomation/streaming/__init__.py

@@ -0,0 +1,5 @@
+# -------------------------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License. See License.txt in the project root for
+# license information.
+# --------------------------------------------------------------------------

sdk/communication/azure-communication-callautomation/azure/communication/callautomation/streaming/models.py

@@ -0,0 +1,119 @@
+# -------------------------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License. See License.txt in the project root for
+# license information.
+# --------------------------------------------------------------------------
+
+from enum import Enum
+from typing import List
+from azure.communication.callautomation._shared.models import CommunicationIdentifier
+
+class ResultStatus(Enum):
+    """
+    The status of the result of transcription.
+    """
+    INTERMEDIATE = "intermediate"
+    FINAL = "final"
+
+class TextFormat(Enum):
+    """
+    The format of transcription text.
+    """
+    DISPLAY = "display"
+
+class WordData:
+    """
+    Text in the phrase.
+
+    :keyword text: Text in the phrase.
+    :paramtype text: str
+    :keyword offset: The word's position within the phrase.
+    :paramtype offset: int
+    :keyword duration: Duration in ticks. 1 tick = 100 nanoseconds.
+    :paramtype duration: int
+    """
+    text:str
+    """ Text in the phrase. """
+    offset:int
+    """ The word's position within the phrase. """
+    duration:int
+    """ Duration in ticks. 1 tick = 100 nanoseconds. """
+    def __init__(self, text: str, offset: int, duration: int):
+        self.text = text
+        self.offset = offset
+        self.duration = duration
+
+class TranscriptionMetadata:
+    """
+    Metadata for Transcription Streaming.
+
+    :keyword subscriptionId: Transcription Subscription Id.
+    :paramtype subscriptionId: str
+    :keyword locale: The target locale in which the translated text needs to be.
+    :paramtype locale: str
+    :keyword callConnectionId: call connection Id.
+    :paramtype callConnectionId: str
+    :keyword correlationId: correlation Id.
+    :paramtype correlationId: str
+    """
+
+    subscriptionId: str
+    """ Transcription Subscription Id. """
+    locale: str
+    """ The target locale in which the translated text needs to be. """
+    callConnectionId: str
+    """ call connection Id. """
+    correlationId: str
+    """ correlation Id. """
+    def __init__(self, subscriptionId: str, locale: str, callConnectionId: str, correlationId: str):
+        self.subscriptionId = subscriptionId
+        self.locale = locale
+        self.callConnectionId = callConnectionId
+        self.correlationId = correlationId
+
+class TranscriptionData:
+    """
+    Streaming Transcription.
+    :keyword  text: The display form of the recognized word.
+    :paramtype  text: str
+    :keyword format: The format of text.
+    :paramtype format: TextFormat
+    :keyword confidence: Confidence of recognition of the whole phrase.
+    :paramtype confidence: float
+    :keyword offset: The position of this payload.
+    :paramtype offset: int
+    :keyword duration: Duration in ticks. 1 tick = 100 nanoseconds.
+    :paramtype duration: int
+    :keyword words: The result for each word of the phrase.
+    :paramtype words: List[WordData]
+    :keyword participant: The identified speaker based on participant raw ID.
+    :paramtype participant: CommunicationIdentifier
+    :keyword resultStatus: Status of the result of transcription.
+    :paramtype resultStatus: ResultStatus
+    """
+    text: str
+    """ The display form of the recognized word. """
+    format: TextFormat
+    """ The format of text. """
+    confidence: float
+    """ Confidence of recognition of the whole phrase, from 0.0 (no confidence) to 1.0 (full confidence). """
+    offset: int
+    """ The position of this payload. """
+    duration: int
+    """ Duration in ticks. 1 tick = 100 nanoseconds. """
+    words: List[WordData]
+    """ The result for each word of the phrase. """
+    participant: CommunicationIdentifier
+    """ The identified speaker based on participant raw ID. """
+    resultStatus: ResultStatus
+    """ Status of the result of transcription. """
+    def __init__(self, text: str, format: TextFormat, confidence: float, offset: int, duration: int,
+    words: List[WordData], participant: CommunicationIdentifier, resultStatus: ResultStatus):
+        self.text = text
+        self.format = format
+        self.confidence = confidence
+        self.offset = offset
+        self.duration = duration
+        self.words = words
+        self.participant = participant
+        self.resultStatus = resultStatus

sdk/communication/azure-communication-callautomation/azure/communication/callautomation/streaming/streaming_data_parser.py

@@ -0,0 +1,51 @@
+# -------------------------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License. See License.txt in the project root for
+# license information.
+# --------------------------------------------------------------------------
+
+from typing import Union
+import json
+from azure.communication.callautomation._shared.models import identifier_from_raw_id
+from azure.communication.callautomation.streaming.models import (TranscriptionMetadata,TranscriptionData,WordData)
+
+class StreamingDataParser:
+    @staticmethod
+    def parse(packet_data: Union[str, bytes]) -> Union[TranscriptionMetadata, TranscriptionData]:
+        """
+        Parse the incoming packets.
+        :keyword packet_data: Transcription packet data.
+        :paramtype packet_data: Union[str, bytes]
+        :return: Union[TranscriptionMetadata, TranscriptionData]
+        :rType: TranscriptionMetadata, TranscriptionData
+        :raises: ValueError
+        """
+        if isinstance(packet_data, str):
+            string_json = packet_data
+        elif isinstance(packet_data,bytes):
+            string_json = packet_data.decode('utf-8')
+        else:
+            ValueError(packet_data)
+
+        json_object = json.loads(string_json)
+        kind = json_object['kind']
+
+        if kind == 'TranscriptionMetadata':
+            transcription_metadata = TranscriptionMetadata(**json_object['transcriptionMetadata'])
+            return transcription_metadata
+        if kind == 'TranscriptionData':
+            participant = identifier_from_raw_id(json_object['transcriptionData']['participantRawID'])
+            word_data_list = json_object['transcriptionData']['words']
+            words = [WordData(entry["text"], entry["offset"], entry["duration"]) for entry in word_data_list]
+            transcription_data = TranscriptionData(
+                text=json_object['transcriptionData']['text'],
+                format=json_object['transcriptionData']['format'],
+                confidence=json_object['transcriptionData']['confidence'],
+                offset=json_object['transcriptionData']['offset'],
+                duration=json_object['transcriptionData']['duration'],
+                words=words,
+                participant=participant,
+                resultStatus=json_object['transcriptionData']['resultStatus']
+            )
+            return transcription_data
+        raise ValueError(string_json)

sdk/communication/azure-communication-callautomation/tests/test_streaming_data_parser.py

@@ -0,0 +1,64 @@
+# -------------------------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# Licensed under the MIT License. See License.txt in the project root for
+# license information.
+# --------------------------------------------------------------------------
+
+import json
+import unittest
+from azure.communication.callautomation.streaming.models import (TranscriptionMetadata,TranscriptionData,WordData,TextFormat,ResultStatus)
+from azure.communication.callautomation.streaming.streaming_data_parser import StreamingDataParser
+
+class TestStreamDataParser(unittest.TestCase):
+    def setUp(self):
+        self.transcriptionMetaDataJson = '{"kind":"TranscriptionMetadata","transcriptionMetadata":{"subscriptionId":"0000a000-9999-5555-ae00-cd00e0bc0000","locale":"en-US","callConnectionId":"6d09449c-6677-4f91-8cb7-012c338e6ec1","correlationId":"6d09449c-6677-4f91-8cb7-012c338e6ec1"}}'
+        self.transcriptionDataJson = '{"kind":"TranscriptionData","transcriptionData":{"text":"Is everything fine.","format":"display","confidence":0.8138430714607239,"offset":868464674,"duration":11600000,"words":[{"text":"is","offset":868464674,"duration":2400000},{"text":"everything","offset":870864674,"duration":5200000},{"text":"fine","offset":876064674,"duration":4000000}],"participantRawID":"4:+910000000000","resultStatus":"Final"}}'
+
+    def test_parse_binary_to_transcription_metadata(self):
+        transcriptionMetaDataBinary = self.transcriptionMetaDataJson.encode('utf-8')
+        parsedData = StreamingDataParser.parse(transcriptionMetaDataBinary)
+        self.assertTrue(isinstance(parsedData, TranscriptionMetadata))
+        self.validate_transcription_metadata(parsedData)
+
+    def test_parse_json_to_transcription_metadata(self):
+        parsedData = StreamingDataParser.parse(self.transcriptionMetaDataJson)
+        self.assertTrue(isinstance(parsedData, TranscriptionMetadata))
+        self.validate_transcription_metadata(parsedData)
+
+    def test_parse_binary_to_transcription_data(self):
+        transcriptionDataBinary = self.transcriptionDataJson.encode('utf-8')
+        parsedData = StreamingDataParser.parse(transcriptionDataBinary)
+        self.assertTrue(isinstance(parsedData, TranscriptionData))
+        self.validate_transcription_data(parsedData)
+
+    def test_parse_json_to_transcription_data(self):
+        parsedData = StreamingDataParser.parse(self.transcriptionDataJson)
+        self.assertTrue(isinstance(parsedData, TranscriptionData))
+        self.validate_transcription_data(parsedData)
+
+    def validate_transcription_metadata(self, transcriptionMetadata):
+        self.assertEqual(transcriptionMetadata.subscriptionId, "0000a000-9999-5555-ae00-cd00e0bc0000")
+        self.assertEqual(transcriptionMetadata.locale, "en-US")
+        self.assertEqual(transcriptionMetadata.correlationId, "6d09449c-6677-4f91-8cb7-012c338e6ec1")
+        self.assertEqual(transcriptionMetadata.callConnectionId, "6d09449c-6677-4f91-8cb7-012c338e6ec1")
+
+    def validate_transcription_data(self, transcriptionData):
+        self.assertEqual(transcriptionData.text, "Is everything fine.")
+        self.assertEqual(transcriptionData.format, "display")
+        self.assertEqual(transcriptionData.resultStatus, "Final")
+        self.assertAlmostEqual(transcriptionData.confidence, 0.8138430714607239)
+        self.assertEqual(transcriptionData.offset, 868464674)
+        self.assertEqual(transcriptionData.duration, 11600000)
+        self.assertEqual(len(transcriptionData.words), 3)
+        self.assertEqual(transcriptionData.words[0].text, "is")
+        self.assertEqual(transcriptionData.words[0].offset, 868464674)
+        self.assertEqual(transcriptionData.words[0].duration, 2400000)
+        self.assertEqual(transcriptionData.words[1].text, "everything")
+        self.assertEqual(transcriptionData.words[1].offset, 870864674)
+        self.assertEqual(transcriptionData.words[1].duration, 5200000)
+        self.assertEqual(transcriptionData.words[2].text, "fine")
+        self.assertEqual(transcriptionData.words[2].offset, 876064674)
+        self.assertEqual(transcriptionData.words[2].duration, 4000000)
+        self.assertEqual(transcriptionData.participant.raw_id, "4:+910000000000")
+
+