pvleopard 1.2.2__py3-none-any.whl → 2.0.1__py3-none-any.whl

pvleopard/_factory.py

@@ -19,7 +19,8 @@ def create(
         access_key: str,
         model_path: Optional[str] = None,
         library_path: Optional[str] = None,
-        enable_automatic_punctuation: bool = False) -> Leopard:
+        enable_automatic_punctuation: bool = False,
+        enable_diarization: bool = False) -> Leopard:
     """
     Factory method for Leopard speech-to-text engine.
 
@@ -28,6 +29,9 @@ def create(
     :param model_path: Absolute path to the file containing model parameters. If not set it will be set to the default
     location.
     :param enable_automatic_punctuation Set to `True` to enable automatic punctuation insertion.
+    :param enable_diarization Set to `true` to enable speaker diarization, which allows Leopard to differentiate
+    speakers as part of the transcription process. Word metadata will include a `speaker_tag` to
+    identify unique speakers.
     :return: An instance of Leopard speech-to-text engine.
     """
 
@@ -41,7 +45,8 @@ def create(
         access_key=access_key,
         model_path=model_path,
         library_path=library_path,
-        enable_automatic_punctuation=enable_automatic_punctuation)
+        enable_automatic_punctuation=enable_automatic_punctuation,
+        enable_diarization=enable_diarization)
 
 
 __all__ = [

pvleopard/_leopard.py

@@ -18,7 +18,27 @@ from typing import *
 
 
 class LeopardError(Exception):
-    pass
+    def __init__(self, message: str = '', message_stack: Sequence[str] = None):
+        super().__init__(message)
+
+        self._message = message
+        self._message_stack = list() if message_stack is None else message_stack
+
+    def __str__(self):
+        message = self._message
+        if len(self._message_stack) > 0:
+            message += ':'
+            for i in range(len(self._message_stack)):
+                message += '\n  [%d] %s' % (i, self._message_stack[i])
+        return message
+
+    @property
+    def message(self) -> str:
+        return self._message
+
+    @property
+    def message_stack(self) -> Sequence[str]:
+        return self._message_stack
 
 
 class LeopardMemoryError(LeopardError):
@@ -119,14 +139,16 @@ class Leopard(object):
             ("word", c_char_p),
             ("start_sec", c_float),
             ("end_sec", c_float),
-            ("confidence", c_float)]
+            ("confidence", c_float),
+            ("speaker_tag", c_int32)]
 
     def __init__(
             self,
             access_key: str,
             model_path: str,
             library_path: str,
-            enable_automatic_punctuation: bool = False) -> None:
+            enable_automatic_punctuation: bool = False,
+            enable_diarization: bool = False) -> None:
         """
         Constructor.
 
@@ -134,6 +156,9 @@ class Leopard(object):
         :param model_path: Absolute path to the file containing model parameters.
         :param library_path: Absolute path to Leopard's dynamic library.
         :param enable_automatic_punctuation Set to `True` to enable automatic punctuation insertion.
+        :param enable_diarization Set to `true` to enable speaker diarization, which allows Leopard to differentiate
+        speakers as part of the transcription process. Word metadata will include a `speaker_tag` to
+        identify unique speakers.
         """
 
         if not isinstance(access_key, str) or len(access_key) == 0:
@@ -147,15 +172,36 @@ class Leopard(object):
 
         library = cdll.LoadLibrary(library_path)
 
+        set_sdk_func = library.pv_set_sdk
+        set_sdk_func.argtypes = [c_char_p]
+        set_sdk_func.restype = None
+
+        set_sdk_func('python'.encode('utf-8'))
+
+        self._get_error_stack_func = library.pv_get_error_stack
+        self._get_error_stack_func.argtypes = [POINTER(POINTER(c_char_p)), POINTER(c_int)]
+        self._get_error_stack_func.restype = self.PicovoiceStatuses
+
+        self._free_error_stack_func = library.pv_free_error_stack
+        self._free_error_stack_func.argtypes = [POINTER(c_char_p)]
+        self._free_error_stack_func.restype = None
+
         init_func = library.pv_leopard_init
-        init_func.argtypes = [c_char_p, c_char_p, c_bool, POINTER(POINTER(self.CLeopard))]
+        init_func.argtypes = [c_char_p, c_char_p, c_bool, c_bool, POINTER(POINTER(self.CLeopard))]
         init_func.restype = self.PicovoiceStatuses
 
         self._handle = POINTER(self.CLeopard)()
 
-        status = init_func(access_key.encode(), model_path.encode(), enable_automatic_punctuation, byref(self._handle))
+        status = init_func(
+            access_key.encode(),
+            model_path.encode(),
+            enable_automatic_punctuation,
+            enable_diarization,
+            byref(self._handle))
         if status is not self.PicovoiceStatuses.SUCCESS:
-            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status]()
+            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status](
+                message='Initialization failed',
+                message_stack=self._get_error_stack())
 
         self._delete_func = library.pv_leopard_delete
         self._delete_func.argtypes = [POINTER(self.CLeopard)]
@@ -201,7 +247,7 @@ class Leopard(object):
         ]
         self._words_delete_func.restype = None
 
-    Word = namedtuple('Word', ['word', 'start_sec', 'end_sec', 'confidence'])
+    Word = namedtuple('Word', ['word', 'start_sec', 'end_sec', 'confidence', 'speaker_tag'])
 
     def process(self, pcm: Sequence[int]) -> Tuple[str, Sequence[Word]]:
         """
@@ -227,7 +273,9 @@ class Leopard(object):
             byref(num_words),
             byref(c_words))
         if status is not self.PicovoiceStatuses.SUCCESS:
-            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status]()
+            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status](
+                message='Process failed',
+                message_stack=self._get_error_stack())
 
         transcript = c_transcript.value.decode('utf-8')
         self._transcript_delete_func(c_transcript)
@@ -238,7 +286,8 @@ class Leopard(object):
                 word=c_words[i].word.decode('utf-8'),
                 start_sec=c_words[i].start_sec,
                 end_sec=c_words[i].end_sec,
-                confidence=c_words[i].confidence)
+                confidence=c_words[i].confidence,
+                speaker_tag=c_words[i].speaker_tag)
             words.append(word)
 
         self._words_delete_func(c_words)
@@ -267,12 +316,9 @@ class Leopard(object):
             byref(num_words),
             byref(c_words))
         if status is not self.PicovoiceStatuses.SUCCESS:
-            if status is self.PicovoiceStatuses.INVALID_ARGUMENT:
-                if not audio_path.lower().endswith(self._VALID_EXTENSIONS):
-                    raise self._PICOVOICE_STATUS_TO_EXCEPTION[status](
-                        "Specified file with extension '%s' is not supported" % pathlib.Path(audio_path).suffix
-                    )
-            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status]()
+            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status](
+                message='Process file failed',
+                message_stack=self._get_error_stack())
 
         transcript = c_transcript.value.decode('utf-8')
         self._transcript_delete_func(c_transcript)
@@ -283,7 +329,8 @@ class Leopard(object):
                 word=c_words[i].word.decode('utf-8'),
                 start_sec=c_words[i].start_sec,
                 end_sec=c_words[i].end_sec,
-                confidence=c_words[i].confidence)
+                confidence=c_words[i].confidence,
+                speaker_tag=c_words[i].speaker_tag)
             words.append(word)
 
         self._words_delete_func(c_words)
@@ -307,6 +354,21 @@ class Leopard(object):
 
         return self._sample_rate
 
+    def _get_error_stack(self) -> Sequence[str]:
+        message_stack_ref = POINTER(c_char_p)()
+        message_stack_depth = c_int()
+        status = self._get_error_stack_func(byref(message_stack_ref), byref(message_stack_depth))
+        if status is not self.PicovoiceStatuses.SUCCESS:
+            raise self._PICOVOICE_STATUS_TO_EXCEPTION[status](message='Unable to get Leopard error state')
+
+        message_stack = list()
+        for i in range(message_stack_depth.value):
+            message_stack.append(message_stack_ref[i].decode('utf-8'))
+
+        self._free_error_stack_func(message_stack_ref)
+
+        return message_stack
+
 
 __all__ = [
     'Leopard',

{pvleopard-1.2.2.dist-info → pvleopard-2.0.1.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: pvleopard
-Version: 1.2.2
+Version: 2.0.1
 Summary: Leopard Speech-to-Text Engine.
 Home-page: https://github.com/Picovoice/leopard
 Author: Picovoice
@@ -17,13 +17,15 @@ Classifier: Topic :: Multimedia :: Sound/Audio :: Speech
 Requires-Python: >=3.5
 Description-Content-Type: text/markdown
 
-# Leopard Speech-to-Text Engine
+# Leopard Binding for Python
+
+## Leopard Speech-to-Text Engine
 
 Made in Vancouver, Canada by [Picovoice](https://picovoice.ai)
 
 Leopard is an on-device speech-to-text engine. Leopard is:
 
-- Private; All voice processing runs locally. 
+- Private; All voice processing runs locally.
 - [Accurate](https://picovoice.ai/docs/benchmark/stt/)
 - [Compact and Computationally-Efficient](https://github.com/Picovoice/speech-to-text-benchmark#rtf)
 - Cross-Platform:
@@ -56,9 +58,9 @@ Create an instance of the engine and transcribe an audio file:
 ```python
 import pvleopard
 
-handle = pvleopard.create(access_key='${ACCESS_KEY}')
+leopard = pvleopard.create(access_key='${ACCESS_KEY}')
 
-transcript, words = handle.process_file('${AUDIO_PATH}')
+transcript, words = leopard.process_file('${AUDIO_PATH}')
 print(transcript)
 for word in words:
     print(
@@ -67,20 +69,24 @@ for word in words:
 ```
 
 Replace `${ACCESS_KEY}` with yours obtained from [Picovoice Console](https://console.picovoice.ai/) and
-`${AUDIO_PATH}` to the path an audio file. Finally, when done be sure to explicitly release the resources using
-`handle.delete()`.
+`${AUDIO_PATH}` to the path an audio file.
+
+Finally, when done be sure to explicitly release the resources:
+```python
+leopard.delete()
+```
 
 ## Language Model
 
-The Leopard Python SDK comes preloaded with a default English language model (`.pv` file). 
-Default models for other supported languages can be found in [lib/common](../../lib/common). 
+The Leopard Python SDK comes preloaded with a default English language model (`.pv` file).
+Default models for other supported languages can be found in [lib/common](../../lib/common).
 
 Create custom language models using the [Picovoice Console](https://console.picovoice.ai/). Here you can train
 language models with custom vocabulary and boost words in the existing vocabulary.
 
 Pass in the `.pv` file via the `model_path` argument:
 ```python
-handle = pvleopard.create(
+leopard = pvleopard.create(
     access_key='${ACCESS_KEY}',
     model_path='${MODEL_PATH}')
 ```

pvleopard-2.0.1.dist-info/RECORD

@@ -0,0 +1,19 @@
+pvleopard/LICENSE,sha256=ZurJwSSRHw99lGaJP88vQREqtZmIABuVKd_rK7k7U70,11344
+pvleopard/__init__.py,sha256=OCI_7jgscvkzJMwT02SfB85_xOoQC8eb1e52HR_GKgo,577
+pvleopard/_factory.py,sha256=1ZotcV9xdcJLRAYajFwFHfB0tV3_9bSi15_cjhedMnM,2075
+pvleopard/_leopard.py,sha256=U6eNGzgQfpKaQxnDaOKHBDi2Kj_ysEHhF3SdVojbmhE,12741
+pvleopard/_util.py,sha256=TQ3fXjUNUDerUbXdC6ntjMXaYD_B5udi__xp41zL9lE,3113
+pvleopard/lib/common/leopard_params.pv,sha256=8jsCUfmOuuTD2MXLduV17hD1vHXuO-wscyJobYCsbfs,37699432
+pvleopard/lib/jetson/cortex-a57-aarch64/libpv_leopard.so,sha256=y431wzz9NuRsOXnhk7ciZ9s6Lc3v9lr5mWdbkqC1bXE,1300032
+pvleopard/lib/linux/x86_64/libpv_leopard.so,sha256=UhcGHHjCnR1Y-AMHVgVIbRoR_yjgRlw7SpcV86JwkX8,1415360
+pvleopard/lib/mac/arm64/libpv_leopard.dylib,sha256=o4WF9nzDvF5K9_AKppz-avXsk-wApmRVKQuUArbvxSc,1545072
+pvleopard/lib/mac/x86_64/libpv_leopard.dylib,sha256=Qwnh-0EI-VSWTjbSXnfmkxmYU8Z1ow-fFkCUQ2K9slc,1697720
+pvleopard/lib/raspberry-pi/cortex-a53/libpv_leopard.so,sha256=DEnIu2Jpe_w9sviYNCLvgC1yqROLdt6A8b4r44d6eiQ,1275760
+pvleopard/lib/raspberry-pi/cortex-a53-aarch64/libpv_leopard.so,sha256=FBqYYJbmtOu4AAD532__IfuJo6O9MnkP933CYK4pF3g,1312320
+pvleopard/lib/raspberry-pi/cortex-a72/libpv_leopard.so,sha256=ALyBJ2I5B_nKQ2Vr7cTrPiFSnxIxsTPhOhcIWp3SOOg,1288048
+pvleopard/lib/raspberry-pi/cortex-a72-aarch64/libpv_leopard.so,sha256=Kn5Ricawg62MGrzL6ZS-bcPOarlSU0PcwSQ-WG7k-Uo,1300032
+pvleopard/lib/windows/amd64/libpv_leopard.dll,sha256=hM6eYFgcQ6p8ZIe8BFw0Yxm3enA2rPSRBCysVzFC8kQ,1524736
+pvleopard-2.0.1.dist-info/METADATA,sha256=hloxBn0p8TABCghj9eSzSBIZxNNK-q1G_MyHR5yoOaY,3141
+pvleopard-2.0.1.dist-info/WHEEL,sha256=g4nMs7d-Xl9-xC9XovUrsDHGXt-FT0E17Yqo92DEfvY,92
+pvleopard-2.0.1.dist-info/top_level.txt,sha256=DAhlor-zWSROmsQCFWDsx_IJSE62zlgJ3sE4quxhEPw,10
+pvleopard-2.0.1.dist-info/RECORD,,

{pvleopard-1.2.2.dist-info → pvleopard-2.0.1.dist-info}/WHEEL

@@ -1,5 +1,5 @@
 Wheel-Version: 1.0
-Generator: bdist_wheel (0.37.1)
+Generator: bdist_wheel (0.34.2)
 Root-Is-Purelib: true
 Tag: py3-none-any
 

pvleopard-1.2.2.dist-info/RECORD

@@ -1,19 +0,0 @@
-pvleopard/LICENSE,sha256=ZurJwSSRHw99lGaJP88vQREqtZmIABuVKd_rK7k7U70,11344
-pvleopard/__init__.py,sha256=OCI_7jgscvkzJMwT02SfB85_xOoQC8eb1e52HR_GKgo,577
-pvleopard/_factory.py,sha256=WzrCwUlU_TCXkzZKPFd0ag4sAdiuwRlpOM6HYl1M-t0,1746
-pvleopard/_leopard.py,sha256=qk9kf7aRETnbEyBOrZ7sg_ESgKm3NEow_6nnQjjl-ms,10366
-pvleopard/_util.py,sha256=TQ3fXjUNUDerUbXdC6ntjMXaYD_B5udi__xp41zL9lE,3113
-pvleopard/lib/common/leopard_params.pv,sha256=CBoaZ8kach9nVD-I-SJakBvbVi2EfxpiumH9Bk_DKCU,19850729
-pvleopard/lib/jetson/cortex-a57-aarch64/libpv_leopard.so,sha256=W9ikHNNpZVIM4lkHABfGJa2WVUK-U5XwZrUz0mUr0es,1290872
-pvleopard/lib/linux/x86_64/libpv_leopard.so,sha256=peVOWhgqeC09UrYSKnduzvkKkm7WZ0-jWbiiTs770MM,1397912
-pvleopard/lib/mac/arm64/libpv_leopard.dylib,sha256=KYM2pXObVPfsitBIl4gvap2IyrDhMpCpNI6k23F_YlE,1513536
-pvleopard/lib/mac/x86_64/libpv_leopard.dylib,sha256=0ImW2M-JTEeAEZ9bfsLV9B0-FClfY1y6EnLR2tFCvs8,1666456
-pvleopard/lib/raspberry-pi/cortex-a53/libpv_leopard.so,sha256=YtVPTq2oVm9DLcK7ruo5ZJkqqe33-OGM1xq5l-v4P0E,1263028
-pvleopard/lib/raspberry-pi/cortex-a53-aarch64/libpv_leopard.so,sha256=w6kYHEuaAZ2tm-aHA_-xkmv-FgrmBFOhYtIDxPXb23c,1303240
-pvleopard/lib/raspberry-pi/cortex-a72/libpv_leopard.so,sha256=LCdTw1mJYY1CjNzXimay8P3kd9LMvUBZap38e8j9Wvw,1275316
-pvleopard/lib/raspberry-pi/cortex-a72-aarch64/libpv_leopard.so,sha256=rtpK3dyk8Kw6Ls4NxwkR4QQiwVo5mbWFVKHgkiITTOo,1295048
-pvleopard/lib/windows/amd64/libpv_leopard.dll,sha256=P4A37VaWU-bMVwoqvuxtKAZfg_Y2Li33mMwf5QjQ3eo,1505792
-pvleopard-1.2.2.dist-info/METADATA,sha256=0gk8fFJY2AbMv7pT-dWdD5YoXlZX-zT-4OhI4jhf2j8,3102
-pvleopard-1.2.2.dist-info/WHEEL,sha256=G16H4A3IeoQmnOrYV4ueZGKSjhipXx8zc8nu9FGlvMA,92
-pvleopard-1.2.2.dist-info/top_level.txt,sha256=DAhlor-zWSROmsQCFWDsx_IJSE62zlgJ3sE4quxhEPw,10
-pvleopard-1.2.2.dist-info/RECORD,,