Merge branch 'houssem_speaker_identify_optim' of https://github.com/l…

…into-ai/linto-diarization into houssem_speaker_identify_optim

http_server/ingress.py

@@ -3,7 +3,6 @@
 import json
 import logging
 from time import time
-import os
 
 from confparser import createParser
 from flask import Flask, Response, abort, json, request
@@ -12,7 +11,6 @@
 
 from diarization.processing import diarizationworker, USE_GPU
 
-
 app = Flask("__diarization-serving__")
 
 logging.basicConfig(

pyannote/README.md

@@ -232,7 +232,7 @@ Diarization worker accepts requests with the following arguments:
 * `file`: (str) Is the relative path of the file in the shared_folder.
 * `speaker_count`: (int, default None) Fixed number of speakers.
 * `max_speaker`: (int, default None) Max number of speaker if speaker_count=None. 
-* `speaker_names`: (string, optional) List of target speaker names, speaker identification (if speaker samples are provided only). Possible values are
+* `speaker_names`: (string, default None) List of target speaker names, speaker identification (if speaker samples are provided only). Possible values are
   * empty string "": no speaker identification
   * wild card "`*`": speaker identification for all speakers
   * list of speaker names in json format (ex: "`["speaker1", ..., "speakerN"]`") or separated by `|` (ex: "`speaker1|...|speakerN`"): speaker identification for the listed speakers only

pyannote/RELEASE.md

@@ -1,5 +1,6 @@
 # 2.0.1
 - Use Qdrant for efficient speaker identification
+- Update pyannote to 3.3.2 (and speechbrain 1.0.0)
 
 # 2.0.0
 - Add speaker identification

simple/README.md

@@ -244,7 +244,7 @@ Diarization worker accepts requests with the following arguments:
 * `file`: (str) Is the relative path of the file in the shared_folder.
 * `speaker_count`: (int, default None) Fixed number of speakers.
 * `max_speaker`: (int, default None) Max number of speaker if speaker_count=None. 
-* `speaker_names`: (string, optional) List of target speaker names, speaker identification (if speaker samples are provided only). Possible values are
+* `speaker_names`: (string, default None) List of target speaker names, speaker identification (if speaker samples are provided only). Possible values are
   * empty string "": no speaker identification
   * wild card "`*`": speaker identification for all speakers
   * list of speaker names in json format (ex: "`["speaker1", ..., "speakerN"]`") or separated by `|` (ex: "`speaker1|...|speakerN`"): speaker identification for the listed speakers only

simple/RELEASE.md

@@ -1,5 +1,6 @@
 # 2.0.1
 - Use Qdrant for efficient speaker identification
+- Specifying max number of speakers is now optional
 
 # 2.0.0
 - Add speaker identification

simple/diarization/processing/speakerdiarization.py

@@ -8,6 +8,7 @@
 import memory_tempfile
 import torch
 import werkzeug
+import warnings
 
 sys.path.append(os.path.join(os.path.dirname(__file__), "simple_diarizer"))
 import simple_diarizer
@@ -210,7 +211,8 @@ def run(
         self.log.info(f"Starting diarization on file {file_path}")
 
         if speaker_count is None and max_speaker is None:
-            raise Exception("Either speaker_count or max_speaker must be set")
+            max_speaker = 50 # default value
+            warnings.warn(f"No speaker count nor maximum specified, using default value {max_speaker=}")
 
         try:                       
             result = self.run_simple_diarizer(