fix: restore previous behavior on live pipeline + audio downscaler (#561)

This commit restore the original behavior with frame cutting. While
silero is used on our gpu for files, look like it's not working great on
the live pipeline. To be investigated, but at the moment, what we keep
is:

- refactored to extract the downscale for further processing in the
pipeline
- remove any downscale implementation from audio_chunker and audio_merge
- removed batching from audio_merge too for now

server/reflector/tools/process.py

@@ -16,7 +16,8 @@ import av
 
 from reflector.logger import logger
 from reflector.processors import (
-    AudioChunkerProcessor,
+    AudioChunkerAutoProcessor,
+    AudioDownscaleProcessor,
     AudioFileWriterProcessor,
     AudioMergeProcessor,
     AudioTranscriptAutoProcessor,
@@ -95,7 +96,8 @@ async def process_audio_file(
 
     # Add the rest of the processors
     processors += [
-        AudioChunkerProcessor(),
+        AudioDownscaleProcessor(),
+        AudioChunkerAutoProcessor(),
         AudioMergeProcessor(),
         AudioTranscriptAutoProcessor.as_threaded(),
         TranscriptLinerProcessor(),
@@ -322,7 +324,8 @@ if __name__ == "__main__":
 
         # Ignore internal processors
         if processor in (
-            "AudioChunkerProcessor",
+            "AudioDownscaleProcessor",
+            "AudioChunkerAutoProcessor",
             "AudioMergeProcessor",
             "AudioFileWriterProcessor",
             "TopicCollectorProcessor",

server/reflector/tools/process_with_diarization.py

@@ -17,7 +17,8 @@ import av
 
 from reflector.logger import logger
 from reflector.processors import (
-    AudioChunkerProcessor,
+    AudioChunkerAutoProcessor,
+    AudioDownscaleProcessor,
     AudioFileWriterProcessor,
     AudioMergeProcessor,
     AudioTranscriptAutoProcessor,
@@ -96,7 +97,8 @@ async def process_audio_file_with_diarization(
 
     # Add the rest of the processors
     processors += [
-        AudioChunkerProcessor(),
+        AudioDownscaleProcessor(),
+        AudioChunkerAutoProcessor(),
         AudioMergeProcessor(),
         AudioTranscriptAutoProcessor.as_threaded(),
     ]
@@ -276,7 +278,8 @@ if __name__ == "__main__":
 
         # Ignore internal processors
         if processor in (
-            "AudioChunkerProcessor",
+            "AudioDownscaleProcessor",
+            "AudioChunkerAutoProcessor",
             "AudioMergeProcessor",
             "AudioFileWriterProcessor",
             "TopicCollectorProcessor",

server/reflector/tools/runpipeline.py

@@ -53,7 +53,7 @@ async def run_single_processor(args):
     async def event_callback(event: PipelineEvent):
         processor = event.processor
         # ignore some processor
-        if processor in ("AudioChunkerProcessor", "AudioMergeProcessor"):
+        if processor in ("AudioChunkerAutoProcessor", "AudioMergeProcessor"):
             return
         print(f"Event: {event}")
         if output_fd: