server: fixes latest implementation details on rtc offer and fastapi

server/reflector/processors/__init__.py

@@ -1,9 +1,9 @@
 from .base import Processor, ThreadedProcessor, Pipeline  # noqa: F401
-from .types import AudioFile, Transcript, Word, TitleSummary  # noqa: F401
+from .types import AudioFile, Transcript, Word, TitleSummary, FinalSummary  # noqa: F401
 from .audio_chunker import AudioChunkerProcessor  # noqa: F401
 from .audio_merge import AudioMergeProcessor  # noqa: F401
 from .audio_transcript import AudioTranscriptProcessor  # noqa: F401
 from .audio_transcript_auto import AudioTranscriptAutoProcessor  # noqa: F401
 from .transcript_liner import TranscriptLinerProcessor  # noqa: F401
-from .transcript_summarizer import TranscriptSummarizerProcessor  # noqa: F401
 from .transcript_topic_detector import TranscriptTopicDetectorProcessor  # noqa: F401
+from .transcript_final_summary import TranscriptFinalSummaryProcessor  # noqa: F401

server/reflector/processors/transcript_final_summary.py

@@ -0,0 +1,30 @@
+from reflector.processors.base import Processor
+from reflector.processors.types import TitleSummary, FinalSummary
+
+
+class TranscriptFinalSummaryProcessor(Processor):
+    """
+    Assemble all summary into a line-based json
+    """
+
+    INPUT_TYPE = TitleSummary
+    OUTPUT_TYPE = FinalSummary
+
+    def __init__(self, **kwargs):
+        super().__init__(**kwargs)
+        self.chunks: list[TitleSummary] = []
+
+    async def _push(self, data: TitleSummary):
+        self.chunks.append(data)
+
+    async def _flush(self):
+        if not self.chunks:
+            self.logger.warning("No summary to output")
+            return
+
+        # FIXME improve final summary
+        result = "\n".join([chunk.summary for chunk in self.chunks])
+        last_chunk = self.chunks[-1]
+        duration = last_chunk.timestamp + last_chunk.duration
+
+        await self.emit(FinalSummary(summary=result, duration=duration))

server/reflector/processors/transcript_summarizer.py

@@ -1,30 +0,0 @@
-from reflector.processors.base import Processor
-from reflector.processors.types import TitleSummary
-from pathlib import Path
-import json
-
-
-class TranscriptSummarizerProcessor(Processor):
-    """
-    Assemble all summary into a line-based json
-    """
-
-    INPUT_TYPE = TitleSummary
-    OUTPUT_TYPE = Path
-
-    def __init__(self, filename: Path, **kwargs):
-        super().__init__(**kwargs)
-        self.filename = filename
-        self.chunkcount = 0
-
-    async def _push(self, data: TitleSummary):
-        with open(self.filename, "a", encoding="utf8") as fd:
-            fd.write(json.dumps(data))
-        self.chunkcount += 1
-
-    async def _flush(self):
-        if self.chunkcount == 0:
-            self.logger.warning("No summary to write")
-            return
-        self.logger.info(f"Writing to {self.filename}")
-        await self.emit(self.filename)

server/reflector/processors/types.py

@@ -66,3 +66,9 @@ class TitleSummary:
     timestamp: float
     duration: float
     transcript: Transcript
+
+
+@dataclass
+class FinalSummary:
+    summary: str
+    duration: float