From 9642d0fd1e201fc804d61a35cbeb0c5a5217e4ab Mon Sep 17 00:00:00 2001
From: Mathieu Virbel <mat@meltingrocks.com>
Date: Thu, 2 Nov 2023 19:40:45 +0100
Subject: [PATCH] hotfix/server: fix duplication of topics

---
 server/reflector/db/transcripts.py               | 6 +++---
 server/reflector/pipelines/main_live_pipeline.py | 9 +++++++--
 server/reflector/processors/types.py             | 6 +++++-
 3 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/server/reflector/db/transcripts.py b/server/reflector/db/transcripts.py
index 89025d53..5d190bdc 100644
--- a/server/reflector/db/transcripts.py
+++ b/server/reflector/db/transcripts.py
@@ -106,9 +106,9 @@ class Transcript(BaseModel):
         return ev
 
     def upsert_topic(self, topic: TranscriptTopic):
-        existing_topic = next((t for t in self.topics if t.id == topic.id), None)
-        if existing_topic:
-            existing_topic.update_from(topic)
+        index = next((i for i, t in enumerate(self.topics) if t.id == topic.id), None)
+        if index is not None:
+            self.topics[index] = topic
         else:
             self.topics.append(topic)
 
diff --git a/server/reflector/pipelines/main_live_pipeline.py b/server/reflector/pipelines/main_live_pipeline.py
index 88e1bffd..bf11bdf3 100644
--- a/server/reflector/pipelines/main_live_pipeline.py
+++ b/server/reflector/pipelines/main_live_pipeline.py
@@ -46,7 +46,9 @@ from reflector.processors import (
     TranscriptTranslatorProcessor,
 )
 from reflector.processors.types import AudioDiarizationInput
-from reflector.processors.types import TitleSummary as TitleSummaryProcessorType
+from reflector.processors.types import (
+    TitleSummaryWithId as TitleSummaryWithIdProcessorType,
+)
 from reflector.processors.types import Transcript as TranscriptProcessorType
 from reflector.settings import settings
 from reflector.ws_manager import WebsocketManager, get_ws_manager
@@ -163,6 +165,8 @@ class PipelineMainBase(PipelineRunner):
             text=data.transcript.text,
             words=data.transcript.words,
         )
+        if isinstance(data, TitleSummaryWithIdProcessorType):
+            topic.id = data.id
         async with self.transaction():
             transcript = await self.get_transcript()
             await transcripts_controller.upsert_topic(transcript, topic)
@@ -302,7 +306,8 @@ class PipelineMainDiarization(PipelineMainBase):
         # XXX translation is lost when converting our data model to the processor model
         transcript = await self.get_transcript()
         topics = [
-            TitleSummaryProcessorType(
+            TitleSummaryWithIdProcessorType(
+                id=topic.id,
                 title=topic.title,
                 summary=topic.summary,
                 timestamp=topic.timestamp,
diff --git a/server/reflector/processors/types.py b/server/reflector/processors/types.py
index b67f84b9..312f5433 100644
--- a/server/reflector/processors/types.py
+++ b/server/reflector/processors/types.py
@@ -167,6 +167,10 @@ class TitleSummary(BaseModel):
         return f"{minutes:02d}:{seconds:02d}.{milliseconds:03d}"
 
 
+class TitleSummaryWithId(TitleSummary):
+    id: str
+
+
 class FinalLongSummary(BaseModel):
     long_summary: str
     duration: float
@@ -386,4 +390,4 @@ class TranslationLanguages(BaseModel):
 
 class AudioDiarizationInput(BaseModel):
     audio_url: str
-    topics: list[TitleSummary]
+    topics: list[TitleSummaryWithId]