add profanity filter, post-process topic/title

server/reflector/processors/audio_transcript.py

@@ -1,4 +1,6 @@
+from profanityfilter import ProfanityFilter
 from prometheus_client import Counter, Histogram
+
 from reflector.processors.base import Processor
 from reflector.processors.types import AudioFile, Transcript
 
@@ -38,6 +40,8 @@ class AudioTranscriptProcessor(Processor):
         self.m_transcript_call = self.m_transcript_call.labels(name)
         self.m_transcript_success = self.m_transcript_success.labels(name)
         self.m_transcript_failure = self.m_transcript_failure.labels(name)
+        self.profanity_filter = ProfanityFilter()
+        self.profanity_filter.set_censor("|*|")
         super().__init__(*args, **kwargs)
 
     async def _push(self, data: AudioFile):
@@ -56,3 +60,11 @@ class AudioTranscriptProcessor(Processor):
 
     async def _transcript(self, data: AudioFile):
         raise NotImplementedError
+
+    def filter_profanity(self, text: str) -> str:
+        """
+        Remove censored words from the transcript
+        """
+        text = self.profanity_filter.censor(text)
+        text = text.replace("|*|", "")
+        return text

server/reflector/processors/audio_transcript_modal.py

@@ -15,6 +15,7 @@ API will be a POST request to TRANSCRIPT_URL:
 from time import monotonic
 
 import httpx
+
 from reflector.processors.audio_transcript import AudioTranscriptProcessor
 from reflector.processors.audio_transcript_auto import AudioTranscriptAutoProcessor
 from reflector.processors.types import AudioFile, Transcript, TranslationLanguages, Word
@@ -86,7 +87,7 @@ class AudioTranscriptModalProcessor(AudioTranscriptProcessor):
             if source_language != target_language and target_language in result["text"]:
                 translation = result["text"][target_language]
             text = result["text"][source_language]
-
+            text = self.filter_profanity(text)
             transcript = Transcript(
                 text=text,
                 translation=translation,

server/reflector/processors/transcript_final_title.py

@@ -60,6 +60,8 @@ class TranscriptFinalTitleProcessor(Processor):
 
         accumulated_titles = ".".join([chunk.title for chunk in self.chunks])
         title_result = await self.get_title(accumulated_titles)
+        final_title = self.llm.ensure_casing(title_result["title"])
+        final_title = self.llm.trim_title(final_title)
 
-        final_title = FinalTitle(title=title_result["title"])
+        final_title = FinalTitle(title=final_title)
         await self.emit(final_title)

server/reflector/processors/transcript_topic_detector.py

@@ -55,8 +55,11 @@ class TranscriptTopicDetectorProcessor(Processor):
         self.logger.info(f"Topic detector got {len(text)} length transcript")
         topic_result = await self.get_topic(text=text)
 
+        title = self.llm.ensure_casing(topic_result["title"])
+        title = self.llm.trim_title(title)
+
         summary = TitleSummary(
-            title=self.llm.ensure_casing(topic_result["title"]),
+            title=title,
             summary=topic_result["summary"],
             timestamp=self.transcript.timestamp,
             duration=self.transcript.duration,