update: LE reasoning prompt

app.py

@@ -489,12 +489,11 @@ CRITICAL: When analyzing timestamps, you must verify that in the duration specif
 2. The speaker talks continuously for at least 20 seconds
 3. The clip ends BEFORE any interruption or speaker change
 """
-            print(" , ".join(speaker_mapping.values()))
             reasoning_prompt = f"""For each Speaker {" , ".join(speaker_mapping.values())}
 in the transcript: {transcript}
 
 Your job is to generate the thinking about the short social media clips for each speaker where they discuss. Think step by step and return a JSON at the end of the thinking.
-You can skip a speaker if they don't have enough content to generate the clips.
+You can skip a speaker if they don't have enough content to generate the clips. Goal of the reasoning is to find the sentence of a specific speaker where they talk about that topic Since I need to make seperate clips of each speaker. SO all the sentences between the start and end of the topic must be from the same speaker.
 Return Format:
 - Name of the Speaker
 - Detailed Step by Step Thinking for each speaker from thier content and the topic they are talking about