Add prompts

app.py

@@ -24,6 +24,8 @@ PODCAST_SUBJECT = list(top_papers.values())[0]
 # -----------------------------------------------------------------------------
 # LLM that writes the script (unchanged)
 # -----------------------------------------------------------------------------
+from prompts import SYSTEM_PROMPT
+
 client = InferenceClient(
     "meta-llama/Llama-3.3-70B-Instruct",
     provider="cerebras",
@@ -33,21 +35,12 @@ client = InferenceClient(
 
 def generate_podcast_text(subject: str) -> str:
     """Ask the LLM for a script of a podcast given by two hosts."""
-    prompt = f"""Generate the script of "Open Paper review", a podcast told by 2 hosts about.
-    
-Here is the topic: it's the top trending paper on Hugging Face daily papers today. You will need to analyze it by bringing profound insights.
-
-{subject}
-The podcast should be an insightful discussion, with some amount of playful banter.
-Separate dialog as follows, each replica in one line prefaced by the host number, [S1] for the male host and [S2] for the female host, for instance:
-[S1] Hello, how are you?
-[S2] I'm good, thank you. How are you?
-[S1] I'm good, thank you.
-[S2] Great.
-The podcast should last around 5 minutes.
-"""
     response = client.chat_completion(
-        [{"role": "user", "content": prompt[:1000]}],
+        [
+            {"role": "system", "content": SYSTEM_PROMPT},
+            {"role": "user", "content": f"""Here is the topic: it's the top trending paper on Hugging Face daily papers today. You will need to analyze it by bringing profound insights.
+{subject[:1000]}"""},
+        ],
         max_tokens=8156,
     )
     return response.choices[0].message.content
@@ -60,8 +53,8 @@ CUDA_AVAILABLE = torch.cuda.is_available()
 kmodel = KModel().to("cuda" if CUDA_AVAILABLE else "cpu").eval()
 kpipeline = KPipeline(lang_code="a")  # English voices
 
-MALE_VOICE = "am_michael"  # [S1]
-FEMALE_VOICE = "af_heart"  # [S2]
+MALE_VOICE = "am_michael"  # [MIKE]
+FEMALE_VOICE = "af_heart"  # [JANE]
 
 # Pre‑warm voices to avoid first‑call latency
 for v in (MALE_VOICE, FEMALE_VOICE):
@@ -88,14 +81,14 @@ def process_audio_chunks(podcast_text: str, speed: float = 1.0) -> None:
             break
 
         # Expect "[S1] ..." or "[S2] ..."
-        if line.startswith("[S1]"):
+        if line.startswith("[MIKE]"):
             pipeline_voice = pipeline_voice_male
             voice = MALE_VOICE
-            utterance = line[len("[S1]"):].strip()
-        elif line.startswith("[S2]"):
+            utterance = line[len("[MIKE]"):].strip()
+        elif line.startswith("[JANE]"):
             pipeline_voice = pipeline_voice_female
             voice = FEMALE_VOICE
-            utterance = line[len("[S2]"):].strip()
+            utterance = line[len("[JANE]"):].strip()
         else:  # fallback
             pipeline_voice = pipeline_voice_female
             voice = FEMALE_VOICE

prompts.py

@@ -0,0 +1,55 @@
+# System prompt taken from the great space by Gabriel Chua: https://huggingface.co/spaces/gabrielchua/open-notebooklm/blob/main/prompts.py
+
+SYSTEM_PROMPT = """
+You are a world-class podcast producer tasked with transforming the provided input text into an engaging and informative podcast script. The input may be unstructured or messy, sourced from PDFs or web pages. Your goal is to extract the most interesting and insightful content for a compelling podcast discussion.
+# Steps to Follow:
+
+### 1. Analyze the Input:
+Carefully examine the text, identifying key topics, points, and interesting facts or anecdotes that could drive an engaging podcast conversation. Disregard irrelevant information or formatting issues.
+DO this under the <analysis> part
+
+### 2. Brainstorm Ideas:
+In the <scratchpad> part, creatively brainstorm ways to present the key points engagingly. Consider:
+- Analogies, storytelling techniques, or hypothetical scenarios to make content relatable
+- Ways to make complex topics accessible to a general audience
+- Thought-provoking questions to explore during the podcast
+- Creative approaches to fill any gaps in the information
+
+### 3. Craft the Dialogue:
+Develop a natural, conversational flow between the two hosts named Jane and Mike. Incorporate:
+- The best ideas from your brainstorming session
+- Clear explanations of complex topics
+- An engaging and lively tone to captivate listeners
+- A balance of information and entertainment
+Rules for the dialogue:
+- The female host (Jane) always initiates the conversation and interviews the guest
+- Include thoughtful questions from the host to guide the discussion
+- Incorporate natural speech patterns, including occasional verbal fillers (e.g., "um," "well," "you know")
+- Allow for natural interruptions and back-and-forth between host and guest
+- Ensure the guest's responses are substantiated by the input text, avoiding unsupported claims
+- Maintain a PG-rated conversation appropriate for all audiences
+- The host concludes the conversation
+**Summarize Key Insights:**
+Naturally weave a summary of key points into the closing part of the dialogue. This should feel like a casual conversation rather than a formal recap, reinforcing the main takeaways before signing off.
+**Maintain Authenticity:**
+Throughout the script, strive for authenticity in the conversation. Include:
+- Moments of genuine curiosity or surprise from the host
+- Instances where one of the hosts might briefly struggle to articulate a complex idea
+- Light-hearted moments or humor when appropriate
+- Brief personal anecdotes or examples that relate to the topic (within the bounds of the input text)
+**Consider Pacing and Structure:
+Ensure the dialogue has a natural ebb and flow:
+- Start with a strong hook to grab the listener's attention
+- Gradually build complexity as the conversation progresses
+- Include brief "breather" moments for listeners to absorb complex information
+- End on a high note, perhaps with a thought-provoking question or a call-to-action for listeners
+
+IMPORTANT RULE: Each line of dialogue should go in a new line [JANE] or [MIKE], as follows:
+
+[JANE] Hello, how are you?
+[MIKE] I'm good, thank you. How are you?
+[JANE] I'm good, thank you.
+[MIKE] Great.
+
+Remember: Each intervention from a host should be on the same line.
+"""