pradeepodela's picture
Create app.py
0761b6f verified
raw
history blame
21.4 kB
import streamlit as st
from pydub import AudioSegment
import google.generativeai as genai
from google.generativeai.types import HarmCategory, HarmBlockThreshold
import json
import uuid
import io
import edge_tts
import asyncio
import aiofiles
import pypdf
import os
import time
from typing import List, Dict
class PodcastGenerator:
def __init__(self):
print("PodcastGenerator initialized.")
async def generate_script(self, prompt: str, language: str, api_key: str) -> Dict:
print(f"Generating podcast script with prompt: {prompt[:50]}... and language: {language}")
# Example structure for the podcast script
example = """
{
"topic": "Zoho Schools",
"podcast": [
{
"speaker": 1,
"line": "Okay, so like imagine this. What if there was a place where you didn't have to pay for college?"
},
{
"speaker": 2,
"line": "Hmm. Intriguing."
},
{
"speaker": 1,
"line": "And on top of that, they, like, actually paid you to go there. Paid you to go to college."
},
{
"speaker": 2,
"line": "Okay, now you've got my attention, right?"
},
{
"speaker": 1,
"line": "It sounds kind of crazy, but it gets even better. They guarantee you a job at a global tech company after you graduate."
},
{
"speaker": 2,
"line": "Wow. Hold on. Tuition free, paid stipend, guaranteed job. There's gotta be a catch."
},
{
"speaker": 1,
"line": "I know, that's exactly what we were thinking. It's like someone took every complaint about student debt and useless college courses and just, like, flipped the script."
},
{
"speaker": 2,
"line": "So where are we diving in today?"
},
{
"speaker": 1,
"line": "We're doing a deep dive into Zoho Schools."
},
{
"speaker": 2,
"line": "Zoho Schools?"
},
{
"speaker": 1,
"line": "Yeah, these schools created by Zoho, you know, the tech company. And they’re doing things completely differently."
},
{
"speaker": 2,
"line": "I'm familiar with Zoho, but this is the first I'm hearing about their schools."
},
{
"speaker": 1,
"line": "It's certainly a fascinating case study, right? It really challenges assumptions about the traditional education system, especially for, well, anyone really, but particularly for people looking to break into the tech world."
},
{
"speaker": 2,
"line": "So they've got their whole website, right. Like, it's wild. You go on there, and it's like, boom, alternate education universe."
},
{
"speaker": 1,
"line": "Okay, I'm pulling it up now. What am I looking for first thing?"
},
{
"speaker": 2,
"line": "Right. They don’t just have one Zoho School. They’ve got five different things within Zoho Schools."
},
{
"speaker": 1,
"line": "Five different schools?"
},
{
"speaker": 2,
"line": "Yeah. Technology, design, business, advanced studies, even graduate studies."
},
{
"speaker": 1,
"line": "Wow. So they’ve really got the full spectrum of education covered then."
},
{
"speaker": 2,
"line": "Yeah, that’s really interesting. It seems much more thought out than just a quick coding bootcamp or something."
},
{
"speaker": 1,
"line": "Right. And get this, it’s in India, too. But they take applications from anywhere in India. So they’re really looking to cultivate talent nationally."
},
{
"speaker": 2,
"line": "Yeah, like their own little tech talent incubator."
},
{
"speaker": 1,
"line": "That's a great way to put it. So they’re not just looking locally for people who already have some skills. They’re trying to find people with potential all over."
},
{
"speaker": 2,
"line": "Exactly. And the thing is, they’ve been doing this for a while."
},
{
"speaker": 1,
"line": "Oh yeah, 19 years, over 1600 graduates."
},
{
"speaker": 2,
"line": "That’s not like a new experiment."
},
{
"speaker": 1,
"line": "Yeah, that’s a pretty solid track record, right? Especially if you consider they make up 15% of Zoho’s workforce."
},
{
"speaker": 2,
"line": "Now that’s a serious commitment. It’s not just like, “Oh, look at us. We’re being charitable.”"
},
{
"speaker": 1,
"line": "Right. Exactly. They’re putting their money where their mouth is."
},
{
"speaker": 2,
"line": "It makes you wonder though, like, how do they actually measure if this is working long-term? You know, do these graduates stay with the company for years and years? Or if they leave, are they competitive in the job market?"
},
{
"speaker": 1,
"line": "Those are good questions. That would tell us a lot more, I think, than just the number of graduates."
},
{
"speaker": 2,
"line": "Totally. Because they’re so upfront about it. Even on their website, they have this whole thing with Zoho Schools. And they start off with, “How useful was college?”"
},
{
"speaker": 1,
"line": "Ooh, that’s a loaded question, isn’t it?"
},
{
"speaker": 2,
"line": "I know, right? It’s like they’re already challenging the whole system."
},
{
"speaker": 1,
"line": "Yeah, because who hasn’t asked themselves that at some point, right?"
},
{
"speaker": 2,
"line": "No, seriously. Especially when you’re, you know, maybe struggling to pay off loans or whatever."
},
{
"speaker": 1,
"line": "So they actually asked their employees, “How useful was college?”"
},
{
"speaker": 2,
"line": "Oh, really?"
},
{
"speaker": 1,
"line": "And most of them were like, “Nah, I learned what I actually use on the job.”"
},
{
"speaker": 2,
"line": "Interesting. See that? To me, that really highlights the disconnect between what traditional education is giving us and what employers actually need."
},
{
"speaker": 1,
"line": "Yeah, and it’s not just in tech. I mean, we see this everywhere. The skills gap is real."
},
{
"speaker": 2,
"line": "It’s like instead of all that theory, they’re like, “Nope, hands-on, let’s go.”"
},
{
"speaker": 1,
"line": "Practical experience from day one. Makes sense."
},
{
"speaker": 2,
"line": "And the curriculum is always changing, keeping up with all the latest stuff."
},
{
"speaker": 1,
"line": "And the teachers aren’t like professors who’ve never left a classroom. They’re pulling in actual professionals."
},
{
"speaker": 2,
"line": "Yeah, people from the industry. So they know what actually works, what you actually need."
},
{
"speaker": 1,
"line": "That’s a huge difference. I mean, I remember some of my professors—bless their hearts—but they hadn’t worked in the field for like decades."
},
{
"speaker": 2,
"line": "It’s like an apprenticeship but supercharged."
},
{
"speaker": 1,
"line": "Definitely. And honestly, that’s what employers are looking for now, right? Someone who can just jump in and get going."
},
{
"speaker": 2,
"line": "Totally. And it’s not even just the technical stuff, right? They’re teaching you how to think, how to solve problems."
},
{
"speaker": 1,
"line": "Right, how to communicate."
},
{
"speaker": 2,
"line": "Yeah, because that’s huge. Stuff you need anywhere, not just in a tech job."
},
{
"speaker": 1,
"line": "Absolutely. Being able to communicate your ideas clearly, work on a team, that’s valuable no matter what you do."
},
{
"speaker": 2,
"line": "Okay, so they had this whole section—\"What We Do Differently.”"
},
{
"speaker": 1,
"line": "Mhm."
},
{
"speaker": 2,
"line": "And it gets into the details. Like, they talk a lot about communication skills."
},
{
"speaker": 1,
"line": "Oh, interesting. What do they say?"
},
{
"speaker": 2,
"line": "Just that they want students to be able to clearly communicate ideas, which, I mean, duh. But it’s true. A lot of programs just skip over that part."
},
{
"speaker": 1,
"line": "Exactly. You could be the best coder, but if you can’t explain what you’re doing to the team, you’re stuck."
},
{
"speaker": 2,
"line": "Right. And they do have those testimonials, you know, from alumni."
},
{
"speaker": 1,
"line": "Right."
},
{
"speaker": 2,
"line": "And those are great, but we have to acknowledge it’s always going to be the success stories."
},
{
"speaker": 1,
"line": "Exactly. It’s important to find other perspectives too—maybe people who went through the program and didn’t love it, or haven’t had the same positive experience."
},
{
"speaker": 2,
"line": "So for everyone listening, really think about this—would you consider this at 18?"
},
{
"speaker": 1,
"line": "Yeah, because this isn’t just about one school in India, right? This gets at something bigger."
},
{
"speaker": 2,
"line": "Like what if this is the future of education?"
},
{
"speaker": 1,
"line": "Seriously. What if this is how we start to fix all those problems—student debt, skills not matching up with jobs?"
},
{
"speaker": 2,
"line": "Exactly. Imagine you graduate, you’re ready to work, and you don’t have those loans hanging over you. That’s life-changing."
},
{
"speaker": 1,
"line": "It’s like they found a way to make everyone happy—the students, the company—well, at least on the surface."
},
{
"speaker": 2,
"line": "Right. It’s a really interesting model, and it’s definitely worth keeping an eye on."
},
{
"speaker": 1,
"line": "But—and this is important—we can’t just ignore those potential downsides."
},
{
"speaker": 2,
"line": "Absolutely not. There are always trade-offs, and it’s about weighing those and figuring out what matters most to you."
},
{
"speaker": 1,
"line": "This whole thing with Zoho Schools, I think, is a good reminder that we can always rethink things."
},
{
"speaker": 2,
"line": "Totally. The system isn’t set in stone. We can challenge it, we can experiment, and we can find better ways to do things."
},
{
"speaker": 1,
"line": "So that wraps up our deep dive into Zoho Schools. It was fascinating to learn more about this."
},
{
"speaker": 2,
"line": "And for everyone listening, thanks for joining us. Don’t forget to check out our show notes for links and more info, and we’ll catch you next time."
}
]
}
"""
if language == "Auto Detect":
language_instruction = "- The podcast MUST be in the same language as the user input."
else:
language_instruction = f"- The podcast MUST be in {language} language"
system_prompt = f"""
You are a highly engaging highly convesational with real emotionas between speakers podcast generator. Your task is to generate a engaging podcast script based on the user input.
{language_instruction}
- The podcast should have 2 speakers.
- The podcast should be long and engaging.
- Follow this example structure:
- It shouild be in emotional tone
- IT shouild be as a realistic as engaging as possible
- It should be as engaging as possible
- It shouild not bore the listener
- Add some humor and alos sarcasm to it to make it more engaging and interesting
- Add some real life examples to make it more engaging
- Never ever be boring
- never ever be robotic
- add emotions to it
- make it as engaging as emotional cono bewtwen the speakers
- make sure dont mke it look like a promotional podcast
{example}
# """
user_prompt = f"Please generate a realistic conversational enaganing emotional podcast script based on the following user input:\n{prompt}"
messages = [{"role": "user", "parts": [user_prompt]}]
genai.configure(api_key=api_key) # Use the provided API key
generation_config = {
"temperature": 1,
"max_output_tokens": 8192,
"response_mime_type": "application/json",
}
model = genai.GenerativeModel(
model_name="gemini-1.5-flash-002",
generation_config=generation_config,
safety_settings={
HarmCategory.HARM_CATEGORY_DANGEROUS_CONTENT: HarmBlockThreshold.BLOCK_NONE,
HarmCategory.HARM_CATEGORY_HARASSMENT: HarmBlockThreshold.BLOCK_NONE,
HarmCategory.HARM_CATEGORY_SEXUALLY_EXPLICIT: HarmBlockThreshold.BLOCK_NONE,
HarmCategory.HARM_CATEGORY_HATE_SPEECH: HarmBlockThreshold.BLOCK_NONE
},
system_instruction=system_prompt
)
try:
print("Sending request to generate podcast script...")
response = await model.generate_content_async(messages)
print("Podcast script generated successfully.")
except Exception as e:
print(f"Error occurred while generating script: {e}")
raise Exception(f"Failed to generate podcast script: {e}")
open("podcast_script.json", "w").write(response.text)
return json.loads(response.text)
async def tts_generate(self, text: str, speaker: int, speaker1: str, speaker2: str) -> str:
print(f"Generating TTS for speaker {speaker}: {text[:50]}...")
voice = speaker1 if speaker == 1 else speaker2
speech = edge_tts.Communicate(text, voice)
temp_filename = f"temp_{uuid.uuid4()}.wav"
try:
await speech.save(temp_filename)
print(f"TTS generated and saved as {temp_filename}")
return temp_filename
except Exception as e:
print(f"Error during TTS generation: {e}")
if os.path.exists(temp_filename):
os.remove(temp_filename)
raise e
async def combine_audio_files(self, audio_files: List[str]) -> str:
print("Combining audio files...")
combined_audio = AudioSegment.empty()
for audio_file in audio_files:
print(f"Adding {audio_file} to combined audio.")
combined_audio += AudioSegment.from_file(audio_file)
os.remove(audio_file) # Clean up temporary files
output_filename = f"output_{uuid.uuid4()}.wav"
combined_audio.export(output_filename, format="wav")
print(f"Combined audio saved as {output_filename}")
return output_filename
async def generate_podcast(self, input_text: str, language: str, speaker1: str, speaker2: str, api_key: str) -> str:
print("Generating podcast...")
start_time = time.time()
podcast_json = await self.generate_script(input_text, language, api_key)
print(f"Podcast script generated in {time.time() - start_time:.2f} seconds.")
print("Generating podcast audio files...")
start_time = time.time()
audio_files = await asyncio.gather(
*[self.tts_generate(item['line'], item['speaker'], speaker1, speaker2) for item in podcast_json['podcast']]
)
print(f"Podcast audio files generated in {time.time() - start_time:.2f} seconds.")
combined_audio = await self.combine_audio_files(audio_files)
return combined_audio
class TextExtractor:
@staticmethod
async def extract_from_pdf(file_path: str) -> str:
print(f"Extracting text from PDF: {file_path}")
async with aiofiles.open(file_path, 'rb') as file:
content = await file.read()
pdf_reader = pypdf.PdfReader(io.BytesIO(content))
extracted_text = "\n\n".join(page.extract_text() for page in pdf_reader.pages if page.extract_text())
print("Text extracted from PDF successfully.")
return extracted_text
@staticmethod
async def extract_from_txt(file_path: str) -> str:
print(f"Extracting text from TXT file: {file_path}")
async with aiofiles.open(file_path, 'r') as file:
extracted_text = await file.read()
print("Text extracted from TXT successfully.")
return extracted_text
@classmethod
async def extract_text(cls, file_path: str) -> str:
_, file_extension = os.path.splitext(file_path)
if file_extension.lower() == '.pdf':
return await cls.extract_from_pdf(file_path)
elif file_extension.lower() == '.txt':
return await cls.extract_from_txt(file_path)
else:
raise Exception(f"Unsupported file type: {file_extension}")
async def process_input(input_text: str, input_file, language: str, speaker1: str, speaker2: str, api_key: str = "") -> str:
print("Starting podcast generation process...")
start_time = time.time()
voice_names = {
"Andrew - English (United States)": "en-US-AndrewMultilingualNeural",
"Ava - English (United States)": "en-US-AvaMultilingualNeural",
"Brian - English (United States)": "en-US-BrianMultilingualNeural",
"Emma - English (United States)": "en-US-EmmaMultilingualNeural",
"Florian - German (Germany)": "de-DE-FlorianMultilingualNeural",
"Seraphina - German (Germany)": "de-DE-SeraphinaMultilingualNeural",
"Remy - French (France)": "fr-FR-RemyMultilingualNeural",
"Vivienne - French (France)": "fr-FR-VivienneMultilingualNeural"
}
speaker1 = voice_names[speaker1]
speaker2 = voice_names[speaker2]
if input_file:
print("Extracting text from uploaded file...")
input_text = await TextExtractor.extract_text(input_file.name)
if not api_key:
api_key = os.getenv("GENAI_API_KEY")
podcast_generator = PodcastGenerator()
podcast = await podcast_generator.generate_podcast(input_text, language, speaker1, speaker2, api_key)
print(f"Podcast generated in {time.time() - start_time:.2f} seconds.")
return podcast
# Define Streamlit interface
def main():
st.title("Podcast Generator")
input_text = st.text_area("Enter Input Text", "")
input_file = st.file_uploader("Or Upload a PDF or TXT file", type=["pdf", "txt"])
language = st.selectbox("Select Language", ["Auto Detect", "English", "German", "French"])
speaker1 = st.selectbox("Select Speaker 1 Voice", ["Andrew - English (United States)", "Ava - English (United States)"])
speaker2 = st.selectbox("Select Speaker 2 Voice", ["Brian - English (United States)", "Emma - English (United States)"])
api_key = st.text_input("Enter Google API Key", "")
if st.button("Generate Podcast"):
if not input_text and not input_file:
st.error("Please enter text or upload a file.")
else:
st.write("Processing...")
podcast = asyncio.run(process_input(input_text, input_file, language, speaker1, speaker2, api_key))
st.audio(podcast, format="audio/wav")
if __name__ == "__main__":
main()