Spaces:
Sleeping
Sleeping
Upload 5 files
Browse files- .env +3 -0
- app.py +269 -0
- chat_history.json +10 -0
- preferences.json +3 -0
- requirements.txt +6 -0
.env
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
GROQ_API_KEY=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
|
| 2 |
+
BRIGHT_API_KEY=ded1bc00-4d84-4dee-b4d6-c896cadf2417
|
| 3 |
+
BRIGHT_ZONE=web_unlocker2
|
app.py
ADDED
|
@@ -0,0 +1,269 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import os
|
| 2 |
+
import json
|
| 3 |
+
import requests
|
| 4 |
+
import re
|
| 5 |
+
from bs4 import BeautifulSoup
|
| 6 |
+
from groq import Groq
|
| 7 |
+
import gradio as gr
|
| 8 |
+
from dotenv import load_dotenv
|
| 9 |
+
from youtube_transcript_api import YouTubeTranscriptApi
|
| 10 |
+
|
| 11 |
+
load_dotenv()
|
| 12 |
+
|
| 13 |
+
|
| 14 |
+
# ENV VARIABLES
|
| 15 |
+
|
| 16 |
+
GROQ_API_KEY = os.getenv("GROQ_API_KEY")
|
| 17 |
+
BRIGHT_API_KEY = os.getenv("BRIGHT_API_KEY")
|
| 18 |
+
BRIGHT_ZONE = os.getenv("BRIGHT_ZONE")
|
| 19 |
+
|
| 20 |
+
client = Groq(api_key=GROQ_API_KEY)
|
| 21 |
+
|
| 22 |
+
CHAT_FILE = "chat_history.json"
|
| 23 |
+
PREF_FILE = "preferences.json"
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
# SAFE LOAD & SAVE
|
| 27 |
+
|
| 28 |
+
def load_json(file, default):
|
| 29 |
+
if os.path.exists(file):
|
| 30 |
+
try:
|
| 31 |
+
with open(file, "r") as f:
|
| 32 |
+
return json.load(f)
|
| 33 |
+
except:
|
| 34 |
+
return default
|
| 35 |
+
return default
|
| 36 |
+
|
| 37 |
+
def save_json(file, data):
|
| 38 |
+
with open(file, "w") as f:
|
| 39 |
+
json.dump(data, f, indent=4)
|
| 40 |
+
|
| 41 |
+
conversation_history = load_json(CHAT_FILE, [])
|
| 42 |
+
user_preferences = load_json(PREF_FILE, {"style": "Default"})
|
| 43 |
+
|
| 44 |
+
|
| 45 |
+
# BRIGHT DATA SCRAPER
|
| 46 |
+
|
| 47 |
+
def brightdata_request(target_url):
|
| 48 |
+
response = requests.post(
|
| 49 |
+
"https://api.brightdata.com/request",
|
| 50 |
+
headers={
|
| 51 |
+
"Content-Type": "application/json",
|
| 52 |
+
"Authorization": f"Bearer {BRIGHT_API_KEY}"
|
| 53 |
+
},
|
| 54 |
+
json={
|
| 55 |
+
"zone": BRIGHT_ZONE,
|
| 56 |
+
"url": target_url,
|
| 57 |
+
"format": "raw"
|
| 58 |
+
}
|
| 59 |
+
)
|
| 60 |
+
return response.text
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
def scrape_goodreads(url):
|
| 64 |
+
html = brightdata_request(url)
|
| 65 |
+
soup = BeautifulSoup(html, "html.parser")
|
| 66 |
+
|
| 67 |
+
books = []
|
| 68 |
+
rows = soup.find_all("tr")
|
| 69 |
+
|
| 70 |
+
for row in rows:
|
| 71 |
+
title_tag = row.find("a", class_="bookTitle")
|
| 72 |
+
author_tag = row.find("a", class_="authorName")
|
| 73 |
+
rating_tag = row.find("span", class_="minirating")
|
| 74 |
+
|
| 75 |
+
if title_tag and author_tag and rating_tag:
|
| 76 |
+
books.append({
|
| 77 |
+
"title": title_tag.get_text(strip=True),
|
| 78 |
+
"author": author_tag.get_text(strip=True),
|
| 79 |
+
"rating": rating_tag.get_text(strip=True)
|
| 80 |
+
})
|
| 81 |
+
|
| 82 |
+
return books[:10]
|
| 83 |
+
|
| 84 |
+
|
| 85 |
+
def qa_bot(url, question):
|
| 86 |
+
books = scrape_goodreads(url)
|
| 87 |
+
|
| 88 |
+
if not books:
|
| 89 |
+
return "No book data found."
|
| 90 |
+
|
| 91 |
+
context = "\n".join(
|
| 92 |
+
[f"{i+1}. {b['title']} by {b['author']} - {b['rating']}"
|
| 93 |
+
for i, b in enumerate(books)]
|
| 94 |
+
)
|
| 95 |
+
|
| 96 |
+
system_prompt = f"""
|
| 97 |
+
You are a helpful assistant.
|
| 98 |
+
Answer ONLY using the following scraped Goodreads data.
|
| 99 |
+
|
| 100 |
+
{context}
|
| 101 |
+
"""
|
| 102 |
+
|
| 103 |
+
response = client.chat.completions.create(
|
| 104 |
+
model="llama-3.1-8b-instant",
|
| 105 |
+
messages=[
|
| 106 |
+
{"role": "system", "content": system_prompt},
|
| 107 |
+
{"role": "user", "content": question}
|
| 108 |
+
]
|
| 109 |
+
)
|
| 110 |
+
|
| 111 |
+
return response.choices[0].message.content
|
| 112 |
+
|
| 113 |
+
# YOUTUBE TRANSCRIPT
|
| 114 |
+
|
| 115 |
+
def extract_video_id(url):
|
| 116 |
+
pattern = r"(?:v=|youtu\.be/)([a-zA-Z0-9_-]{11})"
|
| 117 |
+
match = re.search(pattern, url)
|
| 118 |
+
return match.group(1) if match else None
|
| 119 |
+
|
| 120 |
+
|
| 121 |
+
def get_youtube_transcript(url):
|
| 122 |
+
video_id = extract_video_id(url)
|
| 123 |
+
|
| 124 |
+
if not video_id:
|
| 125 |
+
return "Invalid YouTube URL."
|
| 126 |
+
|
| 127 |
+
try:
|
| 128 |
+
api = YouTubeTranscriptApi()
|
| 129 |
+
transcript = api.fetch(video_id)
|
| 130 |
+
full_text = " ".join([entry.text for entry in transcript])
|
| 131 |
+
return full_text
|
| 132 |
+
except:
|
| 133 |
+
return "No transcript available for this video."
|
| 134 |
+
|
| 135 |
+
|
| 136 |
+
def youtube_qa(video_url, question):
|
| 137 |
+
transcript = get_youtube_transcript(video_url)
|
| 138 |
+
|
| 139 |
+
if transcript.startswith("No") or transcript.startswith("Invalid"):
|
| 140 |
+
return transcript
|
| 141 |
+
|
| 142 |
+
system_prompt = f"""
|
| 143 |
+
You are a helpful assistant.
|
| 144 |
+
Answer ONLY using this transcript.
|
| 145 |
+
|
| 146 |
+
Transcript:
|
| 147 |
+
{transcript[:6000]}
|
| 148 |
+
"""
|
| 149 |
+
|
| 150 |
+
response = client.chat.completions.create(
|
| 151 |
+
model="llama-3.1-8b-instant",
|
| 152 |
+
messages=[
|
| 153 |
+
{"role": "system", "content": system_prompt},
|
| 154 |
+
{"role": "user", "content": question}
|
| 155 |
+
]
|
| 156 |
+
)
|
| 157 |
+
|
| 158 |
+
return response.choices[0].message.content
|
| 159 |
+
|
| 160 |
+
|
| 161 |
+
# VERSION 3 — MEMORY CHAT
|
| 162 |
+
|
| 163 |
+
def chat_with_memory(user_message, preference_text):
|
| 164 |
+
|
| 165 |
+
global conversation_history, user_preferences
|
| 166 |
+
|
| 167 |
+
if preference_text and preference_text.strip():
|
| 168 |
+
user_preferences["style"] = preference_text
|
| 169 |
+
save_json(PREF_FILE, user_preferences)
|
| 170 |
+
|
| 171 |
+
system_prompt = f"""
|
| 172 |
+
You are a helpful AI assistant.
|
| 173 |
+
|
| 174 |
+
User Preferences:
|
| 175 |
+
{user_preferences.get("style", "Default")}
|
| 176 |
+
|
| 177 |
+
Follow the preferred style in all responses.
|
| 178 |
+
Maintain conversational memory.
|
| 179 |
+
"""
|
| 180 |
+
|
| 181 |
+
messages = [{"role": "system", "content": system_prompt}]
|
| 182 |
+
messages.extend(conversation_history)
|
| 183 |
+
messages.append({"role": "user", "content": user_message})
|
| 184 |
+
|
| 185 |
+
response = client.chat.completions.create(
|
| 186 |
+
model="llama-3.1-8b-instant",
|
| 187 |
+
messages=messages
|
| 188 |
+
)
|
| 189 |
+
|
| 190 |
+
assistant_reply = response.choices[0].message.content
|
| 191 |
+
|
| 192 |
+
conversation_history.append({"role": "user", "content": user_message})
|
| 193 |
+
conversation_history.append({"role": "assistant", "content": assistant_reply})
|
| 194 |
+
|
| 195 |
+
save_json(CHAT_FILE, conversation_history)
|
| 196 |
+
|
| 197 |
+
return assistant_reply
|
| 198 |
+
|
| 199 |
+
|
| 200 |
+
def clear_memory():
|
| 201 |
+
global conversation_history
|
| 202 |
+
conversation_history = []
|
| 203 |
+
save_json(CHAT_FILE, [])
|
| 204 |
+
return []
|
| 205 |
+
|
| 206 |
+
|
| 207 |
+
# GRADIO MULTI-TAB UI
|
| 208 |
+
|
| 209 |
+
with gr.Blocks() as demo:
|
| 210 |
+
|
| 211 |
+
gr.Markdown("# 🚀 Full AI Assistant (V1 + V2 + V3)")
|
| 212 |
+
|
| 213 |
+
# TAB 1
|
| 214 |
+
with gr.Tab("🌐 Website Scraper Q&A"):
|
| 215 |
+
url_input = gr.Textbox(label="Enter Goodreads URL",
|
| 216 |
+
value="https://www.goodreads.com/list/show/1.Best_Books_Ever")
|
| 217 |
+
question_input = gr.Textbox(label="Ask your question")
|
| 218 |
+
output1 = gr.Textbox(label="Answer")
|
| 219 |
+
btn1 = gr.Button("Submit")
|
| 220 |
+
|
| 221 |
+
btn1.click(qa_bot, inputs=[url_input, question_input], outputs=output1)
|
| 222 |
+
|
| 223 |
+
# TAB 2
|
| 224 |
+
with gr.Tab("🎥 YouTube Transcript Q&A"):
|
| 225 |
+
video_input = gr.Textbox(label="Enter YouTube URL")
|
| 226 |
+
yt_question = gr.Textbox(label="Ask your question")
|
| 227 |
+
output2 = gr.Textbox(label="Answer")
|
| 228 |
+
btn2 = gr.Button("Submit")
|
| 229 |
+
|
| 230 |
+
btn2.click(youtube_qa, inputs=[video_input, yt_question], outputs=output2)
|
| 231 |
+
|
| 232 |
+
# TAB 3
|
| 233 |
+
with gr.Tab("🧠 Multi-Turn Assistant (Persistent Memory)"):
|
| 234 |
+
|
| 235 |
+
chatbot = gr.Chatbot(label="Conversation", value=conversation_history)
|
| 236 |
+
|
| 237 |
+
preference_input = gr.Textbox(
|
| 238 |
+
label="User Preferences (Optional)",
|
| 239 |
+
placeholder="Example: Respond formally / Be concise / Use bullet points"
|
| 240 |
+
)
|
| 241 |
+
|
| 242 |
+
user_message = gr.Textbox(label="Your Message")
|
| 243 |
+
send_btn = gr.Button("Send")
|
| 244 |
+
clear_btn = gr.Button("Clear Memory")
|
| 245 |
+
|
| 246 |
+
def chat_interface(user_message, preference_text, chat_display):
|
| 247 |
+
if not user_message.strip():
|
| 248 |
+
return "", chat_display
|
| 249 |
+
|
| 250 |
+
assistant_reply = chat_with_memory(user_message, preference_text)
|
| 251 |
+
|
| 252 |
+
chat_display.append({"role": "user", "content": user_message})
|
| 253 |
+
chat_display.append({"role": "assistant", "content": assistant_reply})
|
| 254 |
+
|
| 255 |
+
return "", chat_display
|
| 256 |
+
|
| 257 |
+
send_btn.click(
|
| 258 |
+
chat_interface,
|
| 259 |
+
inputs=[user_message, preference_input, chatbot],
|
| 260 |
+
outputs=[user_message, chatbot]
|
| 261 |
+
)
|
| 262 |
+
|
| 263 |
+
clear_btn.click(
|
| 264 |
+
clear_memory,
|
| 265 |
+
outputs=chatbot
|
| 266 |
+
)
|
| 267 |
+
|
| 268 |
+
if __name__ == "__main__":
|
| 269 |
+
demo.launch()
|
chat_history.json
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
[
|
| 2 |
+
{
|
| 3 |
+
"role": "user",
|
| 4 |
+
"content": "can you tell me my name"
|
| 5 |
+
},
|
| 6 |
+
{
|
| 7 |
+
"role": "assistant",
|
| 8 |
+
"content": "I don't have any information about your name. We just started our conversation, and I'm a new AI each time you interact with me. I don't have any prior knowledge or memory of previous conversations. \n\nHowever, I can ask: If you'd like, you can tell me your name, and I can remember it and use it in our conversation."
|
| 9 |
+
}
|
| 10 |
+
]
|
preferences.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"style": "provide all ways python code"
|
| 3 |
+
}
|
requirements.txt
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
gradio
|
| 2 |
+
groq
|
| 3 |
+
requests
|
| 4 |
+
beautifulsoup4
|
| 5 |
+
python-dotenv
|
| 6 |
+
youtube-transcript-api
|