import streamlit as st import anthropic, openai, base64, cv2, glob, json, math, os, pytz, random, re, requests, textract, time, zipfile import plotly.graph_objects as go import streamlit.components.v1 as components from datetime import datetime from audio_recorder_streamlit import audio_recorder from bs4 import BeautifulSoup from collections import deque from dotenv import load_dotenv from gradio_client import Client from huggingface_hub import InferenceClient from io import BytesIO from PIL import Image from PyPDF2 import PdfReader from urllib.parse import quote from xml.etree import ElementTree as ET from openai import OpenAI import extra_streamlit_components as stx from streamlit.runtime.scriptrunner import get_script_run_ctx import asyncio import edge_tts # ensure this is installed (pip install edge-tts) # 🔧 Config & Setup st.set_page_config( page_title="🚲BikeAI🏆 Claude/GPT Research", page_icon="🚲🏆", layout="wide", initial_sidebar_state="auto", menu_items={ 'Get Help': 'https://huggingface.co/awacke1', 'Report a bug': 'https://huggingface.co/spaces/awacke1', 'About': "🚲BikeAI🏆 Claude/GPT Research AI" } ) load_dotenv() openai.api_key = os.getenv('OPENAI_API_KEY') or st.secrets['OPENAI_API_KEY'] anthropic_key = os.getenv("ANTHROPIC_API_KEY_3") or st.secrets["ANTHROPIC_API_KEY"] claude_client = anthropic.Anthropic(api_key=anthropic_key) openai_client = OpenAI(api_key=openai.api_key, organization=os.getenv('OPENAI_ORG_ID')) HF_KEY = os.getenv('HF_KEY') API_URL = os.getenv('API_URL') st.session_state.setdefault('transcript_history', []) st.session_state.setdefault('chat_history', []) st.session_state.setdefault('openai_model', "gpt-4o-2024-05-13") st.session_state.setdefault('messages', []) st.session_state.setdefault('last_voice_input', "") # 🎨 Minimal Custom CSS st.markdown(""" """, unsafe_allow_html=True) # 🔑 Common Utilities def generate_filename(prompt, file_type="md"): ctz = pytz.timezone('US/Central') date_str = datetime.now(ctz).strftime("%m%d_%H%M") safe = re.sub(r'[<>:"/\\\\|?*\n]', ' ', prompt) safe = re.sub(r'\s+', ' ', safe).strip()[:90] return f"{date_str}_{safe}.{file_type}" def create_file(filename, prompt, response): with open(filename, 'w', encoding='utf-8') as f: f.write(prompt + "\n\n" + response) def get_download_link(file): with open(file, "rb") as f: b64 = base64.b64encode(f.read()).decode() return f'📂 Download {os.path.basename(file)}' @st.cache_resource def speech_synthesis_html(result): # This old function can remain as a fallback, but we won't use it after integrating EdgeTTS. html_code = f""" """ components.html(html_code, height=0) #------------add EdgeTTS # --- NEW FUNCTIONS FOR EDGE TTS --- async def edge_tts_generate_audio(text, voice="en-US-AriaNeural", rate=0, pitch=0): """ Generate audio from text using Edge TTS and return the path to the MP3 file. """ if not text.strip(): return None rate_str = f"{rate:+d}%" pitch_str = f"{pitch:+d}Hz" communicate = edge_tts.Communicate(text, voice, rate=rate_str, pitch=pitch_str) out_fn = generate_filename(text,"mp3") await communicate.save(out_fn) return out_fn def speak_with_edge_tts(text, voice="en-US-AriaNeural", rate=0, pitch=0): """ Synchronous wrapper to call the async TTS generation and return the file path. """ return asyncio.run(edge_tts_generate_audio(text, voice, rate, pitch)) def play_and_download_audio(file_path): """ Display an audio player and a download link for the generated MP3 file. """ if file_path and os.path.exists(file_path): st.audio(file_path) st.markdown(get_download_link(file_path), unsafe_allow_html=True) #--------------------------- def process_image(image_path, user_prompt): with open(image_path, "rb") as imgf: image_data = imgf.read() b64img = base64.b64encode(image_data).decode("utf-8") resp = openai_client.chat.completions.create( model=st.session_state["openai_model"], messages=[ {"role": "system", "content": "You are a helpful assistant."}, {"role": "user", "content": [ {"type": "text", "text": user_prompt}, {"type": "image_url", "image_url": {"url": f"data:image/png;base64,{b64img}"}} ]} ], temperature=0.0, ) return resp.choices[0].message.content def process_audio(audio_path): with open(audio_path, "rb") as f: transcription = openai_client.audio.transcriptions.create(model="whisper-1", file=f) st.session_state.messages.append({"role": "user", "content": transcription.text}) return transcription.text def process_video(video_path, seconds_per_frame=1): vid = cv2.VideoCapture(video_path) total = int(vid.get(cv2.CAP_PROP_FRAME_COUNT)) fps = vid.get(cv2.CAP_PROP_FPS) skip = int(fps*seconds_per_frame) frames_b64 = [] for i in range(0, total, skip): vid.set(cv2.CAP_PROP_POS_FRAMES, i) ret, frame = vid.read() if not ret: break _, buf = cv2.imencode(".jpg", frame) frames_b64.append(base64.b64encode(buf).decode("utf-8")) vid.release() return frames_b64 def process_video_with_gpt(video_path, prompt): frames = process_video(video_path) resp = openai_client.chat.completions.create( model=st.session_state["openai_model"], messages=[ {"role":"system","content":"Analyze video frames."}, {"role":"user","content":[ {"type":"text","text":prompt}, *[{"type":"image_url","image_url":{"url":f"data:image/jpeg;base64,{fr}"}} for fr in frames] ]} ] ) return resp.choices[0].message.content def search_arxiv(query): st.write("🔍 Searching ArXiv...") client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern") r1 = client.predict(prompt=query, llm_model_picked="mistralai/Mixtral-8x7B-Instruct-v0.1", stream_outputs=True, api_name="/ask_llm") st.markdown("### Mistral-8x7B-Instruct-v0.1 Result") st.markdown(r1) r2 = client.predict(prompt=query, llm_model_picked="mistralai/Mistral-7B-Instruct-v0.2", stream_outputs=True, api_name="/ask_llm") st.markdown("### Mistral-7B-Instruct-v0.2 Result") st.markdown(r2) return f"{r1}\n\n{r2}" def perform_ai_lookup(q): start = time.time() client = Client("awacke1/Arxiv-Paper-Search-And-QA-RAG-Pattern") # Perform a RAG-based search r = client.predict(q,20,"Semantic Search","mistralai/Mixtral-8x7B-Instruct-v0.1",api_name="/update_with_rag_md") refs = r[0] # Ask model for answer r2 = client.predict(q,"mistralai/Mixtral-8x7B-Instruct-v0.1",True,api_name="/ask_llm") result = f"### 🔎 {q}\n\n{r2}\n\n{refs}" #--------------------------------------------------------------- # Speak results #speech_synthesis_html(r2) # Instead of speech_synthesis_html, use EdgeTTS now: st.markdown(result) # Speak main result audio_file_main = speak_with_edge_tts(r2, voice="en-US-AriaNeural", rate=0, pitch=0) st.write("### Audio Output for Main Result") play_and_download_audio(audio_file_main) # Speak references summaries summaries_text = "Here are the summaries from the references: " + refs.replace('"','') audio_file_refs = speak_with_edge_tts(summaries_text, voice="en-US-AriaNeural", rate=0, pitch=0) st.write("### Audio Output for References Summaries") play_and_download_audio(audio_file_refs) # Extract titles from refs and speak them titles = [] for line in refs.split('\n'): m = re.search(r"\[([^\]]+)\]", line) if m: titles.append(m.group(1)) if titles: titles_text = "Here are the titles of the papers: " + ", ".join(titles) audio_file_titles = speak_with_edge_tts(titles_text, voice="en-US-AriaNeural", rate=0, pitch=0) st.write("### Audio Output for Paper Titles") play_and_download_audio(audio_file_titles) # -------------------------------------------- st.markdown(result) elapsed = time.time()-start st.write(f"Elapsed: {elapsed:.2f} s") fn = generate_filename(q,"md") create_file(fn,q,result) return result def process_with_gpt(text): if not text: return st.session_state.messages.append({"role":"user","content":text}) with st.chat_message("user"): st.markdown(text) with st.chat_message("assistant"): c = openai_client.chat.completions.create( model=st.session_state["openai_model"], messages=st.session_state.messages, stream=False ) ans = c.choices[0].message.content st.write("GPT-4o: " + ans) create_file(generate_filename(text,"md"),text,ans) st.session_state.messages.append({"role":"assistant","content":ans}) return ans def process_with_claude(text): if not text: return with st.chat_message("user"): st.markdown(text) with st.chat_message("assistant"): r = claude_client.messages.create( model="claude-3-sonnet-20240229", max_tokens=1000, messages=[{"role":"user","content":text}] ) ans = r.content[0].text st.write("Claude: " + ans) create_file(generate_filename(text,"md"),text,ans) st.session_state.chat_history.append({"user":text,"claude":ans}) return ans def create_zip_of_files(files): zip_name = "all_files.zip" with zipfile.ZipFile(zip_name,'w') as z: for f in files: z.write(f) return zip_name def get_media_html(p,typ="video",w="100%"): d = base64.b64encode(open(p,'rb').read()).decode() if typ=="video": return f'

' else: return f'

' def create_media_gallery(): st.header("🎬 Media Gallery") tabs = st.tabs(["🖼️ Images", "🎵 Audio", "🎥 Video"]) with tabs[0]: imgs = glob.glob("*.png")+glob.glob("*.jpg") if imgs: c = st.slider("Cols",1,5,3) cols = st.columns(c) for i,f in enumerate(imgs): with cols[i%c]: st.image(Image.open(f),use_container_width=True) if st.button(f"👀 Analyze {os.path.basename(f)}"): a = process_image(f,"Describe this image.") st.markdown(a) with tabs[1]: auds = glob.glob("*.mp3")+glob.glob("*.wav") for a in auds: with st.expander(f"🎵 {os.path.basename(a)}"): st.markdown(get_media_html(a,"audio"),unsafe_allow_html=True) if st.button(f"Transcribe {os.path.basename(a)}"): t = process_audio(a) st.write(t) with tabs[2]: vids = glob.glob("*.mp4") for v in vids: with st.expander(f"🎥 {os.path.basename(v)}"): st.markdown(get_media_html(v,"video"),unsafe_allow_html=True) if st.button(f"Analyze {os.path.basename(v)}"): a = process_video_with_gpt(v,"Describe video.") st.markdown(a) def display_file_manager(): st.sidebar.title("📁 File Management") files = sorted(glob.glob("*.md"),reverse=True) if st.sidebar.button("🗑 Delete All"): for f in files: os.remove(f) st.rerun() if st.sidebar.button("⬇️ Download All"): z= create_zip_of_files(files) st.sidebar.markdown(get_download_link(z),unsafe_allow_html=True) for f in files: col1,col2,col3,col4 = st.sidebar.columns([1,3,1,1]) with col1: if st.button("🌐",key="v"+f): st.session_state.current_file=f c=open(f,'r',encoding='utf-8').read() st.write(c) with col2: st.markdown(get_download_link(f),unsafe_allow_html=True) with col3: if st.button("📂",key="e"+f): st.session_state.current_file=f st.session_state.file_content=open(f,'r',encoding='utf-8').read() with col4: if st.button("🗑",key="d"+f): os.remove(f) st.experimental_rerun() def main(): st.sidebar.markdown("### 🚲BikeAI🏆 Multi-Agent Research AI") tab_main = st.radio("Action:",["🎤 Voice Input","📸 Media Gallery","🔍 Search ArXiv","📝 File Editor"],horizontal=True) model_choice = st.sidebar.radio("AI Model:", ["Arxiv","GPT-4o","Claude-3","GPT+Claude+Arxiv"], index=0) # Declare the component mycomponent = components.declare_component("mycomponent", path="mycomponent") val = mycomponent(my_input_value="Hello") if val: # Strip whitespace and newlines from the end of the user input user_input = val.strip() if user_input: if model_choice == "GPT-4o": process_with_gpt(user_input) elif model_choice == "Claude-3": process_with_claude(user_input) elif model_choice == "Arxiv": st.subheader("Arxiv Only Results:") perform_ai_lookup(user_input) else: col1,col2,col3=st.columns(3) with col1: st.subheader("GPT-4o Omni:") try: process_with_gpt(user_input) except: st.write('GPT 4o error') with col2: st.subheader("Claude-3 Sonnet:") try: process_with_claude(user_input) except: st.write('Claude error') with col3: st.subheader("Arxiv + Mistral:") try: r = perform_ai_lookup(user_input) st.markdown(r) except: st.write("Arxiv error") if tab_main == "🎤 Voice Input": st.subheader("🎤 Voice Recognition") user_text = st.text_area("Message:", height=100) # Strip whitespace and newlines user_text = user_text.strip() if st.button("Send 📨"): if user_text: if model_choice == "GPT-4o": process_with_gpt(user_text) elif model_choice == "Claude-3": process_with_claude(user_text) elif model_choice == "Arxiv": st.subheader("Arxiv Only Results:") perform_ai_lookup(user_text) else: col1,col2,col3=st.columns(3) with col1: st.subheader("GPT-4o Omni:") process_with_gpt(user_text) with col2: st.subheader("Claude-3 Sonnet:") process_with_claude(user_text) with col3: st.subheader("Arxiv & Mistral:") res = perform_ai_lookup(user_text) st.markdown(res) st.subheader("📜 Chat History") t1,t2=st.tabs(["Claude History","GPT-4o History"]) with t1: for c in st.session_state.chat_history: st.write("**You:**", c["user"]) st.write("**Claude:**", c["claude"]) with t2: for m in st.session_state.messages: with st.chat_message(m["role"]): st.markdown(m["content"]) elif tab_main == "📸 Media Gallery": create_media_gallery() elif tab_main == "🔍 Search ArXiv": q=st.text_input("Research query:") if q: q = q.strip() # Strip whitespace and newlines if q: r=search_arxiv(q) st.markdown(r) elif tab_main == "📝 File Editor": if getattr(st.session_state,'current_file',None): st.subheader(f"Editing: {st.session_state.current_file}") new_text = st.text_area("Content:", st.session_state.file_content, height=300) # Here also you can strip if needed, but usually for file editing you might not want to. if st.button("Save"): with open(st.session_state.current_file,'w',encoding='utf-8') as f: f.write(new_text) st.success("Updated!") display_file_manager() if __name__=="__main__": main()