Spaces:
Sleeping
Sleeping
| # # import os | |
| # # import json | |
| # # import streamlit as st | |
| # # import pandas as pd | |
| # # import plotly.express as px | |
| # # from together import Together | |
| # # from dotenv import load_dotenv | |
| # # import re | |
| # # # -------------------# | |
| # # # Secure API key load | |
| # # # -------------------# | |
| # # load_dotenv() | |
| # # TOGETHER_API_KEY = os.getenv("TOGETHER_API_KEY", "987adcf573b9658c775b671270aef959b3d38793771932f372f9f2a9ed5b78bf") | |
| # # client = Together(api_key=TOGETHER_API_KEY) | |
| # # # -------------------# | |
| # # # Streamlit UI setup | |
| # # # -------------------# | |
| # # st.set_page_config(page_title="FutureScope: Research Direction Explorer", layout="wide") | |
| # # st.markdown(""" | |
| # # <style> | |
| # # body { | |
| # # background: linear-gradient(135deg, #0f2027, #203a43, #2c5364); | |
| # # color: #FFFFFF; | |
| # # } | |
| # # h1, h2, h3 { | |
| # # text-align: center; | |
| # # color: #FFD700; | |
| # # font-family: 'Poppins', sans-serif; | |
| # # } | |
| # # .footer { | |
| # # position: fixed; | |
| # # left: 0; | |
| # # bottom: 0; | |
| # # width: 100%; | |
| # # color: white; | |
| # # text-align: center; | |
| # # padding: 10px; | |
| # # background-color: rgba(0,0,0,0.4); | |
| # # } | |
| # # .stButton > button { | |
| # # background-color: #FFD700 !important; | |
| # # color: black !important; | |
| # # font-weight: bold; | |
| # # border-radius: 10px; | |
| # # } | |
| # # </style> | |
| # # """, unsafe_allow_html=True) | |
| # # # -------------------# | |
| # # # App Title | |
| # # # -------------------# | |
| # # st.markdown("<h1>🧭 FutureScope: Research Direction Explorer</h1>", unsafe_allow_html=True) | |
| # # st.markdown("<p style='text-align:center;'>Discover how your research area evolved and where it's heading next 🚀</p>", unsafe_allow_html=True) | |
| # # # -------------------# | |
| # # # User Input | |
| # # # -------------------# | |
| # # user_topic = st.text_input("🔍 Enter your research topic", placeholder="e.g. Graph Neural Networks for Drug Discovery") | |
| # # # -------------------# | |
| # # # Main Logic | |
| # # # -------------------# | |
| # # if st.button("Generate Research Insights"): | |
| # # if not user_topic.strip(): | |
| # # st.warning("⚠️ Please enter a valid research topic.") | |
| # # else: | |
| # # with st.spinner("Analyzing topic evolution and forecasting future directions... ⏳"): | |
| # # # Prompt Design | |
| # # prompt = f""" | |
| # # You are a world-class AI research assistant specialized in analyzing research trends. | |
| # # Given the topic: "{user_topic}", perform the following: | |
| # # 1. Summarize how this research area has evolved in the past 10–15 years. | |
| # # 2. Identify key milestones and subfields in a timeline format. | |
| # # 3. Predict 3–5 future research directions and explain why each matters. | |
| # # Return the output strictly in JSON format like this: | |
| # # {{ | |
| # # "evolution_summary": "...", | |
| # # "timeline": [{{"year": ..., "trend": "..."}}, ...], | |
| # # "future_directions": [{{"title": "...", "reason": "..."}}, ...] | |
| # # }} | |
| # # """ | |
| # # # Call Together API | |
| # # response = client.chat.completions.create( | |
| # # model="meta-llama/Llama-3.3-70B-Instruct-Turbo-Free", | |
| # # messages=[{"role": "user", "content": prompt}] | |
| # # ) | |
| # # raw_content = response.choices[0].message.content | |
| # # # -------------------# | |
| # # # JSON Cleaning & Parsing | |
| # # # -------------------# | |
| # # def extract_json(text): | |
| # # """Extract valid JSON portion from the model response.""" | |
| # # text = text.strip() | |
| # # text = re.sub(r"^```json|```$", "", text).strip() # remove code fences | |
| # # match = re.search(r'\{.*\}', text, re.DOTALL) | |
| # # if match: | |
| # # return match.group(0) | |
| # # return text | |
| # # cleaned = extract_json(raw_content) | |
| # # try: | |
| # # data = json.loads(cleaned) | |
| # # except Exception as e: | |
| # # st.error(f"⚠️ Failed to parse JSON: {e}") | |
| # # st.text_area("Raw Response", raw_content, height=300) | |
| # # st.stop() | |
| # # # -------------------# | |
| # # # Display Results | |
| # # # -------------------# | |
| # # st.markdown("## 🧩 Evolution Summary") | |
| # # st.markdown(f"<div style='background:#1e2a38;padding:15px;border-radius:10px;'>{data['evolution_summary']}</div>", unsafe_allow_html=True) | |
| # # # Timeline Chart | |
| # # if "timeline" in data and len(data["timeline"]) > 0: | |
| # # df = pd.DataFrame(data["timeline"]) | |
| # # if "year" in df.columns and "trend" in df.columns: | |
| # # fig = px.scatter(df, x="year", y="trend", title="📈 Topic Evolution Over Time", | |
| # # size=[10]*len(df), text="trend", color_discrete_sequence=["gold"]) | |
| # # fig.update_traces(textposition='top center', marker=dict(symbol="circle")) | |
| # # fig.update_layout(template="plotly_dark", height=500) | |
| # # st.plotly_chart(fig, use_container_width=True) | |
| # # else: | |
| # # st.warning("Timeline data invalid — showing raw table:") | |
| # # st.dataframe(df) | |
| # # # Future Directions | |
| # # st.markdown("## 🔮 Predicted Future Directions") | |
| # # for item in data.get("future_directions", []): | |
| # # st.markdown(f""" | |
| # # <div style='background:#142733;padding:15px;margin:10px;border-radius:10px;'> | |
| # # <h4>🧠 {item['title']}</h4> | |
| # # <p>{item['reason']}</p> | |
| # # </div> | |
| # # """, unsafe_allow_html=True) | |
| # # # Tools: Copy / Download | |
| # # col1, col2 = st.columns(2) | |
| # # with col1: | |
| # # if st.button("📋 Copy Insights"): | |
| # # st.write("Copied to clipboard! (Use Ctrl+C manually to copy)") | |
| # # with col2: | |
| # # st.download_button( | |
| # # label="💾 Download JSON", | |
| # # data=json.dumps(data, indent=2), | |
| # # file_name=f"{user_topic.replace(' ','_')}_future_directions.json", | |
| # # mime="application/json" | |
| # # ) | |
| # # # -------------------# | |
| # # # Footer | |
| # # # -------------------# | |
| # # st.markdown("<div class='footer'>© Group 6 ILP TCS Research ", unsafe_allow_html=True) | |
| import os | |
| import json | |
| import streamlit as st | |
| import pandas as pd | |
| import plotly.express as px | |
| from together import Together | |
| from dotenv import load_dotenv | |
| import re | |
| # -------------------# | |
| # Secure API key load | |
| # -------------------# | |
| load_dotenv() | |
| TOGETHER_API_KEY = os.getenv("TOGETHER_API_KEY", "987adcf573b9658c775b671270aef959b3d38793771932f372f9f2a9ed5b78bf") | |
| client = Together(api_key=TOGETHER_API_KEY) | |
| # -------------------# | |
| # Streamlit UI setup | |
| # -------------------# | |
| st.set_page_config(page_title="PreSearch : Research Direction Explorer", layout="wide") | |
| # Detect Streamlit theme | |
| try: | |
| theme = st.get_option("theme.base") | |
| except: | |
| theme = "dark" | |
| # Define theme colors | |
| if theme == "light": | |
| BACKGROUND_GRADIENT = "linear-gradient(135deg, #f9f9f9, #eaeaea, #dddddd)" | |
| TEXT_COLOR = "#000000" | |
| TITLE_COLOR = "#DAA520" | |
| CARD_BG = "#ffffff" | |
| FOOTER_BG = "rgba(0, 0, 0, 0.1)" | |
| else: | |
| BACKGROUND_GRADIENT = "linear-gradient(135deg, #0f2027, #203a43, #2c5364)" | |
| TEXT_COLOR = "#FFFFFF" | |
| TITLE_COLOR = "#FFD700" | |
| CARD_BG = "#1e2a38" | |
| FOOTER_BG = "rgba(0, 0, 0, 0.4)" | |
| # -------------------# | |
| # Dynamic CSS Styling | |
| # -------------------# | |
| st.markdown(f""" | |
| <style> | |
| body {{ | |
| background: {BACKGROUND_GRADIENT}; | |
| color: {TEXT_COLOR}; | |
| font-family: 'Poppins', sans-serif; | |
| }} | |
| h1, h2, h3 {{ | |
| text-align: center; | |
| color: {TITLE_COLOR}; | |
| }} | |
| .footer {{ | |
| position: fixed; | |
| left: 0; | |
| bottom: 0; | |
| width: 100%; | |
| color: {TEXT_COLOR}; | |
| text-align: center; | |
| padding: 10px; | |
| background-color: {FOOTER_BG}; | |
| }} | |
| .stButton > button {{ | |
| background-color: {TITLE_COLOR} !important; | |
| color: black !important; | |
| font-weight: bold; | |
| border-radius: 10px; | |
| }} | |
| div[data-testid="stMarkdownContainer"] p {{ | |
| color: {TEXT_COLOR}; | |
| }} | |
| </style> | |
| """, unsafe_allow_html=True) | |
| # -------------------# | |
| # App Title | |
| # -------------------# | |
| st.markdown("<h1>🧭 PreSearch: Research Direction Explorer</h1>", unsafe_allow_html=True) | |
| st.markdown("<p style='text-align:center;'>Discover how your research area evolved and where it's heading next 🚀</p>", unsafe_allow_html=True) | |
| # -------------------# | |
| # User Input | |
| # -------------------# | |
| user_topic = st.text_input("🔍 Enter your research topic", placeholder="e.g. Graph Neural Networks for Drug Discovery") | |
| # -------------------# | |
| # Main Logic | |
| # -------------------# | |
| if st.button("Generate Research Insights"): | |
| if not user_topic.strip(): | |
| st.warning("⚠️ Please enter a valid research topic.") | |
| else: | |
| with st.spinner("Analyzing topic evolution and forecasting future directions... ⏳"): | |
| # Prompt Design | |
| prompt = f""" | |
| You are a world-class AI research assistant specialized in analyzing research trends. | |
| Given the topic: "{user_topic}", perform the following: | |
| 1. Summarize how this research area has evolved in the past 10–15 years. | |
| 2. Identify key milestones and subfields in a timeline format. | |
| 3. Predict 3–5 future research directions and explain why each matters. | |
| Return the output strictly in JSON format like this: | |
| {{ | |
| "evolution_summary": "...", | |
| "timeline": [{{"year": ..., "trend": "..."}}, ...], | |
| "future_directions": [{{"title": "...", "reason": "..."}}, ...] | |
| }} | |
| """ | |
| # Call Together API | |
| response = client.chat.completions.create( | |
| model="meta-llama/Llama-3.3-70B-Instruct-Turbo-Free", | |
| messages=[{"role": "user", "content": prompt}] | |
| ) | |
| raw_content = response.choices[0].message.content | |
| # -------------------# | |
| # JSON Cleaning & Parsing | |
| # -------------------# | |
| def extract_json(text): | |
| text = text.strip() | |
| text = re.sub(r"^```json|```$", "", text).strip() | |
| match = re.search(r'\{.*\}', text, re.DOTALL) | |
| return match.group(0) if match else text | |
| cleaned = extract_json(raw_content) | |
| try: | |
| data = json.loads(cleaned) | |
| except Exception as e: | |
| st.error(f"⚠️ Failed to parse JSON: {e}") | |
| st.text_area("Raw Response", raw_content, height=300) | |
| st.stop() | |
| # -------------------# | |
| # Display Results | |
| # -------------------# | |
| st.markdown("## 🧩 Evolution Summary") | |
| st.markdown(f"<div style='background:{CARD_BG};padding:15px;border-radius:10px;color:{TEXT_COLOR};'>{data['evolution_summary']}</div>", unsafe_allow_html=True) | |
| # Timeline Chart | |
| if "timeline" in data and len(data["timeline"]) > 0: | |
| df = pd.DataFrame(data["timeline"]) | |
| if "year" in df.columns and "trend" in df.columns: | |
| fig = px.scatter(df, x="year", y="trend", title="📈 Topic Evolution Over Time", | |
| size=[10]*len(df), text="trend", color_discrete_sequence=["gold"]) | |
| fig.update_traces(textposition='top center', marker=dict(symbol="circle")) | |
| fig.update_layout(template="plotly_dark" if theme == "dark" else "plotly_white", height=500) | |
| st.plotly_chart(fig, use_container_width=True) | |
| else: | |
| st.warning("Timeline data invalid — showing raw table:") | |
| st.dataframe(df) | |
| # Future Directions | |
| st.markdown("## 🔮 Predicted Future Directions") | |
| for item in data.get("future_directions", []): | |
| st.markdown(f""" | |
| <div style='background:{CARD_BG};padding:15px;margin:10px;border-radius:10px;color:{TEXT_COLOR};'> | |
| <h4>🧠 {item['title']}</h4> | |
| <p>{item['reason']}</p> | |
| </div> | |
| """, unsafe_allow_html=True) | |
| # Tools: Copy / Download | |
| col1, col2 = st.columns(2) | |
| with col1: | |
| if st.button("📋 Copy Insights"): | |
| st.write("Copied to clipboard! (Use Ctrl+C manually to copy)") | |
| with col2: | |
| st.download_button( | |
| label="💾 Download JSON", | |
| data=json.dumps(data, indent=2), | |
| file_name=f"{user_topic.replace(' ','_')}_future_directions.json", | |
| mime="application/json" | |
| ) | |
| # -------------------# | |
| # Footer | |
| # -------------------# | |
| st.markdown(f"<div class='footer'>© Group 6 ILP TCS Research</div>", unsafe_allow_html=True) | |
| # import os, json, time, re, requests, random | |
| # import pandas as pd | |
| # import streamlit as st | |
| # from together import Together | |
| # # ========================= | |
| # # 0️⃣ Configuration & Setup | |
| # # ========================= | |
| # st.set_page_config(page_title="📚 pResearch Retrieval", layout="wide", page_icon=":books:") | |
| # st.title("🤖 **pResearch: Multi-Agent Research Retrieval System**") | |
| # st.caption("Built with LLM-based reasoning, multi-agent intelligence, and human-in-loop control.") | |
| # st.markdown("---") | |
| # TOGETHER_API_KEY = os.getenv("TOGETHER_API_KEY", "987adcf573b9658c775b671270aef959b3d38793771932f372f9f2a9ed5b78bf") | |
| # SEMANTIC_API_KEY = os.getenv("SEMANTIC_API_KEY", "b2EsaPVVN1890PXdCeum37K9zKq4AYY46n8QyLvp") | |
| # client = Together(api_key=TOGETHER_API_KEY) | |
| # # ========================= | |
| # # Unified LLM Call | |
| # # ========================= | |
| # @st.cache_data(show_spinner=False) | |
| # def llm_call(prompt: str, temperature=0.2, max_retries=3): | |
| # for attempt in range(max_retries): | |
| # try: | |
| # resp = client.chat.completions.create( | |
| # model="meta-llama/Llama-3.3-70B-Instruct-Turbo", | |
| # messages=[{"role": "user", "content": prompt}], | |
| # temperature=temperature | |
| # ) | |
| # return resp.choices[0].message.content.strip() | |
| # except Exception as e: | |
| # time.sleep(1 + attempt) | |
| # return "LLM error (see logs)" | |
| # # ============================================================ | |
| # # 1️⃣ Query Reformulator Agent | |
| # # ============================================================ | |
| # def agent_query_reformulator(query: str): | |
| # prompt = f""" | |
| # You are an expert academic assistant. | |
| # Reformulate the query below into 5 semantically diverse and rich alternatives | |
| # that explore different perspectives (methods, datasets, applications, etc.) | |
| # Query: "{query}" | |
| # Respond in JSON format: | |
| # {{ | |
| # "reformulated_queries": [ | |
| # {{ "id": 1, "query": "..." }}, | |
| # {{ "id": 2, "query": "..." }}, | |
| # {{ "id": 3, "query": "..." }}, | |
| # {{ "id": 4, "query": "..." }}, | |
| # {{ "id": 5, "query": "..." }} | |
| # ] | |
| # }} | |
| # """ | |
| # output = llm_call(prompt) | |
| # cleaned = re.sub(r"```json|```", "", output).strip() | |
| # try: | |
| # data = json.loads(cleaned) | |
| # queries = [q["query"] for q in data.get("reformulated_queries", []) if "query" in q] | |
| # except Exception: | |
| # queries = [] | |
| # # fallback diversity | |
| # while len(queries) < 5: | |
| # alt = llm_call(f"Generate a diverse reformulation of: {query}") | |
| # queries.append(alt[:300]) | |
| # queries = list(dict.fromkeys(queries))[:5] | |
| # return {"original_query": query, "reformulated_queries": [{"id": i+1, "query": q} for i, q in enumerate(queries)]} | |
| # # ============================================================ | |
| # # 2️⃣ Retriever Agent (Semantic Scholar) | |
| # # ============================================================ | |
| # def agent_retriever(query, top_k=20): | |
| # url = "https://api.semanticscholar.org/graph/v1/paper/search" | |
| # headers = {"x-api-key": SEMANTIC_API_KEY} | |
| # params = { | |
| # "query": query, "limit": top_k, | |
| # "fields": "paperId,title,abstract,year,authors,url,venue,citationCount" | |
| # } | |
| # resp = requests.get(url, headers=headers, params=params) | |
| # if resp.status_code != 200: | |
| # return [] | |
| # return resp.json().get("data", []) | |
| # # ============================================================ | |
| # # 3️⃣ Reranker Agent | |
| # # ============================================================ | |
| # def agent_reranker(query, papers): | |
| # cleaned_papers = [p for p in papers if p.get("abstract")] | |
| # random.shuffle(cleaned_papers) | |
| # for batch_start in range(0, len(cleaned_papers), 5): | |
| # batch = cleaned_papers[batch_start:batch_start+5] | |
| # papers_str = "\n\n".join([ | |
| # f"[{i+1}] Title: {p.get('title','N/A')}\nAbstract: {p.get('abstract','')[:500]}" | |
| # for i,p in enumerate(batch) | |
| # ]) | |
| # prompt = f""" | |
| # You are a relevance scoring agent. | |
| # Given a research query and 5 papers, assign a score (0–1) to each paper for its relevance. | |
| # Query: {query} | |
| # Papers: | |
| # {papers_str} | |
| # Respond strictly in JSON: | |
| # {{ "results": [{{"id": 1, "score": 0.85}}, ...] }} | |
| # """ | |
| # output = llm_call(prompt) | |
| # cleaned = re.sub(r"```json|```", "", output).strip() | |
| # try: | |
| # results = json.loads(cleaned)["results"] | |
| # for i,r in enumerate(results): | |
| # cleaned_papers[batch_start+i]["semantic_score"] = r.get("score",0) | |
| # except: | |
| # for i in range(len(batch)): | |
| # cleaned_papers[batch_start+i]["semantic_score"] = 0.0 | |
| # return sorted(cleaned_papers, key=lambda x: x.get("semantic_score",0), reverse=True) | |
| # # ============================================================ | |
| # # 4️⃣ Weighting Agent (Meta Scorer) | |
| # # ============================================================ | |
| # def agent_weighting(papers): | |
| # prompt = """ | |
| # You are an expert in bibliometrics. | |
| # Assign importance weights (sum=1.0) for how papers should be ranked based on: | |
| # - semantic_score (LLM relevance) | |
| # - citationCount | |
| # - recency | |
| # - venue quality | |
| # Return JSON: | |
| # {"weights":{"semantic_score":0.55,"citations":0.25,"recency":0.15,"venue":0.05}} | |
| # """ | |
| # output = llm_call(prompt) | |
| # cleaned = re.sub(r"```json|```", "", output).strip() | |
| # try: | |
| # weights = json.loads(cleaned)["weights"] | |
| # except: | |
| # weights = {"semantic_score":0.55,"citations":0.25,"recency":0.15,"venue":0.05} | |
| # total = sum(weights.values()) | |
| # return {k:v/total for k,v in weights.items()} | |
| # # ============================================================ | |
| # # 5️⃣ Meta-Scoring and Ranking | |
| # # ============================================================ | |
| # def agent_meta_scorer(papers, weights): | |
| # current_year = 2025 | |
| # prestige = {"CVPR":1.0,"ICCV":0.95,"ECCV":0.9,"NEURIPS":0.9,"ICML":0.85,"AAAI":0.8,"IJCAI":0.8,"ARXIV":0.4} | |
| # for p in papers: | |
| # sem = p.get("semantic_score",0) | |
| # cit = min(p.get("citationCount",0)/1000,1.0) | |
| # rec = max(0, 1 - (current_year - p.get("year",2000))/10) | |
| # venue_name = (p.get("venue") or "").upper() | |
| # ven = next((v for k,v in prestige.items() if k in venue_name), 0.3) | |
| # p["final_score"] = ( | |
| # weights["semantic_score"]*sem + | |
| # weights["citations"]*cit + | |
| # weights["recency"]*rec + | |
| # weights["venue"]*ven | |
| # ) | |
| # return sorted(papers, key=lambda x: x["final_score"], reverse=True) | |
| # # ============================================================ | |
| # # 6️⃣ Critique Agent | |
| # # ============================================================ | |
| # def agent_critique(papers, query): | |
| # top_titles = [p["title"] for p in papers[:5]] | |
| # prompt = f""" | |
| # As a research critic, evaluate whether these top papers are relevant to: | |
| # "{query}" | |
| # Papers: {json.dumps(top_titles, indent=2)} | |
| # Respond as JSON: | |
| # {{ "critique": "...", "relevance_score": 0–1 }} | |
| # """ | |
| # output = llm_call(prompt) | |
| # cleaned = re.sub(r"```json|```", "", output).strip() | |
| # try: | |
| # return json.loads(cleaned) | |
| # except: | |
| # return {"critique":"Automatic check fallback.","relevance_score":0.7} | |
| # # ============================================================ | |
| # # 7️⃣ Human-in-Loop Fallback | |
| # # ============================================================ | |
| # def human_feedback_loop(papers): | |
| # st.warning("⚠️ Low relevance detected — human feedback required.") | |
| # for i,p in enumerate(papers[:3]): | |
| # st.markdown(f"**{i+1}. {p['title']}** *(Score: {p['final_score']:.3f})*") | |
| # st.caption(f"{p.get('abstract','')[:250]}...") | |
| # choice = st.radio("Approve ranking?", ["Yes","No"], index=0) | |
| # if choice == "No": | |
| # st.info("🔄 Re-ranking by citation count.") | |
| # papers = sorted(papers, key=lambda x: x.get("citationCount",0), reverse=True) | |
| # return papers | |
| # # ============================================================ | |
| # # 8️⃣ Streamlit Master Orchestrator | |
| # # ============================================================ | |
| # def run_pipeline(query, top_k=10): | |
| # st.markdown("## 🧩 Stage 1: Query Reformulation") | |
| # with st.spinner("Generating diverse reformulations..."): | |
| # q_data = agent_query_reformulator(query) | |
| # queries = [query] + [q["query"] for q in q_data["reformulated_queries"]] | |
| # for q in queries[1:]: | |
| # st.markdown(f"🔹 *{q}*") | |
| # st.markdown("## 🔍 Stage 2: Retrieval") | |
| # all_papers = [] | |
| # progress = st.progress(0) | |
| # for i, q in enumerate(queries): | |
| # new_papers = agent_retriever(q, top_k) | |
| # all_papers.extend(new_papers) | |
| # progress.progress((i+1)/len(queries)) | |
| # time.sleep(0.3) | |
| # progress.empty() | |
| # st.success(f"✅ Retrieved {len(all_papers)} papers in total.") | |
| # st.markdown("## 🧠 Stage 3: Semantic Reranking") | |
| # with st.spinner("Reranking papers semantically..."): | |
| # reranked = agent_reranker(query, all_papers) | |
| # st.info(f"Top paper after rerank: **{reranked[0]['title']}**") | |
| # st.markdown("## ⚖️ Stage 4: Weighting Agent & Meta-Scoring") | |
| # weights = agent_weighting(reranked) | |
| # st.json(weights) | |
| # meta_ranked = agent_meta_scorer(reranked, weights) | |
| # st.markdown("## 🔎 Stage 5: Critique Agent") | |
| # critique = agent_critique(meta_ranked, query) | |
| # st.info(f"**Critique:** {critique['critique']} | Relevance: {critique['relevance_score']:.2f}") | |
| # if critique["relevance_score"] < 0.6: | |
| # meta_ranked = human_feedback_loop(meta_ranked) | |
| # df = pd.DataFrame(meta_ranked) | |
| # st.download_button( | |
| # label="💾 Download Results as CSV", | |
| # data=df.to_csv(index=False), | |
| # file_name="final_ranked_papers.csv", | |
| # mime="text/csv" | |
| # ) | |
| # st.dataframe(df.head(20)) | |
| # st.success("🎯 Pipeline completed successfully!") | |
| # # ============================================================ | |
| # # 9️⃣ Streamlit UI | |
| # # ============================================================ | |
| # with st.form("research_form"): | |
| # query = st.text_input("Enter your research query:", "spatio-temporal action detection and localization") | |
| # top_k = st.slider("Number of papers per reformulation", 5, 50, 10) | |
| # run = st.form_submit_button("🚀 Run Multi-Agent Retrieval") | |
| # if run: | |
| # run_pipeline(query, top_k) | |