Spaces:
Sleeping
Sleeping
File size: 19,916 Bytes
e2ef653 622a72f e2ef653 622a72f | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 | import streamlit as st
import chromadb
from sentence_transformers import SentenceTransformer
import fitz # PyMuPDF
import os
import requests
import re
import hashlib
# βββ Page Config ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
st.set_page_config(
page_title="PDF RAG Β· Upload & Ask",
page_icon="π",
layout="wide",
initial_sidebar_state="expanded"
)
# βββ CSS ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
st.markdown("""
<style>
@import url('https://fonts.googleapis.com/css2?family=IBM+Plex+Sans:wght@300;400;500;600&family=IBM+Plex+Mono:wght@400;500&display=swap');
html, body, [class*="css"] { font-family: 'IBM Plex Sans', sans-serif; }
.main { background-color: #0b0f1a; }
.hero {
background: linear-gradient(160deg, #0d1424 0%, #0b0f1a 100%);
border: 1px solid #1e2a3e;
border-top: 3px solid #22d3ee;
border-radius: 12px;
padding: 28px 32px;
margin-bottom: 24px;
}
.hero h1 { font-size: 1.8rem; font-weight: 600; color: #e2e8f0; margin: 0 0 6px 0; }
.hero p { color: #64748b; font-size: 0.95rem; margin: 0; }
.phase-bar {
display: flex; gap: 0; margin-bottom: 28px;
border: 1px solid #1e2a3e; border-radius: 10px; overflow: hidden;
}
.phase {
flex: 1; padding: 10px 6px; text-align: center;
font-size: 0.75rem; color: #4b5563; background: #0d1117;
border-right: 1px solid #1e2a3e; line-height: 1.5;
}
.phase:last-child { border-right: none; }
.phase.done { color: #22d3ee; background: rgba(34,211,238,0.05); }
.phase.active { color: #f8fafc; background: rgba(34,211,238,0.1); font-weight: 600; }
.phase-icon { font-size: 1.1rem; display: block; margin-bottom: 2px; }
.pdf-card {
background: #0d1424;
border: 1px solid #1e2a3e;
border-radius: 10px;
padding: 14px 16px;
margin: 8px 0;
display: flex;
align-items: center;
justify-content: space-between;
}
.pdf-name { font-size: 0.85rem; color: #e2e8f0; font-weight: 500; }
.pdf-meta { font-family: 'IBM Plex Mono', monospace; font-size: 0.72rem; color: #475569; margin-top: 3px; }
.pdf-badge {
font-size: 0.72rem; font-family: 'IBM Plex Mono', monospace;
background: rgba(34,211,238,0.1); color: #22d3ee;
border: 1px solid rgba(34,211,238,0.25); padding: 3px 10px; border-radius: 20px;
}
.answer-box {
background: #0d1424;
border: 1px solid #1e3a4a;
border-left: 3px solid #22d3ee;
border-radius: 10px;
padding: 22px 24px;
color: #e2e8f0;
line-height: 1.75;
font-size: 0.96rem;
margin: 12px 0 20px 0;
}
.chunk-card {
background: #0d1117;
border: 1px solid #1e2a3e;
border-radius: 9px;
padding: 14px 18px;
margin: 8px 0;
}
.chunk-top {
display: flex; justify-content: space-between;
align-items: center; margin-bottom: 8px;
}
.chunk-source { font-size: 0.77rem; font-weight: 600; color: #22d3ee; text-transform: uppercase; letter-spacing: 0.05em; }
.chunk-page { font-family: 'IBM Plex Mono', monospace; font-size: 0.72rem; color: #475569; }
.score-bar-wrap { display: flex; align-items: center; gap: 8px; }
.score-bar {
height: 4px; border-radius: 2px; background: #1e2a3e; width: 80px; overflow: hidden;
}
.score-fill { height: 100%; border-radius: 2px; background: #22d3ee; }
.score-num { font-family: 'IBM Plex Mono', monospace; font-size: 0.72rem; color: #22d3ee; }
.chunk-text { font-size: 0.86rem; color: #94a3b8; line-height: 1.65; }
.stat-row { display: flex; gap: 10px; margin: 16px 0; }
.stat-box {
flex: 1; background: #0d1424; border: 1px solid #1e2a3e;
border-radius: 8px; padding: 12px; text-align: center;
}
.stat-val { font-size: 1.35rem; font-weight: 600; color: #22d3ee; }
.stat-lbl { font-size: 0.7rem; color: #475569; margin-top: 2px; }
.section-label {
font-size: 0.7rem; text-transform: uppercase; letter-spacing: 0.1em;
color: #374151; font-weight: 600; margin: 18px 0 8px 0;
}
section[data-testid="stSidebar"] {
background-color: #080c14; border-right: 1px solid #131c2e;
}
.empty-state {
text-align: center; padding: 48px 24px;
border: 2px dashed #1e2a3e; border-radius: 12px; color: #374151;
}
.empty-state .icon { font-size: 2.5rem; margin-bottom: 12px; }
.empty-state p { font-size: 0.9rem; line-height: 1.6; }
</style>
""", unsafe_allow_html=True)
# βββ Session State ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
if "indexed_files" not in st.session_state:
st.session_state.indexed_files = {} # filename β {chunks, pages, size}
if "chroma_collection" not in st.session_state:
st.session_state.chroma_collection = None
if "chroma_client" not in st.session_state:
st.session_state.chroma_client = None
if "total_chunks" not in st.session_state:
st.session_state.total_chunks = 0
# βββ Load embedding model (cached globally) βββββββββββββββββββββββββββββββββββ
@st.cache_resource(show_spinner=False)
def load_embed_model():
return SentenceTransformer('all-MiniLM-L6-v2')
# βββ PDF Extraction βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def extract_text_from_pdf(pdf_bytes: bytes) -> list[dict]:
"""Returns list of {page, text} dicts."""
doc = fitz.open(stream=pdf_bytes, filetype="pdf")
pages = []
for page_num, page in enumerate(doc, start=1):
text = page.get_text("text").strip()
if text:
pages.append({"page": page_num, "text": text})
doc.close()
return pages
# βββ Chunking βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def chunk_text(pages: list[dict], chunk_size: int = 400, overlap: int = 60) -> list[dict]:
"""Splits page text into overlapping word-based chunks."""
chunks = []
for p in pages:
words = p["text"].split()
start = 0
while start < len(words):
end = start + chunk_size
chunk_words = words[start:end]
chunk_text_str = " ".join(chunk_words).strip()
if len(chunk_text_str) > 60:
chunks.append({"text": chunk_text_str, "page": p["page"]})
start += chunk_size - overlap
return chunks
# βββ Index PDF into ChromaDB ββββββββββββββββββββββββββββββββββββββββββββββββββ
def index_pdf(filename: str, pdf_bytes: bytes, embed_model):
# Init or reuse ChromaDB
if st.session_state.chroma_client is None:
st.session_state.chroma_client = chromadb.Client()
st.session_state.chroma_collection = st.session_state.chroma_client.get_or_create_collection(
name="pdf_rag", metadata={"hnsw:space": "cosine"}
)
collection = st.session_state.chroma_collection
# Extract & chunk
pages = extract_text_from_pdf(pdf_bytes)
chunks = chunk_text(pages)
if not chunks:
return 0, 0
# Embed & add
texts = [c["text"] for c in chunks]
embeddings = embed_model.encode(texts, batch_size=32, show_progress_bar=False).tolist()
ids, docs, metas, embeds = [], [], [], []
for i, (chunk, emb) in enumerate(zip(chunks, embeddings)):
chunk_id = f"{hashlib.md5(filename.encode()).hexdigest()[:8]}_chunk_{i}"
ids.append(chunk_id)
docs.append(chunk["text"])
metas.append({"filename": filename, "page": chunk["page"]})
embeds.append(emb)
collection.add(ids=ids, embeddings=embeds, documents=docs, metadatas=metas)
st.session_state.indexed_files[filename] = {
"chunks": len(chunks),
"pages": len(pages),
"size_kb": round(len(pdf_bytes) / 1024, 1)
}
st.session_state.total_chunks += len(chunks)
return len(chunks), len(pages)
# βββ RAG Query ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
def rag_query(question: str, embed_model, top_k: int, api_key: str):
collection = st.session_state.chroma_collection
q_emb = embed_model.encode(question).tolist()
results = collection.query(query_embeddings=[q_emb], n_results=top_k)
chunks = []
for i in range(len(results["documents"][0])):
distance = results["distances"][0][i]
chunks.append({
"text": results["documents"][0][i],
"filename": results["metadatas"][0][i]["filename"],
"page": results["metadatas"][0][i]["page"],
"relevance": round((1 - distance) * 100, 1),
})
context = "\n\n".join([
f"[Source: {c['filename']}, Page {c['page']}]\n{c['text']}" for c in chunks
])
prompt = f"""You are a helpful assistant. Answer the user's question using ONLY the document context provided below. Be concise and clear. Always mention the source filename and page number when referencing specific information. If the answer cannot be found in the provided context, say "I couldn't find that information in the uploaded documents."
Document Context:
{context}
Question: {question}
Answer:"""
headers = {"Authorization": f"Bearer {api_key}", "Content-Type": "application/json"}
payload = {
"model": "llama-3.3-70b-versatile",
"messages": [{"role": "user", "content": prompt}],
"max_tokens": 600,
"temperature": 0.2,
}
r = requests.post("https://api.groq.com/openai/v1/chat/completions", headers=headers, json=payload, timeout=30)
r.raise_for_status()
answer = r.json()["choices"][0]["message"]["content"]
return answer, chunks
# βββ Determine current phase ββββββββββββββββββββββββββββββββββββββββββββββββββ
has_docs = len(st.session_state.indexed_files) > 0
phase = 1 if not has_docs else 2
# βββ Sidebar ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
with st.sidebar:
st.markdown("## π PDF RAG Demo")
st.markdown("<div style='color:#374151;font-size:0.8rem'>Upload β Extract β Index β Ask</div>", unsafe_allow_html=True)
st.markdown("---")
env_key = os.environ.get("GROQ_API_KEY", "")
if env_key:
api_key = env_key
st.success("β
Groq key loaded from secrets")
else:
api_key = st.text_input("π Groq API Key", type="password", placeholder="gsk_...", help="Free at console.groq.com")
if not api_key:
st.caption("Get free key β [console.groq.com](https://console.groq.com)")
st.markdown("---")
st.markdown("<div class='section-label'>Indexed Documents</div>", unsafe_allow_html=True)
if st.session_state.indexed_files:
for fname, info in st.session_state.indexed_files.items():
st.markdown(f"""
<div style='padding:6px 0;border-bottom:1px solid #131c2e'>
<div style='font-size:0.8rem;color:#e2e8f0'>π {fname}</div>
<div style='font-size:0.72rem;color:#475569;font-family:IBM Plex Mono,monospace'>
{info["pages"]}p Β· {info["chunks"]} chunks Β· {info["size_kb"]}KB
</div>
</div>""", unsafe_allow_html=True)
st.markdown("---")
if st.button("ποΈ Clear all & reset", use_container_width=True):
for key in ["indexed_files", "chroma_collection", "chroma_client", "total_chunks"]:
del st.session_state[key]
st.rerun()
else:
st.markdown("<div style='color:#374151;font-size:0.82rem'>No documents indexed yet.</div>", unsafe_allow_html=True)
st.markdown("---")
st.markdown("""
<div style='font-size:0.77rem;color:#374151;line-height:1.9'>
<b style='color:#4b5563'>Stack</b><br>
π PDF parsing: PyMuPDF<br>
βοΈ Chunking: word-overlap (400w)<br>
π’ Embeddings: all-MiniLM-L6-v2<br>
ποΈ Vector DB: ChromaDB in-memory<br>
π§ LLM: Groq Β· Llama 3.3 70B<br>
π Hosting: HuggingFace Spaces
</div>
""", unsafe_allow_html=True)
# βββ Hero βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
st.markdown("""
<div class='hero'>
<h1>π PDF RAG β Upload & Ask</h1>
<p>Upload any PDF documents Β· They get extracted, chunked, embedded, and indexed Β· Then ask questions across all of them</p>
</div>
""", unsafe_allow_html=True)
# Phase bar
st.markdown(f"""
<div class='phase-bar'>
<div class='phase {"done" if phase > 1 else "active"}'>
<span class='phase-icon'>π€</span>Upload PDFs
</div>
<div class='phase {"active" if phase == 1 else "done"}'>
<span class='phase-icon'>π</span>Extract Text
</div>
<div class='phase {"active" if phase == 1 else "done"}'>
<span class='phase-icon'>βοΈ</span>Chunk
</div>
<div class='phase {"active" if phase == 1 else "done"}'>
<span class='phase-icon'>π’</span>Embed
</div>
<div class='phase {"active" if phase == 1 else "done"}'>
<span class='phase-icon'>ποΈ</span>Index
</div>
<div class='phase {"active" if phase == 2 else ""}'>
<span class='phase-icon'>π¬</span>Ask Questions
</div>
</div>
""", unsafe_allow_html=True)
# βββ Load model βββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
with st.spinner("βοΈ Loading embedding model..."):
embed_model = load_embed_model()
# ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
# PHASE 1 β Upload & Index
# ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
st.markdown("<div class='section-label'>Step 1 β Upload PDF Documents</div>", unsafe_allow_html=True)
uploaded_files = st.file_uploader(
"Drop your PDF files here",
type=["pdf"],
accept_multiple_files=True,
label_visibility="collapsed"
)
if uploaded_files:
new_files = [f for f in uploaded_files if f.name not in st.session_state.indexed_files]
if new_files:
st.markdown(f"**{len(new_files)} new file(s) ready to index:**")
for f in new_files:
st.markdown(f"<div class='pdf-card'><div><div class='pdf-name'>π {f.name}</div><div class='pdf-meta'>{round(f.size/1024,1)} KB</div></div><div class='pdf-badge'>ready</div></div>", unsafe_allow_html=True)
if st.button(f"β‘ Extract & Index {len(new_files)} PDF(s)", type="primary", use_container_width=True):
progress = st.progress(0, text="Starting...")
for idx, f in enumerate(new_files):
progress.progress((idx) / len(new_files), text=f"Processing: {f.name}")
pdf_bytes = f.read()
with st.spinner(f"Extracting & indexing **{f.name}**..."):
n_chunks, n_pages = index_pdf(f.name, pdf_bytes, embed_model)
st.success(f"β
**{f.name}** β {n_pages} pages Β· {n_chunks} chunks indexed")
progress.progress(1.0, text="Done!")
st.balloons()
st.rerun()
else:
st.info("All uploaded files are already indexed. Upload new files or ask questions below.")
elif not has_docs:
st.markdown("""
<div class='empty-state'>
<div class='icon'>π</div>
<p><b style='color:#94a3b8'>No documents uploaded yet</b><br>
Upload one or more PDF files above to get started.<br>
Any topic works β reports, manuals, research papers, policies.</p>
</div>
""", unsafe_allow_html=True)
# ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
# PHASE 2 β Stats & Query
# ββββββββββββββββββββββββββββββββββββββββββββββββββββββββββββ
if has_docs:
total_pages = sum(v["pages"] for v in st.session_state.indexed_files.values())
st.markdown("<div class='section-label' style='margin-top:24px'>Index Summary</div>", unsafe_allow_html=True)
st.markdown(f"""
<div class='stat-row'>
<div class='stat-box'><div class='stat-val'>{len(st.session_state.indexed_files)}</div><div class='stat-lbl'>Documents</div></div>
<div class='stat-box'><div class='stat-val'>{total_pages}</div><div class='stat-lbl'>Pages Parsed</div></div>
<div class='stat-box'><div class='stat-val'>{st.session_state.total_chunks}</div><div class='stat-lbl'>Chunks Indexed</div></div>
<div class='stat-box'><div class='stat-val'>384</div><div class='stat-lbl'>Embedding Dims</div></div>
</div>
""", unsafe_allow_html=True)
if not api_key:
st.warning("π Enter your Groq API key in the sidebar to start asking questions.")
st.stop()
st.markdown("---")
st.markdown("<div class='section-label'>Step 2 β Ask a Question</div>", unsafe_allow_html=True)
col1, col2 = st.columns([5, 1])
with col1:
question = st.text_input("", placeholder="What does the document say about...?", label_visibility="collapsed")
with col2:
top_k = st.selectbox("Top K", [2, 3, 4, 5], index=1, help="Number of chunks to retrieve")
ask_btn = st.button("π Search & Answer", type="primary", use_container_width=True)
if ask_btn and question:
with st.spinner("π Searching index and generating answer..."):
try:
answer, chunks = rag_query(question, embed_model, top_k, api_key)
st.markdown(f"<div class='section-label'>Answer</div>", unsafe_allow_html=True)
st.markdown(f"<div class='answer-box'>{answer}</div>", unsafe_allow_html=True)
st.markdown("<div class='section-label'>Retrieved Chunks (context sent to LLM)</div>", unsafe_allow_html=True)
for i, chunk in enumerate(chunks):
bar_width = int(chunk['relevance'])
st.markdown(f"""
<div class='chunk-card'>
<div class='chunk-top'>
<div>
<div class='chunk-source'>π {chunk['filename']}</div>
<div class='chunk-page'>Page {chunk['page']}</div>
</div>
<div class='score-bar-wrap'>
<div class='score-bar'><div class='score-fill' style='width:{bar_width}%'></div></div>
<div class='score-num'>{chunk['relevance']}%</div>
</div>
</div>
<div class='chunk-text'>{chunk['text']}</div>
</div>
""", unsafe_allow_html=True)
except requests.HTTPError as e:
if e.response.status_code == 401:
st.error("β Invalid Groq API key.")
else:
st.error(f"β API error: {str(e)}")
except Exception as e:
st.error(f"β Error: {str(e)}")
elif ask_btn and not question:
st.warning("Please enter a question.") |