from huggingface_hub import list_repo_files, hf_hub_download, snapshot_download
from pathlib import Path
from huggingface_hub import list_repo_files, hf_hub_download
import streamlit as st
import os
from PIL import Image
import pandas as pd
import json
from datetime import datetime
from typing import List, Dict, Any, Optional
from RAG import EnhancedMultimodalRAGSystem
from config import *
# Page config
st.set_page_config(
page_title="DTMI UGM Academic Assistant",
page_icon="🎓",
layout="wide",
initial_sidebar_state="expanded"
)
st.markdown("""
""", unsafe_allow_html=True)
@st.cache_resource
def initialize_rag_system():
try:
return EnhancedMultimodalRAGSystem()
except Exception as e:
st.error(f"❌ Error initializing RAG system: {e}")
st.stop()
def display_example_queries():
"""Display clickable example queries"""
st.markdown("### 💡 Contoh Pertanyaan")
for category, queries in EXAMPLE_QUERIES.items():
with st.expander(f"{category}", expanded=True):
for query in queries:
if st.button(f"💬 {query}", key=f"example_{hash(query)}", use_container_width=True):
st.session_state.user_input = query
st.rerun()
def display_tables_in_chat(table_data: List[Dict]):
"""Display tables directly in chat"""
if not table_data:
return
st.markdown("### 📊 Tabel Data")
for i, table_info in enumerate(table_data, 1):
with st.expander(f"📊 {table_info['title']} (Hal. {table_info['page']}, {table_info['year']})", expanded=True):
# Table metadata
col1, col2, col3 = st.columns(3)
with col1:
st.metric("📄 Halaman", table_info['page'])
with col2:
st.metric("📅 Tahun", table_info['year'])
with col3:
st.metric("📊 Score", f"{table_info['score']:.3f}")
# Display table data
try:
if table_info.get("data_type") == "dataframe" and isinstance(table_info["data"], pd.DataFrame):
st.dataframe(table_info["data"], use_container_width=True)
# Download CSV
csv_data = table_info["data"].to_csv(index=False)
st.download_button(
label="💾 Download CSV",
data=csv_data,
file_name=f"table_{i}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.csv",
mime="text/csv"
)
elif table_info.get("data_type") == "json":
st.json(table_info["data"])
# Download JSON
json_str = json.dumps(table_info["data"], indent=2, ensure_ascii=False)
st.download_button(
label="💾 Download JSON",
data=json_str,
file_name=f"data_{i}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.json",
mime="application/json"
)
# Show description
if table_info.get('description'):
st.markdown("**📝 Deskripsi:**")
st.text(table_info['description'])
except Exception as e:
st.error(f"❌ Error displaying table: {e}")
def display_single_image_compact(img_info: Dict, index: int):
"""Display single image in compact format - CLEAN VERSION"""
try:
image_path = img_info["path"]
# Check if file exists
if not os.path.exists(image_path):
st.error(f"❌ Gambar {index} tidak ditemukan")
return
# Load and display image
image = Image.open(image_path)
# Display image with nice styling
st.image(image,
caption=f"📖 {img_info.get('title', 'Gambar')} - Hal. {img_info.get('page', 'N/A')} ({img_info.get('year', 'N/A')})",
use_container_width=True)
# Compact metadata
col1, col2 = st.columns(2)
with col1:
st.metric("📊 Relevance Score", f"None")
# {img_info.get('score', 0):.2f}")
with col2:
st.metric("📐 Ukuran", f"{image.width}×{image.height}px")
# Expandable details
with st.expander(f"📝 Detail Gambar {index}", expanded=False):
if img_info.get('description'):
st.markdown("**📄 Deskripsi:**")
st.text(img_info['description'])
if img_info.get('caption'):
st.markdown("**💬 Caption:**")
st.text(img_info['caption'])
except Exception as e:
st.error(f"❌ Error loading image {index}: {str(e)}")
def display_single_image_full(img_info: Dict):
"""Display single image in full format - CLEAN VERSION"""
try:
image_path = img_info["path"]
if not os.path.exists(image_path):
st.error("❌ Gambar tidak ditemukan")
return
# Load image
image = Image.open(image_path)
# Display with title
st.markdown(f"### 🖼️ {img_info.get('title', 'Gambar')}")
# Create columns for image and metadata
col1, col2 = st.columns([3, 1])
with col1:
st.image(image, use_column_width=True)
with col2:
st.markdown("**📋 Informasi Gambar**")
st.metric("📄 Halaman", img_info.get('page', 'N/A'))
st.metric("📅 Tahun", img_info.get('year', 'N/A'))
# st.metric("📊 Score", f"{img_info.get('score', 0):.3f}")
st.metric("📐 Dimensi", f"{image.width} × {image.height}")
# Download button
with open(image_path, "rb") as file:
st.download_button(
label="💾 Download Gambar",
data=file.read(),
file_name=os.path.basename(image_path),
mime="image/png",
use_container_width=True
)
# Description below image
if img_info.get('description'):
st.markdown("**📝 Deskripsi Gambar:**")
st.info(img_info['description'])
if img_info.get('caption'):
st.markdown("**💬 Caption:**")
st.info(img_info['caption'])
except Exception as e:
st.error(f"❌ Error loading image: {str(e)}")
def display_images_in_chat(image_data: List[Dict], show_details: bool = True):
"""Display images directly in chat - CLEAN VERSION"""
if not image_data:
return
st.markdown("### 🖼️ Gambar Terkait")
if len(image_data) == 1:
st.markdown(f"*Ditemukan 1 gambar relevan*")
else:
st.markdown(f"*Ditemukan {len(image_data)} gambar relevan*")
if len(image_data) > 1:
cols = st.columns(min(len(image_data), 2)) # Max 2 columns
for i, img_info in enumerate(image_data):
with cols[i % 2]:
display_single_image_compact(img_info, i+1)
else:
display_single_image_full(image_data[0])
def enhanced_chat_interface():
if 'messages' not in st.session_state:
st.session_state.messages = []
if 'user_input' not in st.session_state:
st.session_state.user_input = ""
rag_system = initialize_rag_system()
st.markdown("""
🎓 DTMI UGM Academic Assistant
Asisten Cerdas Multimodal untuk Informasi Akademik DTMI UGM
💬 Tanyakan apapun tentang kurikulum, silabus, gambar, dan tabel data
""", unsafe_allow_html=True)
# Sidebar with controls
with st.sidebar:
st.markdown("### ⚙️ Pengaturan")
# Content type preferences
st.markdown("### 🎯 Preferensi Konten")
content_preferences = []
for content_type, description in CONTENT_TYPE_DESCRIPTIONS.items():
if st.checkbox(description, key=f"pref_{content_type}"):
content_preferences.append(content_type)
# Retrieval settings
st.markdown("### 🔍 Pengaturan Pencarian")
max_results = st.slider("Jumlah Konteks Maksimal", 5, 20, 10)
# Display settings
st.markdown("### 🎭 Tampilan")
show_images_inline = st.checkbox("🖼️ Tampilkan Gambar", value=True)
show_tables_inline = st.checkbox("📊 Tampilkan Tabel", value=True)
compact_mode = st.checkbox("📱 Mode Kompak", value=False)
# Chat statistics
if st.session_state.messages:
st.markdown("### 📊 Statistik")
total_messages = len(st.session_state.messages)
st.metric("💬 Total Pesan", total_messages)
st.metric("🗣️ Percakapan", total_messages // 2)
# Clear chat
if st.button("🗑️ Hapus Chat", type="secondary", use_container_width=True):
st.session_state.messages = []
st.rerun()
# Main chat area
col1, col2 = st.columns([3, 1] if not compact_mode else [1, 0])
with col1:
# Display chat history
for message in st.session_state.messages:
if message["role"] == "user":
st.markdown(f"""
👤 Anda:
{message["content"]}
""", unsafe_allow_html=True)
else:
st.markdown(f"""
🤖 Assistant:
{message["content"]}
""", unsafe_allow_html=True)
# 🎯 DISPLAY MULTIMODAL CONTENT
if "result_data" in message:
result_data = message["result_data"]
# Show quick stats if has multimodal content
if result_data.get("has_images") or result_data.get("has_tables"):
st.markdown("---") # Separator
col_stats1, col_stats2, col_stats3 = st.columns(3)
with col_stats1:
st.metric("🖼️ Gambar", len(result_data.get("image_data", [])))
with col_stats2:
st.metric("📊 Tabel", len(result_data.get("table_data", [])))
with col_stats3:
st.metric("📚 Sumber", result_data.get("total_sources", 0))
# 🖼️ DISPLAY IMAGES
if show_images_inline and result_data.get("has_images"):
display_images_in_chat(result_data.get("image_data", []))
# 📊 DISPLAY TABLES
if show_tables_inline and result_data.get("has_tables"):
display_tables_in_chat(result_data.get("table_data", []))
# Collapsible sources
if "sources" in message and message["sources"]:
with st.expander("📚 Lihat Sumber Informasi", expanded=False):
for i, source in enumerate(message["sources"][:3], 1):
content_type = source['metadata']['content_type']
year = source['metadata'].get('year', 'N/A')
page = source['metadata'].get('page', 'N/A')
st.markdown(f"""
**📋 Sumber {i}:** {CONTENT_TYPE_DESCRIPTIONS.get(content_type, content_type)}
**📅 Tahun:** {year} | **📄 Halaman:** {page}
**📝 Preview:** {source['content'][:150]}...
""")
st.markdown("---")
# Chat input
user_input = st.chat_input(
"💬 Tanyakan tentang kurikulum, gambar, tabel, atau informasi lainnya...", key="chat_input")
# Handle example query selection
if st.session_state.user_input:
user_input = st.session_state.user_input
st.session_state.user_input = ""
# 🚀 PROCESS USER INPUT
if user_input:
# Add user message
st.session_state.messages.append({"role": "user", "content": user_input})
# Show loading
with st.spinner("🔍 Mencari informasi relevan..."):
try:
result_data = rag_system.query(
user_input,
k=max_results,
content_filter=content_preferences if content_preferences else None
)
# Save assistant message with complete data
assistant_message = {
"role": "assistant",
"content": result_data["answer"],
"sources": result_data["sources"],
"result_data": result_data
}
st.session_state.messages.append(assistant_message)
except Exception as e:
st.error(f"❌ Terjadi kesalahan: {e}")
st.session_state.messages.append({
"role": "assistant",
"content": "Maaf, terjadi kesalahan dalam memproses pertanyaan Anda. Silakan coba lagi."
})
st.rerun()
# Sidebar dengan example queries (only if not compact)
if not compact_mode:
with col2:
display_example_queries()
# Quick actions
st.markdown("### ⚡ Aksi Cepat")
quick_actions = [
("🖼️ Cari Gambar", "Tampilkan gambar formulir atau diagram"),
("📊 Lihat Tabel", "Tabel kurikulum semester 1"),
("🎓 Info Program", "Informasi program studi teknik mesin"),
("📚 Silabus", "Silabus mata kuliah wajib")
]
for label, query in quick_actions:
if st.button(label, use_container_width=True):
st.session_state.user_input = query
st.rerun()
if st.session_state.messages:
st.markdown("### 📤 Export")
if st.button("💾 Download Chat", use_container_width=True):
chat_export = ""
for msg in st.session_state.messages:
role = "User" if msg["role"] == "user" else "Assistant"
chat_export += f"**{role}:** {msg['content']}\n\n"
st.download_button(
label="📄 Download Markdown",
data=chat_export,
file_name=f"chat_dtmi_{datetime.now().strftime('%Y%m%d_%H%M%S')}.md",
mime="text/markdown",
use_container_width=True
)
REPO_ID = "wicaksonolxn/dataoptim"
REPO_TYPE = "dataset"
LOCALROOT = Path("./repo_data") # Changed to a specific subfolder
MARKER = LOCALROOT / ".hf_mirror_complete"
def ensure_cache_dir():
cache_dir = Path.home() / ".cache" / "huggingface"
try:
cache_dir.mkdir(parents=True, exist_ok=True)
test_file = cache_dir / "test_write"
test_file.touch()
test_file.unlink()
return True
except Exception as e:
st.warning(f"Cannot create cache dir at {cache_dir}: {e}")
# Try alternative cache location
alt_cache = Path("./cache/huggingface")
try:
alt_cache.mkdir(parents=True, exist_ok=True)
os.environ["HF_HOME"] = str(alt_cache.parent)
return True
except Exception as e2:
st.error(f"Cannot create alternative cache: {e2}")
return False
def mirror_repo_fixed(repo_id: str, repo_type: str, local_root: Path) -> bool:
try:
if not ensure_cache_dir():
return False
# Ensure local root exists
local_root.mkdir(parents=True, exist_ok=True)
st.write(f"⏬ Downloading repository '{repo_id}' to {local_root}")
downloaded_path = snapshot_download(
repo_id=repo_id,
repo_type=repo_type,
local_dir=str(local_root),
local_dir_use_symlinks=False,
resume_download=True,
force_download=False,
)
# Create completion marker
MARKER.touch()
st.success(f"✅ Repository downloaded successfully to {downloaded_path}")
return True
except Exception as e:
st.error(f"❌ Error downloading repository: {e}")
return False
def mirror_repo_file_by_file(repo_id: str, repo_type: str, local_root: Path) -> bool:
"""
Alternative file-by-file download method - more control but slower
"""
try:
# Ensure cache directory exists
if not ensure_cache_dir():
return False
# Get list of files
remote_files = list_repo_files(repo_id=repo_id, repo_type=repo_type)
if not remote_files:
st.error(f"No files found in repository '{repo_id}'")
return False
st.write(f"⏬ Copying **{len(remote_files)}** files from Hugging Face Hub...")
bar = st.progress(0.0)
# Ensure local root exists
local_root.mkdir(parents=True, exist_ok=True)
for i, remote_file in enumerate(remote_files, 1):
try:
# Download to local_root directly
downloaded_file = hf_hub_download(
repo_id=repo_id,
filename=remote_file,
repo_type=repo_type,
local_dir=str(local_root), # Download directly to target directory
local_dir_use_symlinks=False,
force_download=False,
resume_download=True,
)
# Update progress
bar.progress(i / len(remote_files))
except Exception as file_error:
st.warning(f"Failed to download {remote_file}: {file_error}")
continue
# Create completion marker
MARKER.touch()
st.success("✅ Download completed – all files are now local.")
return True
except Exception as e:
st.error(f"❌ Error in file-by-file download: {e}")
return False
def verify_download(local_root: Path) -> dict:
"""Verify the downloaded repository structure"""
if not local_root.exists():
return {"status": "missing", "files": 0, "size": 0}
files = list(local_root.rglob("*"))
file_count = len([f for f in files if f.is_file()])
total_size = sum(f.stat().st_size for f in files if f.is_file())
return {
"status": "exists",
"files": file_count,
"size": total_size,
"size_mb": round(total_size / (1024 * 1024), 2)
}
@st.cache_resource(show_spinner=False)
def ensure_repo_mirrored() -> bool:
"""
Ensure repository is mirrored locally with better error handling
"""
# Check if already mirrored
if MARKER.exists():
verification = verify_download(LOCALROOT)
if verification["files"] > 0:
st.success(
f"✅ Repository already available locally ({verification['files']} files, {verification['size_mb']} MB)")
return True
else:
# Marker exists but no files - remove marker and re-download
MARKER.unlink()
# Show current status
verification = verify_download(LOCALROOT)
if verification["status"] == "exists" and verification["files"] > 0:
st.info(f"🔄 Found partial download ({verification['files']} files). Resuming...")
# Try snapshot download first (recommended)
st.info("🚀 Attempting full repository download...")
if mirror_repo_fixed(REPO_ID, REPO_TYPE, LOCALROOT):
return True
# Fallback to file-by-file if snapshot fails
st.warning("⚠️ Snapshot download failed. Trying file-by-file download...")
if mirror_repo_file_by_file(REPO_ID, REPO_TYPE, LOCALROOT):
return True
# If both methods fail
st.error("❌ All download methods failed. Please check your internet connection and repository access.")
return False
def cleanup_partial_download():
"""Clean up partial downloads and reset"""
if LOCALROOT.exists():
import shutil
shutil.rmtree(LOCALROOT)
if MARKER.exists():
MARKER.unlink()
st.success("🗑️ Cleaned up partial downloads")
# Enhanced main function with better error handling
def main():
"""Main function with robust repository handling"""
# Add cleanup option in sidebar for debugging
with st.sidebar:
st.markdown("### 🔧 Debug Options")
if st.button("🗑️ Clean & Re-download"):
cleanup_partial_download()
st.rerun()
# Show download status
verification = verify_download(LOCALROOT)
if verification["status"] == "exists":
st.success(f"📁 Local files: {verification['files']}")
st.success(f"💾 Size: {verification['size_mb']} MB")
try:
if ensure_repo_mirrored():
enhanced_chat_interface()
else:
st.error("❌ Failed to download repository. Please try again or check your connection.")
if st.button("🔄 Retry Download"):
st.rerun()
except Exception as e:
st.error(f"❌ Critical error in main(): {e}")
st.info("💡 Try using the 'Clean & Re-download' button in the sidebar")
def get_repo_info():
try:
files = list_repo_files(repo_id=REPO_ID, repo_type=REPO_TYPE)
return {
"total_files": len(files),
"file_types": list(set(Path(f).suffix for f in files if Path(f).suffix)),
"sample_files": files[:5]
}
except Exception as e:
return {"error": str(e)}
if __name__ == "__main__":
main()