Update app.py

app.py

@@ -1,76 +1,135 @@
 import streamlit as st
 import pandas as pd
 import openpyxl
-import faiss
-import spacy
-from sentence_transformers import SentenceTransformer
-
-# ✅ Load NLP Model
-nlp = spacy.load("en_core_web_sm")
-
-# ✅ Streamlit UI
-st.set_page_config(page_title="AI-Powered Timetable", layout="wide")
-st.markdown("<h1 style='text-align: center; color: #4CAF50;'>📅 AI-Powered Timetable</h1>", unsafe_allow_html=True)
-
-# ✅ File Upload Section
-st.sidebar.markdown("## 📂 Upload Your Timetable Files")
-uploaded_timetable = st.sidebar.file_uploader("Upload Timetable (Excel)", type=["xlsx"])
-
-# ✅ Load Timetable Data
-def load_timetable(file):
-    if not file:
-        return None
-    df = pd.read_excel(file)
-    return df
-
-timetable_df = load_timetable(uploaded_timetable)
-
-# ✅ Extract Timetable as Text Data
-if timetable_df is not None:
-    timetable_texts = timetable_df.astype(str).apply(lambda x: " | ".join(x), axis=1).tolist()
-else:
-    timetable_texts = []
-
-# ✅ Initialize Embeddings
-embedder = SentenceTransformer("all-MiniLM-L6-v2")
-if timetable_texts:
-    embeddings = embedder.encode(timetable_texts, convert_to_tensor=True).cpu().numpy()
-    dimension = embeddings.shape[1]  # FIXED: Ensure valid shape
-    index = faiss.IndexFlatL2(dimension)
-    index.add(embeddings)
-
-# ✅ Retrieve Relevant Texts
-def retrieve_timetable_data(query, top_k=2):
-    query_embedding = embedder.encode([query], convert_to_tensor=True).cpu().numpy()
-    distances, indices = index.search(query_embedding, top_k)
-
-    retrieved_texts = [timetable_texts[idx] for idx in indices[0] if idx < len(timetable_texts)]
-    return "\n".join(retrieved_texts)
-
-# ✅ NLP-Based Response System
-def generate_response(query):
-    doc = nlp(query.lower())
-
-    if "who" in query:
-        return "Searching for faculty information..."
-    elif "when" in query:
-        return "Retrieving schedule details..."
-    elif "where" in query:
-        return "Locating classrooms..."
+import docx
+from io import BytesIO
+from reportlab.pdfgen import canvas
+from reportlab.lib.pagesizes import letter
+from huggingface_hub import InferenceClient
+
+# Hugging Face API Key
+API_KEY = "hf_xxxxxxxxxxxxxxxxxxxxxxxx"  # Replace with your Hugging Face API Key
+
+# Load Hugging Face TinyLlama Model
+@st.cache_resource
+def load_huggingface_client():
+    return InferenceClient(model="TinyLlama/TinyLlama-1.1B-Chat-v1.0", token=API_KEY)
+
+client = load_huggingface_client()
+
+# Function to generate AI text suggestions using TinyLlama
+def generate_suggestion(prompt):
+    response = client.text_generation(prompt, max_new_tokens=100)
+    return response.strip()
+
+# Function to load Word document
+def load_docx(file):
+    doc = docx.Document(file)
+    return "\n".join([para.text for para in doc.paragraphs])
+
+# Function to save Word document
+def save_docx(text):
+    output = BytesIO()
+    doc = docx.Document()
+    for line in text.split("\n"):
+        doc.add_paragraph(line)
+    doc.save(output)
+    return output
+
+# Function to load Excel file
+def load_excel(file):
+    return pd.ExcelFile(file)
+
+# Function to edit Excel file with full tool support
+def edit_excel(df):
+    st.write("### Excel Editor with Full Editing Tools")
+    
+    # Display DataFrame and allow modifications
+    edited_df = st.data_editor(df, num_rows="dynamic", use_container_width=True)
+
+    # Provide additional Excel formatting options
+    col_to_format = st.selectbox("Select Column to Format:", df.columns)
+    if st.button("Apply Bold Format"):
+        edited_df[col_to_format] = edited_df[col_to_format].apply(lambda x: f"**{x}**")
+
+    return edited_df
+
+# Function to generate PDF
+def save_pdf(text):
+    output = BytesIO()
+    c = canvas.Canvas(output, pagesize=letter)
+    c.setFont("Helvetica", 12)
+    y_position = 750
+    for line in text.split("\n"):
+        c.drawString(100, y_position, line)
+        y_position -= 20
+    c.save()
+    return output
+
+# Streamlit UI
+st.title("Advanced RAG + TinyLlama NLP File Manager")
+st.write("Upload, edit, and manage `.docx`, `.xls` files, with full Excel tools and AI-powered NLP.")
+
+# File Uploaders
+uploaded_files = [None] * 4
+file_types = ["docx", "xls", "docx", "xls"]
+
+for i in range(4):
+    uploaded_files[i] = st.file_uploader(f"Upload File {i+1} ({file_types[i]})", type=[file_types[i]])
+
+# File Editors
+edited_docs = {}
+edited_excels = {}
+
+for i, file in enumerate(uploaded_files):
+    if file:
+        st.subheader(f"Editing File {i+1}: {file.name}")
+
+        if file.name.endswith(".docx"):
+            text = load_docx(file)
+            edited_text = st.text_area(f"Edit {file.name}:", text, height=300)
+
+            # AI text suggestion
+            if st.button(f"Generate AI Suggestion for {file.name}"):
+                suggestion = generate_suggestion(edited_text)
+                st.text_area("AI Suggestion:", suggestion, height=200)
+
+            edited_docs[file.name] = edited_text
+
+        elif file.name.endswith(".xls"):
+            xls = load_excel(file)
+            sheet_name = st.selectbox(f"Select Sheet in {file.name}:", xls.sheet_names)
+            df = pd.read_excel(xls, sheet_name=sheet_name)
+            edited_df = edit_excel(df)
+            edited_excels[file.name] = edited_df
+
+# Download Buttons
+for file_name, text in edited_docs.items():
+    st.download_button(
+        label=f"Download {file_name}",
+        data=save_docx(text).getvalue(),
+        file_name=file_name,
+        mime="application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+    )
+
+    # PDF download option
+    st.download_button(
+        label=f"Download {file_name} as PDF",
+        data=save_pdf(text).getvalue(),
+        file_name=f"{file_name}.pdf",
+        mime="application/pdf"
+    )
+
+for file_name, df in edited_excels.items():
+    output = BytesIO()
+    with pd.ExcelWriter(output, engine="openpyxl") as writer:
+        df.to_excel(writer, index=False, sheet_name="EditedSheet")
+        writer.close()
     
-    return "Let me check the timetable."
-
-# ✅ AI Query Section
-st.markdown("## 🤖 Ask AI About Your Timetable")
-user_query = st.text_input("Type your question here")
-
-if st.button("Ask AI"):
-    retrieved_info = retrieve_timetable_data(user_query)
-    response = generate_response(user_query)
-    st.write(f"📝 **Retrieved Timetable Data:**\n{retrieved_info}")
-    st.write(f"🤖 **AI Suggests:** {response}")
-
-# ✅ Display Uploaded Timetable
-if timetable_df is not None:
-    st.markdown("## 📜 Uploaded Timetable")
-    st.dataframe(timetable_df)
+    st.download_button(
+        label=f"Download {file_name}",
+        data=output.getvalue(),
+        file_name=file_name,
+        mime="application/vnd.openxmlformats-officedocument.spreadsheetml.sheet"
+    )
+