Create app.py

app.py

@@ -0,0 +1,93 @@
+import streamlit as st
+from transformers import pipeline
+from PyPDF2 import PdfReader
+from sentence_transformers import SentenceTransformer, util
+import os
+
+# Load Hugging Face model (e.g., FLAN-T5 or GPT-like model)
+@st.cache_resource
+def load_model():
+    return pipeline("text2text-generation", model="google/flan-t5-base")
+
+# Extract text from PDF files in the dataset folder
+def extract_text_from_pdfs(folder_path):
+    pdf_texts = []
+    for filename in os.listdir(folder_path):
+        if filename.endswith(".pdf"):
+            pdf_path = os.path.join(folder_path, filename)
+            reader = PdfReader(pdf_path)
+            text = " ".join([page.extract_text() for page in reader.pages])
+            pdf_texts.append(text)
+    return pdf_texts
+
+# Embed PDF texts for similarity matching
+@st.cache_resource
+def create_embeddings(texts):
+    embedder = SentenceTransformer('all-MiniLM-L6-v2')
+    embeddings = embedder.encode(texts, convert_to_tensor=True)
+    return embedder, embeddings
+
+# Query matching function
+def get_relevant_content(query, texts, embedder, embeddings):
+    query_embedding = embedder.encode(query, convert_to_tensor=True)
+    scores = util.pytorch_cos_sim(query_embedding, embeddings)
+    best_idx = scores.argmax().item()
+    return texts[best_idx]
+
+# Streamlit UI
+def main():
+    st.title("Educational Assistant Chatbot")
+    st.markdown("Welcome! Ask me about academic programs, visa requirements, and more!")
+
+    # Sidebar for Admin Options
+    st.sidebar.header("Admin Options")
+    dataset_folder = st.sidebar.text_input("Dataset Folder Path", "./pdf_dataset")
+    email_to_send = st.sidebar.text_input("Email for Profile Submission", "application@aspireec.pk")
+
+    # Load model and dataset
+    st.sidebar.markdown("### Data Loading")
+    if st.sidebar.button("Load Dataset"):
+        st.session_state.pdf_texts = extract_text_from_pdfs("chatbot")
+        st.session_state.embedder, st.session_state.embeddings = create_embeddings(st.session_state.pdf_texts)
+        st.sidebar.success("Dataset Loaded Successfully!")
+
+    # Chat interface
+    user_query = st.text_input("Your Query:")
+    if st.button("Ask"):
+        if "embedder" not in st.session_state or "pdf_texts" not in st.session_state:
+            st.error("Dataset is not loaded. Please load the dataset in the sidebar.")
+        else:
+            relevant_content = get_relevant_content(
+                user_query, st.session_state.pdf_texts, st.session_state.embedder, st.session_state.embeddings
+            )
+            model = load_model()
+            response = model(f"Question: {user_query} Context: {relevant_content}", max_length=200)
+            st.success(response[0]['generated_text'])
+
+    # Profile creation
+    st.markdown("### Create a Student Profile")
+    name = st.text_input("Name:")
+    email = st.text_input("Email:")
+    contact_number = st.text_input("Contact Number (Optional):")
+    study_level = st.selectbox("Level of Study", ["Undergraduate", "Postgraduate", "PhD"])
+    field_of_interest = st.text_input("Field of Interest:")
+    career_goal = st.text_area("Career Goals:")
+    visa_query = st.text_area("Visa Concerns or Questions:")
+
+    if st.button("Submit Profile"):
+        profile = f"""
+        Name: {name}
+        Email: {email}
+        Contact Number: {contact_number}
+        Level of Study: {study_level}
+        Field of Interest: {field_of_interest}
+        Career Goals: {career_goal}
+        Visa Queries: {visa_query}
+        """
+        # Simulate email sending (for simplicity, just display the profile)
+        st.markdown("### Profile Summary")
+        st.code(profile)
+        st.success(f"Profile submitted to {email_to_send}!")
+
+if __name__ == "__main__":
+    main()