added all files

agent.py

@@ -0,0 +1,87 @@
+import os
+import fitz
+import faiss
+import torch
+import sqlite3
+from transformers import AutoTokenizer, AutoModelForCausalLM, pipeline
+from sentence_transformers import SentenceTransformer
+
+class CodingAgent:
+    def __init__(self):
+        # Load TinyLlama (CPU-friendly)
+        model_id = "TinyLlama/TinyLlama-1.1B-Chat-v1.0"
+        self.tokenizer = AutoTokenizer.from_pretrained(model_id)
+        self.model = AutoModelForCausalLM.from_pretrained(model_id)
+        self.llm = pipeline("text-generation", model=self.model, tokenizer=self.tokenizer, max_new_tokens=512, device=-1)
+
+        # Embedding model + FAISS index
+        self.embedder = SentenceTransformer("all-MiniLM-L6-v2")
+        self.index = faiss.IndexFlatL2(384)
+        self.docs = []
+        self.id_map = []
+
+        # SQLite for session memory
+        self.conn = sqlite3.connect("memory.db", check_same_thread=False)
+        self.conn.execute("""CREATE TABLE IF NOT EXISTS memory (id INTEGER PRIMARY KEY, query TEXT, response TEXT)""")
+
+    def embed_chunks(self, texts):
+        return self.embedder.encode(texts)
+
+    def ingest_file(self, filepath):
+        chunks = []
+        if filepath.endswith(".pdf"):
+            doc = fitz.open(filepath)
+            for page in doc:
+                text = page.get_text()
+                words = text.split()
+                for i in range(0, len(words), 300):
+                    chunk = " ".join(words[i:i+300])
+                    if len(chunk) > 100:
+                        chunks.append(chunk)
+        elif filepath.endswith(".py"):
+            with open(filepath, 'r', encoding='utf-8') as f:
+                code = f.read()
+                lines = code.splitlines()
+                for i in range(0, len(lines), 20):
+                    chunk = "\n".join(lines[i:i+20])
+                    chunks.append(chunk)
+        else:
+            return "Unsupported file format."
+
+        embeddings = self.embed_chunks(chunks)
+        self.index.add(embeddings)
+        self.docs.extend(chunks)
+        self.id_map.extend(range(len(self.docs)-len(chunks), len(self.docs)))
+        return f"Added {len(chunks)} chunks."
+
+    def retrieve_context(self, query, top_k=3):
+        if self.index.ntotal == 0:
+            return ""
+        query_emb = self.embed_chunks([query])[0]
+        D, I = self.index.search([query_emb], top_k)
+        return "\n\n".join([self.docs[i] for i in I[0]])
+
+    def answer(self, query):
+        # Check memory
+        cursor = self.conn.execute("SELECT response FROM memory WHERE query = ?", (query,))
+        result = cursor.fetchone()
+        if result:
+            return f"[From memory] {result[0]}"
+
+        context = self.retrieve_context(query)
+        prompt = f"You are a coding assistant. Answer the following:\n\nContext:\n{context}\n\nQuestion: {query}\nAnswer:"
+        result = self.llm(prompt)[0]['generated_text'].split("Answer:")[-1].strip()
+
+        self.conn.execute("INSERT INTO memory (query, response) VALUES (?, ?)", (query, result))
+        self.conn.commit()
+        return result
+
+    def clear_context(self):
+        self.conn.execute("DELETE FROM memory")
+        self.conn.commit()
+        return "Cleared memory."
+
+    def get_stats(self):
+        cursor = self.conn.execute("SELECT COUNT(*) FROM memory")
+        count = cursor.fetchone()[0]
+        return f"Stored answers: {count}\nDocuments: {len(self.docs)}"

app.py

@@ -0,0 +1,46 @@
+import gradio as gr
+from agent import CodingAgent
+
+agent = CodingAgent()
+
+def handle_query(message, history):
+    response = agent.answer(message)
+    history.append((message, response))
+    return history, ""
+
+def upload_file(file):
+    return agent.ingest_file(file.name)
+
+def clear_memory():
+    return agent.clear_context()
+
+def get_info():
+    return agent.get_stats()
+
+with gr.Blocks(title="LLaMA-3 Coding Agent") as demo:
+    gr.Markdown("# 🦙 TinyLlama Coding Agent\nSupports code Q&A + PDF/code file RAG")
+
+    with gr.Tab("Chat"):
+        chatbot = gr.Chatbot()
+        with gr.Row():
+            msg = gr.Textbox(placeholder="Ask a coding question")
+            send = gr.Button("Send")
+        send.click(handle_query, [msg, chatbot], [chatbot, msg])
+        msg.submit(handle_query, [msg, chatbot], [chatbot, msg])
+
+    with gr.Tab("Upload PDF / .py"):
+        file_input = gr.File(label="Upload PDF or Python File", file_types=[".pdf", ".py"])
+        upload_btn = gr.Button("Upload")
+        output = gr.Textbox()
+        upload_btn.click(upload_file, file_input, output)
+
+    with gr.Tab("System"):
+        info_btn = gr.Button("Get Info")
+        clear_btn = gr.Button("Clear Memory")
+        info_box = gr.Textbox()
+        status_box = gr.Textbox()
+        info_btn.click(get_info, outputs=info_box)
+        clear_btn.click(clear_memory, outputs=status_box)
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,6 @@
+transformers
+torch
+gradio
+sentence-transformers
+faiss-cpu
+PyMuPDF