Code Walkthrough

Core Components

The assistant has three main components:

class SimpleLocalAssistant:
    def __init__(self):
        # ChromaDB for vector storage
        self.client = chromadb.PersistentClient(path="./chroma_db")
        self.collection = self.client.get_or_create_collection("local_docs")
    
    def add_document(self, file_path: str):
        # Load, chunk, and embed documents
        text = self.load_text_file(file_path)
        chunks = self.chunk_text(text)
        self.collection.add(documents=chunks, ids=ids)
    
    def ask(self, question: str):
        # Search, retrieve, and generate
        results = self.collection.query(query_texts=[question])
        context = "\n".join(results['documents'][0])
        answer = self.ask_ollama(question, context)
        return answer

Document Chunking

Documents are split into overlapping chunks to preserve context:

def chunk_text(self, text, chunk_size=1000, overlap=200):
    chunks = []
    start = 0
    while start < len(text):
        end = start + chunk_size
        chunks.append(text[start:end])
        start = end - overlap  # Overlap preserves context
    return chunks

Query Processing

When you ask a question:

def ask(self, question):
    # 1. Search for relevant chunks
    results = self.collection.query(
        query_texts=[question],
        n_results=4  # Top 4 most relevant chunks
    )
    
    # 2. Combine context
    context = "\n\n".join(results['documents'][0])
    
    # 3. Generate answer with LLM
    prompt = f"""Context: {context}
    
    Question: {question}
    
    Answer:"""
    
    response = ollama.chat(model="llama3.1:8b", ...)
    return response