Local Knowledge Base AI: Private Q&A with Ollama

Local Knowledge Base AI: Private Q&A with Ollama

💡 What You Will Learn

Local Knowledge Base AI: Private Q&A with Ollama

|:----|:--------|:----|:------------------|:--------|

# Linux/Mac
curl -fsSL https://ollama.com/install.sh | sh
ollama pull qwen2.5:7b
# ollama pull nomic-embed-textEmbedding

# requirements.txt
# chromadb==0.5.0
# ollama==0.3.0

import chromadb
from chromadb.utils import embedding_functions
import ollama

# Chroma
client = chromadb.PersistentClient(path="./local_kb")
collection = client.get_or_create_collection(
    name="company_docs",
    embedding_function=embedding_functions.OllamaEmbeddingFunction(
        model_name="nomic-embed-text"
    )
)

# 
def index_documents(docs):
    """docs = [{'id':'1','text':'...','metadata':{...}},...]"""
    ids = [d['id'] for d in docs]
    texts = [d['text'] for d in docs]
    metadatas = [d.get('metadata',{}) for d in docs]
    collection.add(documents=texts, ids=ids, metadatas=metadatas)

# Retrieve+Generate
def query_kb(question, top_k=3):
    results = collection.query(query_texts=[question], n_results=top_k)
    context = '\n'.join(results['documents'][0])
    prompt = f"\n{context}\n\n{question}"
    resp = ollama.chat(model='qwen2.5:7b', messages=[{'role':'user','content':prompt}])
    return resp['message']['content']

# Testing
sample_docs = [
    {'id':'1', 'text':'2025Q4800035%', 'metadata':{'type':''}},
    {'id':'2', 'text':'155101015', 'metadata':{'type':'HR'}},
]
index_documents(sample_docs)
print(query_kb(""))

import streamlit as st
st.title("📚 ")
question = st.text_input("")
if question:
    with st.spinner("..."):
        answer = query_kb(question)
    st.markdown(answer)