demo5

Demo5: Vector Store Chunk Explorer

This demo provides a Streamlit UI to explore vector_stores.search(...) results.

Given a query, it searches the configured vector store and renders in the central area:

• retrieved chunk text,
• score,
• file/chunk identifiers,
• page metadata and full additional properties.

Prerequisites

• Existing project configuration (config.py, config_private.py)
• VECTOR_STORE_ID configured in the active .env profile
• Active production profile selected in the same shell (source ./set_env.sh prod-chicago or source ./set_env.sh prod-frankfurt)

Install demo dependency

pip install -r demos/demo5/requirements.txt

Run

From the repository root:

streamlit run demos/demo5/app.py

Sidebar configuration

The sidebar always shows the effective runtime configuration:

• REGION
• VECTOR_STORE_ID

Notes

• Search is executed with client.vector_stores.search(...).