Tool use with local context in DSPy

This example uses DSPy with a local knowledge base.

The module calls a tiny retrieval tool first, then passes the retrieved context into a dspy.Predict step that drafts the answer.

Requirements

• Python 3.10+

Steps for running this example in the shell

1. Install dependencies:

python -m venv .venv

# macOS/Linux:
source .venv/bin/activate

# Windows Command Prompt:
.venv\Scripts\activate.bat

# Windows PowerShell:
.venv\Scripts\Activate.ps1

pip install -r requirements.txt

2. Run the script:

python context_tooling.py

What it demonstrates

• A local tool that retrieves relevant context from an offline knowledge base
• A DSPy module that threads that context into the model call
• A repeatable local LM stub, so no external service is needed

Output

Question 1: How do I reset my password?
Answer: Use the self-service portal to reset the password, then retry the login.
Source: Reset password

Question 2: I cannot connect to the VPN from home.
Answer: Open the local VPN client, select the office profile, and reconnect.
Source: VPN access

Question 3: Where can I download the invoice copy?
Answer: Download the latest invoice from the billing page.
Source: Invoice copy