Overview
The LiveDesign Python Client (LDClient) Tutorial series bridges the gap between Schrödinger’s powerful enterprise informatics platform and the scientists who use it. By reframing technical API documentation through the lens of scientific workflows, these notebooks empowered LiveDesign users at partner organizations to programmatically interact with the platform for the first time.
The Problem
Schrödinger’s LiveDesign Python Client (LDClient) is a powerful tool for programmatically interacting with the LiveDesign enterprise informatics platform. However, the existing documentation was:
- Written in source-code format — comprehensive but inaccessible to non-developers
- Organized by API endpoint — not by what a scientist would actually want to accomplish
- Lacking context — individual request examples without scientific motivation
- Missing workflow examples — no end-to-end scenarios that mirror real use cases
As a result, even technically inclined scientists at partner organizations were unable to leverage the client effectively.
The Solution
Workflow-First Organization
Rather than documenting the API alphabetically or by class hierarchy, the notebooks are organized around what scientists want to do:
| Notebook | Scientific Workflow |
|---|---|
| Getting Started | Authentication, environment setup, session management |
| Working with Compounds | Querying, registering, and managing chemical structures |
| LiveReports & Data | Reading and writing assay data, properties, and results |
| Model Integration | Interfacing with predictive ML models and running predictions |
| Bulk Operations | Handling large dataset uploads and exports efficiently |
| Advanced Workflows | Combining multiple operations for automated pipelines |
Key Features of Each Notebook
- Layered object model explained — visualizing how LiveDesign’s data structures relate to each other before showing code
- Annotated examples — every code block has scientific context explaining why this operation is useful
- Common pitfalls — real issues encountered in production use, with solutions
- Copy-paste ready — snippets designed to work with minimal modification
Distribution
The notebooks were:
- Internally distributed across Schrödinger’s deployment and support teams
- Externally shared with LiveDesign partner organizations and customers
- Used as the foundation for customer onboarding sessions and training workshops
Impact
- Enabled data scientists at partner pharma organizations to automate workflows they previously had to perform manually in the UI
- Reduced support ticket volume related to programmatic API usage
- Became the de facto reference resource for LDClient usage across the LiveDesign user community
- Facilitated custom pipeline development at multiple top-20 pharma partners