OpenParable
Skip to content
← Back to skills

Openevidence Local Dev Loop

Development
openevidence-local-dev-loop

Openevidence Local Dev Loop is an Agent Skill for development published in the Dicklesworthstone/pi_agent_rust repository on GitHub: a SKILL.md instruction file an AI agent can load to do this task.

Set up local development environment for OpenEvidence integration. Use when configuring development workflow, setting up testing environment, or creating a rapid iteration loop for clinical AI development. Trigger with phrases like "openevidence dev setup", "openevidence local", "openevidence development", "openevidence testing environment".

View on GitHub ↗
QUICK START

How to use this skill

Bring this guide into your coding agent with a prompt tailored to the tool you use.

  1. Open your project in Codex.
  2. Copy and paste the prompt below.
  3. Review files and risks before approving any action.
Prompt to paste
Inspect the Agent Skill at https://github.com/Dicklesworthstone/pi_agent_rust/blob/HEAD/tests/ext_conformance/artifacts/plugins-community/plugins/saas-packs/openevidence-pack/skills/openevidence-local-dev-loop/SKILL.md and all supporting files. Report provenance, permissions, scripts, dependencies, and security risks. Propose an exact installation into .agents/skills/openevidence-local-dev-loop, but wait for my approval before writing files or running scripts.

Copying this prompt does not install or run the skill. Review third-party files before use.

OpenEvidence Local Dev Loop

Overview

Set up a rapid development loop for building and testing OpenEvidence integrations locally.

Prerequisites

  • Completed openevidence-install-auth setup
  • Node.js 18+ with npm/pnpm or Python 3.10+
  • Text editor or IDE configured
  • Git repository initialized

Instructions

Step 1: Project Structure

mkdir -p src/{openevidence,services,utils}
mkdir -p tests/{unit,integration}
mkdir -p config

# Create essential files
touch src/openevidence/{client,types,errors}.ts
touch src/services/clinical-query.ts
touch tests/integration/openevidence.test.ts
touch config/{development,test}.json

Step 2: Environment Configuration

# .env.development
cat > .env.development << 'EOF'
OPENEVIDENCE_API_KEY=oe_sandbox_***
OPENEVIDENCE_ORG_ID=org_sandbox_***
OPENEVIDENCE_BASE_URL=https://api.sandbox.openevidence.com
OPENEVIDENCE_TIMEOUT=60000
LOG_LEVEL=debug
EOF

# .env.test
cat > .env.test << 'EOF'
OPENEVIDENCE_API_KEY=oe_test_***
OPENEVIDENCE_ORG_ID=org_test_***
OPENEVIDENCE_BASE_URL=https://api.sandbox.openevidence.com
OPENEVIDENCE_TIMEOUT=30000
LOG_LEVEL=error
EOF

Step 3: Dev Dependencies

# TypeScript project
npm install -D typescript ts-node vitest @types/node dotenv-cli

# Add scripts to package.json
npm pkg set scripts.dev="dotenv -e .env.development -- ts-node src/index.ts"
npm pkg set scripts.test="dotenv -e .env.test -- vitest"
npm pkg set scripts.test:integration="dotenv -e .env.test -- vitest run tests/integration"
npm pkg set scripts.watch="ts-node-dev --respawn src/index.ts"

Step 4: Development Client Wrapper

// src/openevidence/client.ts
import { OpenEvidenceClient } from '@openevidence/sdk';

interface ClientConfig {
  apiKey: string;
  orgId: string;
  baseUrl?: string;
  timeout?: number;
}

let clientInstance: OpenEvidenceClient | null = null;

export function getClient(config?: Partial<ClientConfig>): OpenEvidenceClient {
  if (!clientInstance || config) {
    clientInstance = new OpenEvidenceClient({
      apiKey: config?.apiKey || process.env.OPENEVIDENCE_API_KEY!,
      orgId: config?.orgId || process.env.OPENEVIDENCE_ORG_ID!,
      baseUrl: config?.baseUrl || process.env.OPENEVIDENCE_BASE_URL,
      timeout: config?.timeout || parseInt(process.env.OPENEVIDENCE_TIMEOUT || '30000'),
    });
  }
  return clientInstance;
}

// For testing - allows resetting singleton
export function resetClient(): void {
  clientInstance = null;
}

Step 5: Mock Client for Unit Tests

// tests/mocks/openevidence.ts
import { vi } from 'vitest';

export const mockClinicalQuery = vi.fn().mockResolvedValue({
  answer: 'Mock clinical answer for testing',
  citations: [
    { source: 'Test Journal 2025', title: 'Test Article', authors: ['Test Author'] }
  ],
  confidence: 0.95,
  lastUpdated: '2025-01-01',
});

export const mockOpenEvidenceClient = {
  query: mockClinicalQuery,
  deepConsult: vi.fn(),
  health: { check: vi.fn().mockResolvedValue({ status: 'healthy' }) },
};

vi.mock('@openevidence/sdk', () => ({
  OpenEvidenceClient: vi.fn(() => mockOpenEvidenceClient),
}));

Step 6: Hot Reload Development Script

// scripts/dev-server.ts
import { watch } from 'chokidar';
import { spawn, ChildProcess } from 'child_process';

let serverProcess: ChildProcess | null = null;

function startServer() {
  if (serverProcess) {
    serverProcess.kill();
  }

  console.log('[dev] Starting server...');
  serverProcess = spawn('npx', ['ts-node', 'src/index.ts'], {
    stdio: 'inherit',
    env: { ...process.env, NODE_ENV: 'development' },
  });
}

// Watch for changes
watch('src/**/*.ts', { ignoreInitial: false })
  .on('all', (event, path) => {
    console.log(`[dev] ${event}: ${path}`);
    startServer();
  });

Output

  • Configured development environment with hot reload
  • Separate environment files for dev/test
  • Mock client for unit testing
  • Integration test setup for sandbox API

Development Workflow

# Start development with hot reload
npm run dev

# Run unit tests in watch mode
npm test

# Run integration tests against sandbox
npm run test:integration

# Type check
npx tsc --noEmit

Error Handling

IssueCauseSolution
Env vars undefined.env not loadedCheck dotenv-cli is running
Sandbox unavailableWrong base URLVerify OPENEVIDENCE_BASE_URL
Tests timing outNetwork issuesMock client for unit tests
Hot reload not workingFile watcher issueRestart dev script

Examples

Integration Test Example

// tests/integration/openevidence.test.ts
import { describe, it, expect, beforeAll } from 'vitest';
import { getClient, resetClient } from '../../src/openevidence/client';

describe('OpenEvidence Integration', () => {
  beforeAll(() => {
    resetClient();
  });

  it('should successfully query clinical evidence', async () => {
    const client = getClient();

    const response = await client.query({
      question: 'What is the half-life of aspirin?',
      context: { specialty: 'pharmacology' },
    });

    expect(response.answer).toBeDefined();
    expect(response.citations.length).toBeGreaterThan(0);
    expect(response.confidence).toBeGreaterThan(0.5);
  }, 30000); // 30s timeout for API calls
});

Unit Test with Mocks

// tests/unit/clinical-query.test.ts
import { describe, it, expect, vi, beforeEach } from 'vitest';
import { mockClinicalQuery, mockOpenEvidenceClient } from '../mocks/openevidence';
import { queryClinicalEvidence } from '../../src/services/clinical-query';

describe('Clinical Query Service', () => {
  beforeEach(() => {
    vi.clearAllMocks();
  });

  it('should return formatted clinical answer', async () => {
    const result = await queryClinicalEvidence('test question');

    expect(mockClinicalQuery).toHaveBeenCalledWith(
      expect.objectContaining({ question: 'test question' })
    );
    expect(result.answer).toBe('Mock clinical answer for testing');
  });
});

Resources

Next Steps

For SDK patterns and best practices, see openevidence-sdk-patterns.