@democratize-quality/mcp-server 1.1.0 → 1.1.2

package/src/tools/api/prompts/generation-prompts.js

@@ -0,0 +1,599 @@
+/**
+ * Code Generation Prompts for API Test Generator
+ * 
+ * This module contains all prompts used for AI-based test code generation.
+ * Separated for easy iteration, versioning, and prompt engineering.
+ */
+
+module.exports = {
+  /**
+   * Playwright test generation prompts
+   */
+  playwright: {
+    system: `You are an expert Playwright test engineer. Generate clean, idiomatic, production-ready Playwright API tests.
+
+CRITICAL - File Generation Rules:
+- Generate ONLY the ONE requested test file
+- DO NOT create any additional files (README, SUMMARY, GUIDE, NOTES, etc.)
+- DO NOT create documentation files unless explicitly requested
+- DO NOT create helper files unless explicitly requested
+- Output ONLY the test code for the specified file
+
+Requirements:
+- Use Playwright's request fixture directly (no wrappers)
+- Use proper TypeScript types when language=typescript
+- Include proper async/await patterns
+- Use template literals for URL construction with \${baseUrl}
+- Generate proper assertions with expect()
+- Handle both single and chained requests
+- Include proper error handling
+- Add descriptive comments for complex logic
+
+Code Style:
+- 2-space indentation
+- Single quotes for strings (except template literals)
+- Semicolons at end of statements
+- Clear variable naming
+- Proper spacing and formatting`,
+
+    /**
+     * Generate a section test file with multiple scenarios
+     */
+    generateTest: (context) => {
+      const { testPlanTitle, baseUrl, sectionTitle, scenarios, language, isTypeScript, sessionId } = context;
+      
+      let prompt = `Generate a complete Playwright test file for this section:
+
+**Test Plan:** ${testPlanTitle}
+**Section:** ${sectionTitle}
+**Base URL:** ${baseUrl}
+**Language:** ${language}
+**Session ID:** ${sessionId}
+
+**Scenarios to test:**
+`;
+
+      scenarios.forEach((scenario, idx) => {
+        prompt += `
+${idx + 1}. **${scenario.title}**
+   - Method: ${scenario.method}
+   - Endpoint: ${scenario.endpoint}
+   - Expected Status: ${scenario.expectedStatus || 200}
+`;
+        
+        if (scenario.requestBody) {
+          prompt += `   - Request Body: ${JSON.stringify(scenario.requestBody, null, 2).substring(0, 200)}...
+`;
+        }
+        
+        if (scenario.expectedBody) {
+          prompt += `   - Expected Body Fields: ${Object.keys(scenario.expectedBody).join(', ')}
+`;
+        }
+
+        if (scenario.steps && scenario.steps.length > 0) {
+          prompt += `   - Multi-step test with ${scenario.steps.length} sequential requests
+`;
+        }
+      });
+
+      prompt += `
+**Instructions:**
+1. Generate a COMPLETE test file including imports and test structure
+2. ${isTypeScript ? 'Use TypeScript with proper types' : 'Use JavaScript'}
+3. Import from '@playwright/test' for Playwright
+4. Create test.describe() for the section: "${sectionTitle}"
+5. Create one test() for each scenario above
+6. Use \`const baseUrl = '${baseUrl}';\` at the top of describe block
+7. Use Playwright's \`request\` fixture directly (no helper classes)
+8. Construct URLs as template literals: \`\${baseUrl}/endpoint\`
+9. Include proper assertions for status codes and response body structure
+10. For multi-step tests, chain requests and pass data between steps
+11. Add error handling where appropriate
+12. Include comments for complex logic
+
+**CRITICAL - Output Restrictions:**
+- Generate ONLY this ONE test file: ${sectionTitle}
+- DO NOT generate README, SUMMARY, GUIDE, or any other files
+- DO NOT suggest creating additional files
+- DO NOT include file creation instructions
+- Return ONLY the test file code as plain text
+
+**File structure:**
+\`\`\`
+// Header comment
+import/require statements
+test.describe('${sectionTitle}', () => {
+  const baseUrl = '${baseUrl}';
+  
+  test('scenario 1', async ({ request }) => {
+    // test code
+  });
+  
+  test('scenario 2', async ({ request }) => {
+    // test code
+  });
+});
+\`\`\`
+
+Return the COMPLETE file content for THIS FILE ONLY.`;
+
+      return prompt;
+    },
+
+    /**
+     * Generate main test file that runs all sections
+     */
+    generateMainTest: (context) => {
+      const { testPlanTitle, baseUrl, sections, language, isTypeScript, sessionId } = context;
+      
+      let prompt = `Generate a main Playwright test file that orchestrates all test sections:
+
+**Test Plan:** ${testPlanTitle}
+**Base URL:** ${baseUrl}
+**Language:** ${language}
+**Session ID:** ${sessionId}
+
+**Sections:**
+`;
+
+      sections.forEach((section, idx) => {
+        prompt += `${idx + 1}. ${section.title} (${section.scenarioCount} scenarios)
+`;
+      });
+
+      prompt += `
+**Instructions:**
+1. Generate a COMPLETE main test file
+2. ${isTypeScript ? 'Use TypeScript' : 'Use JavaScript'}
+3. Import from '@playwright/test'
+4. Create one main test.describe() for "${testPlanTitle}"
+5. Set \`const baseUrl = '${baseUrl}';\`
+6. Import or reference individual section test files
+7. Keep it simple - this is just an orchestration file
+8. Add a comment noting individual sections are in separate files
+
+**CRITICAL - Output Restrictions:**
+- Generate ONLY this ONE main test file
+- DO NOT generate README, SUMMARY, GUIDE, or any other files
+- DO NOT create section files (they're generated separately)
+- Return ONLY the main test file code
+
+Return the COMPLETE main file content for THIS FILE ONLY.`;
+
+      return prompt;
+    }
+  },
+
+  /**
+   * Jest test generation prompts
+   */
+  jest: {
+    system: `You are an expert Jest test engineer. Generate clean, idiomatic, production-ready Jest API tests using axios.
+
+CRITICAL - File Generation Rules:
+- Generate ONLY the ONE requested test file
+- DO NOT create any additional files (README, SUMMARY, GUIDE, NOTES, etc.)
+- DO NOT create documentation files unless explicitly requested
+- Output ONLY the test code for the specified file
+
+Requirements:
+- Use axios for API calls
+- Use proper TypeScript types when language=typescript
+- Include proper async/await patterns
+- Generate proper assertions with expect()
+- Handle both single and chained requests
+- Include proper error handling
+
+Code Style:
+- 2-space indentation
+- Single quotes for strings (except template literals)
+- Semicolons at end of statements
+- Clear variable naming`,
+
+    /**
+     * Generate a section test file with multiple scenarios
+     */
+    generateTest: (context) => {
+      const { testPlanTitle, baseUrl, sectionTitle, scenarios, language, isTypeScript, sessionId } = context;
+      
+      let prompt = `Generate a complete Jest test file for this section:
+
+**Test Plan:** ${testPlanTitle}
+**Section:** ${sectionTitle}
+**Base URL:** ${baseUrl}
+**Language:** ${language}
+**Session ID:** ${sessionId}
+
+**Scenarios to test:**
+`;
+
+      scenarios.forEach((scenario, idx) => {
+        prompt += `
+${idx + 1}. **${scenario.title}**
+   - Method: ${scenario.method}
+   - Endpoint: ${scenario.endpoint}
+   - Expected Status: ${scenario.expectedStatus || 200}
+`;
+        
+        if (scenario.requestBody) {
+          prompt += `   - Request Body: ${JSON.stringify(scenario.requestBody, null, 2).substring(0, 200)}...
+`;
+        }
+        
+        if (scenario.expectedBody) {
+          prompt += `   - Expected Body Fields: ${Object.keys(scenario.expectedBody).join(', ')}
+`;
+        }
+
+        if (scenario.steps && scenario.steps.length > 0) {
+          prompt += `   - Multi-step test with ${scenario.steps.length} sequential requests
+`;
+        }
+      });
+
+      prompt += `
+**Instructions:**
+1. Generate a COMPLETE test file including imports and test structure
+2. ${isTypeScript ? 'Use TypeScript with proper types' : 'Use JavaScript'}
+3. Import axios for HTTP requests
+4. Create describe() for the section: "${sectionTitle}"
+5. Create one test() for each scenario above
+6. Use \`const baseUrl = '${baseUrl}';\` at the top
+7. Use axios methods (axios.get, axios.post, etc.)
+8. Construct URLs as template literals: \`\${baseUrl}/endpoint\`
+9. Include proper assertions with Jest's expect()
+10. For multi-step tests, chain requests and pass data between steps
+11. Add error handling where appropriate
+
+**CRITICAL - Output Restrictions:**
+- Generate ONLY this ONE test file
+- DO NOT generate README, SUMMARY, GUIDE, or any other files
+- DO NOT suggest creating additional files
+- Return ONLY the test file code as plain text
+
+**File structure:**
+\`\`\`
+// Header comment
+import/require statements
+describe('${sectionTitle}', () => {
+  const baseUrl = '${baseUrl}';
+  
+  test('scenario 1', async () => {
+    // test code
+  });
+  
+  test('scenario 2', async () => {
+    // test code
+  });
+});
+\`\`\`
+
+Return the COMPLETE file content for THIS FILE ONLY.`;
+
+      return prompt;
+    },
+
+    /**
+     * Generate main test file
+     */
+    generateMainTest: (context) => {
+      const { testPlanTitle, baseUrl, sections, language, isTypeScript, sessionId } = context;
+      
+      let prompt = `Generate a main Jest test file that orchestrates all test sections:
+
+**Test Plan:** ${testPlanTitle}
+**Base URL:** ${baseUrl}
+**Language:** ${language}
+**Session ID:** ${sessionId}
+
+**Sections:**
+`;
+
+      sections.forEach((section, idx) => {
+        prompt += `${idx + 1}. ${section.title} (${section.scenarioCount} scenarios)
+`;
+      });
+
+      prompt += `
+**Instructions:**
+1. Generate a COMPLETE main test file
+2. ${isTypeScript ? 'Use TypeScript' : 'Use JavaScript'}
+3. Import axios
+4. Create one main describe() for "${testPlanTitle}"
+5. Set \`const baseUrl = '${baseUrl}';\`
+6. Add a comment noting individual sections are in separate files
+
+**CRITICAL - Output Restrictions:**
+- Generate ONLY this ONE main test file
+- DO NOT generate README, SUMMARY, GUIDE, or any other files
+- Return ONLY the main test file code
+
+Return the COMPLETE main file content for THIS FILE ONLY.`;
+
+      return prompt;
+    }
+  },
+
+  /**
+   * Legacy scenario-level prompts (kept for backward compatibility)
+   */
+  playwrightScenario: {
+    system: `You are an expert Playwright test engineer. Generate clean, idiomatic, production-ready Playwright API tests.
+
+Requirements:
+- Use Playwright's request fixture directly (no wrappers)
+- Use proper TypeScript types when language=typescript
+- Include proper async/await patterns
+- Use template literals for URL construction with \${baseUrl}
+- Generate proper assertions with expect()
+- Handle both single and chained requests
+- Include proper error handling
+- Add descriptive comments for complex logic
+
+Code Style:
+- 2-space indentation
+- Single quotes for strings (except template literals)
+- Semicolons at end of statements
+- Clear variable naming
+- Proper spacing and formatting`,
+
+    user: (context) => {
+      const { scenario, baseUrl, language, options } = context;
+      
+      let prompt = `Generate a Playwright test for this scenario:
+
+**Test Details:**
+- Title: "${scenario.title}"
+- Method: ${scenario.method}
+- Endpoint: ${scenario.endpoint}
+- Base URL: ${baseUrl}
+- Language: ${language}
+
+`;
+
+      // Add request details
+      if (scenario.data && Object.keys(scenario.data).length > 0) {
+        prompt += `**Request Body:**
+\`\`\`json
+${JSON.stringify(scenario.data, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      if (scenario.headers && Object.keys(scenario.headers).length > 0) {
+        prompt += `**Headers:**
+\`\`\`json
+${JSON.stringify(scenario.headers, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      if (scenario.query && Object.keys(scenario.query).length > 0) {
+        prompt += `**Query Parameters:**
+\`\`\`json
+${JSON.stringify(scenario.query, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      if (scenario.pathParams && Object.keys(scenario.pathParams).length > 0) {
+        prompt += `**Path Parameters:**
+\`\`\`json
+${JSON.stringify(scenario.pathParams, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      // Add expectations
+      if (scenario.expect) {
+        prompt += `**Expected Response:**
+- Status Code: ${scenario.expect.status || 200}
+`;
+        if (scenario.expect.body) {
+          prompt += `- Response Body Structure:
+\`\`\`json
+${JSON.stringify(scenario.expect.body, null, 2)}
+\`\`\`
+`;
+        }
+      }
+
+      // Add chained requests if present
+      if (scenario.chain && scenario.chain.length > 0) {
+        prompt += `\n**Chained Requests:**
+This is a multi-step test with ${scenario.chain.length} sequential API calls.
+`;
+        scenario.chain.forEach((step, index) => {
+          prompt += `
+Step ${index + 1}: ${step.name}
+- Method: ${step.method}
+- Endpoint: ${step.endpoint}
+`;
+          if (step.extract) {
+            prompt += `- Extract: ${JSON.stringify(step.extract)}
+`;
+          }
+        });
+      }
+
+      prompt += `
+**Instructions:**
+1. Generate ONLY the test body code (the code inside the test() function)
+2. Start at the correct indentation level (6 spaces for test body)
+3. Use \`await request.${scenario.method.toLowerCase()}()\` for the API call
+4. Construct URL as: \`\${baseUrl}${scenario.endpoint}\`
+5. Include all request options (headers, params, data)
+6. Add proper assertions for status code and response body
+7. For TypeScript, use proper types
+8. For chained requests, store results and use them in subsequent calls
+
+DO NOT include:
+- test.describe() wrapper
+- test() function declaration
+- import statements
+- Extra blank lines at start/end
+
+Return ONLY the test body code, properly indented.`;
+
+      return prompt;
+    }
+  },
+
+  /**
+   * Generate a complete Jest test for a single scenario
+   */
+  jestScenario: {
+    system: `You are an expert Jest test engineer. Generate clean, idiomatic, production-ready Jest API tests using axios.
+
+Requirements:
+- Use axios for API calls
+- Use proper TypeScript types when language=typescript
+- Include proper async/await patterns
+- Generate proper assertions with expect()
+- Handle both single and chained requests
+- Include proper error handling
+- Add descriptive comments for complex logic
+
+Code Style:
+- 2-space indentation
+- Single quotes for strings
+- Semicolons at end of statements
+- Clear variable naming
+- Proper spacing and formatting`,
+
+    user: (context) => {
+      const { scenario, baseUrl, language, options } = context;
+      
+      let prompt = `Generate a Jest/axios test for this scenario:
+
+**Test Details:**
+- Title: "${scenario.title}"
+- Method: ${scenario.method}
+- Endpoint: ${scenario.endpoint}
+- Base URL: ${baseUrl}
+- Language: ${language}
+
+`;
+
+      // Add request details (similar to Playwright)
+      if (scenario.data && Object.keys(scenario.data).length > 0) {
+        prompt += `**Request Body:**
+\`\`\`json
+${JSON.stringify(scenario.data, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      if (scenario.headers && Object.keys(scenario.headers).length > 0) {
+        prompt += `**Headers:**
+\`\`\`json
+${JSON.stringify(scenario.headers, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      if (scenario.query && Object.keys(scenario.query).length > 0) {
+        prompt += `**Query Parameters:**
+\`\`\`json
+${JSON.stringify(scenario.query, null, 2)}
+\`\`\`
+
+`;
+      }
+
+      // Add expectations
+      if (scenario.expect) {
+        prompt += `**Expected Response:**
+- Status Code: ${scenario.expect.status || 200}
+`;
+        if (scenario.expect.body) {
+          prompt += `- Response Body Structure:
+\`\`\`json
+${JSON.stringify(scenario.expect.body, null, 2)}
+\`\`\`
+`;
+        }
+      }
+
+      prompt += `
+**Instructions:**
+1. Generate ONLY the test body code (the code inside the test() function)
+2. Start at the correct indentation level (6 spaces for test body)
+3. Use \`await apiUtils.makeRequest()\` helper for API calls
+4. Include all request options (method, url, headers, params, data)
+5. Add proper assertions for status code and response body
+6. For TypeScript, use proper types
+7. Handle errors appropriately
+
+DO NOT include:
+- describe() wrapper
+- test() function declaration
+- import statements
+- Extra blank lines at start/end
+
+Return ONLY the test body code, properly indented.`;
+
+      return prompt;
+    }
+  },
+
+  /**
+   * Generate Postman test scripts
+   */
+  postmanTestScript: {
+    system: `You are a Postman automation expert. Generate clear, effective Postman test scripts.
+
+Requirements:
+- Use pm.test() for assertions
+- Use pm.response for response validation
+- Use pm.expect() for assertions
+- Include proper error messages
+- Add descriptive test names`,
+
+    user: (context) => {
+      const { expect, extract } = context;
+      
+      let prompt = `Generate Postman test script for this validation:
+
+`;
+
+      if (expect && expect.status) {
+        prompt += `**Expected Status:** ${expect.status}
+`;
+      }
+
+      if (expect && expect.body) {
+        prompt += `**Expected Response Body:**
+\`\`\`json
+${JSON.stringify(expect.body, null, 2)}
+\`\`\`
+`;
+      }
+
+      if (extract) {
+        prompt += `**Extract Variables:**
+${Object.entries(extract).map(([key, field]) => `- ${key} from ${field}`).join('\n')}
+`;
+      }
+
+      prompt += `
+**Instructions:**
+1. Generate test script as array of strings
+2. Each line should be a separate array element
+3. Include status code validation
+4. Include response body validation
+5. Extract and store variables if specified
+
+Return JavaScript array of test script lines.`;
+
+      return prompt;
+    }
+  }
+};