Fix MCQ functionality issues: consistent handler signatures and legacy fallback

[Copilot]

This PR addresses the critical issues identified in the MCQ functionality review comments:

Issues Fixed

1. Inconsistent Parameter Passing Between Handlers

Problem: OpenAI and Gemini handlers were receiving different parameters, making the code inconsistent and harder to maintain.

Before:

// OpenAI handler
OpenAIHandler.generateSolutionResponses(solutionModel, problemInfo.coding_data, signal);

// Gemini handler  
GeminiHandler.generateSolutionResponses(solutionModel, problemInfo);

After: Both handlers now receive the same parameters:

// Both handlers now consistent
OpenAIHandler.generateSolutionResponses(solutionModel, problemInfo.coding_data, signal);
GeminiHandler.generateSolutionResponses(solutionModel, problemInfo.coding_data, signal);

2. Gemini Handler Function Signature Mismatch

Problem: Gemini handler expected UnifiedProblemSchema while OpenAI handler expected ProblemSchema directly.

Before:

// Gemini handler signature
export async function generateSolutionResponses(
  modelName: GeminiModel,
  problemInfo: UnifiedProblemSchema,
): Promise<SolutionSchema>

// Had to validate internally and extract coding_data
if (problemInfo.type !== 'coding' || !problemInfo.coding_data) {
  throw new Error('Solution generation is only available for coding problems');
}
const codingData = problemInfo.coding_data;

After: Now matches OpenAI handler signature:

// Gemini handler signature now consistent
export async function generateSolutionResponses(
  modelName: GeminiModel,
  problemInfo: ProblemSchema,
  signal: AbortSignal,
): Promise<SolutionSchema>

// Direct usage without internal validation
const codingData = problemInfo;

3. Incomplete Legacy Fallback Logic

Problem: Legacy problemInfo without type property would be set to null even for valid coding problems.

Before:

const codingData = problemInfo?.type === 'coding' ? problemInfo.coding_data : null;
// Legacy format would always result in null

After: Proper fallback for legacy format:

const codingData = problemInfo?.type === 'coding' 
  ? problemInfo.coding_data 
  : (problemInfo?.type ? null : problemInfo as unknown as ProblemSchema);
// Legacy format without 'type' is treated as ProblemSchema

Benefits

• Consistency: Both handlers now have identical function signatures
• Maintainability: Reduced code duplication and complexity
• Backward Compatibility: Legacy problemInfo format continues to work
• Type Safety: Proper TypeScript types and imports added
• No Breaking Changes: All existing functionality preserved

Testing

• ✅ TypeScript compilation passes
• ✅ Build process successful
• ✅ No test failures (existing test suite passes)

This addresses all review comments from the upstream PR and ensures the MCQ functionality works consistently across both OpenAI and Gemini handlers while maintaining backward compatibility.

This pull request was created as a result of the following prompt from Copilot chat.

Fix Issues in MCQ Functionality PR

Based on the review comments in the upstream PR (xhoantran#7), there are several critical issues that need to be fixed:

Issues to Fix:

1. Inconsistent Parameter Passing Between Handlers

• Problem: OpenAI handler receives problemInfo.coding_data and signal, but Gemini handler receives modelName and problemInfo
• Location: src/main/processor/index.ts line 42
• Fix: Make both handlers consistent by passing the same parameters

2. Gemini Handler Function Signature Mismatch

• Problem: Gemini handler expects UnifiedProblemSchema but should receive ProblemSchema directly like OpenAI handler
• Location: src/main/processor/GeminiHandler.ts around line 390
• Fix: Change function signature to accept ProblemSchema directly and remove internal validation

3. Incomplete Legacy Fallback Logic

• Problem: Legacy problemInfo without type property will be set to null even for valid coding problems
• Location: src/renderer/features/solution/index.tsx line 50
• Fix: Add proper fallback for legacy format: const codingData = problemInfo?.type === 'coding' ? problemInfo.coding_data : (problemInfo?.type ? null : problemInfo as ProblemSchema);

Specific Changes Needed:

File: src/main/processor/index.ts

• Ensure consistent parameter passing to both OpenAI and Gemini handlers
• Both should receive the same data structure

File: src/main/processor/GeminiHandler.ts

• Update generateSolutionResponses function signature to match OpenAI handler
• Accept ProblemSchema directly instead of UnifiedProblemSchema
• Remove internal type validation since it's done in the processor index

File: src/renderer/features/solution/index.tsx

• Fix legacy fallback logic to properly handle old problemInfo format
• Ensure backward compatibility is maintained

Expected Outcome:

• Both handlers have consistent function signatures
• Legacy problemInfo format continues to work
• No breaking changes for existing functionality
• All review comments are addressed

---

💬 Share your feedback on Copilot coding agent for the chance to win a $200 gift card! Click here to start the survey.