docs: add context gathering rule and update existing rules
- Create comprehensive context_gathering.mdc rule documenting ContextGatherer utility patterns, FuzzyTaskSearch integration, token breakdown display, code block syntax highlighting, and enhanced result display patterns - Update new_features.mdc to include context gathering step - Update commands.mdc with context-aware command pattern - Update ui.mdc with enhanced display patterns and syntax highlighting - Update utilities.mdc to document new context gathering utilities - Update glossary.mdc to include new context_gathering rule - Establishes standardized patterns for building intelligent, context-aware commands that can leverage project knowledge for better AI assistance
This commit is contained in:
Eyal Toledano
268
.cursor/rules/context_gathering.mdc
Normal file
268
.cursor/rules/context_gathering.mdc
Normal file
@@ -0,0 +1,268 @@
|
||||
---
|
||||
description: Standardized patterns for gathering and processing context from multiple sources in Task Master commands, particularly for AI-powered features.
|
||||
globs:
|
||||
alwaysApply: false
|
||||
---
|
||||
# Context Gathering Patterns and Utilities
|
||||
|
||||
This document outlines the standardized patterns for gathering and processing context from multiple sources in Task Master commands, particularly for AI-powered features.
|
||||
|
||||
## Core Context Gathering Utility
|
||||
|
||||
The `ContextGatherer` class (`scripts/modules/utils/contextGatherer.js`) provides a centralized, reusable utility for extracting context from multiple sources:
|
||||
|
||||
### **Key Features**
|
||||
- **Multi-source Context**: Tasks, files, custom text, project file tree
|
||||
- **Token Counting**: Detailed breakdown using `gpt-tokens` library
|
||||
- **Format Support**: Different output formats (research, chat, system-prompt)
|
||||
- **Error Handling**: Graceful handling of missing files, invalid task IDs
|
||||
- **Performance**: File size limits, depth limits for tree generation
|
||||
|
||||
### **Usage Pattern**
|
||||
```javascript
|
||||
import { ContextGatherer } from '../utils/contextGatherer.js';
|
||||
|
||||
// Initialize with project paths
|
||||
const gatherer = new ContextGatherer(projectRoot, tasksPath);
|
||||
|
||||
// Gather context with detailed token breakdown
|
||||
const result = await gatherer.gather({
|
||||
tasks: ['15', '16.2'], // Task and subtask IDs
|
||||
files: ['src/api.js', 'README.md'], // File paths
|
||||
customContext: 'Additional context text',
|
||||
includeProjectTree: true, // Include file tree
|
||||
format: 'research', // Output format
|
||||
includeTokenCounts: true // Get detailed token breakdown
|
||||
});
|
||||
|
||||
// Access results
|
||||
const contextString = result.context;
|
||||
const tokenBreakdown = result.tokenBreakdown;
|
||||
```
|
||||
|
||||
### **Token Breakdown Structure**
|
||||
```javascript
|
||||
{
|
||||
customContext: { tokens: 150, characters: 800 },
|
||||
tasks: [
|
||||
{ id: '15', type: 'task', title: 'Task Title', tokens: 245, characters: 1200 },
|
||||
{ id: '16.2', type: 'subtask', title: 'Subtask Title', tokens: 180, characters: 900 }
|
||||
],
|
||||
files: [
|
||||
{ path: 'src/api.js', tokens: 890, characters: 4500, size: '4.5 KB' }
|
||||
],
|
||||
projectTree: { tokens: 320, characters: 1600 },
|
||||
total: { tokens: 1785, characters: 8000 }
|
||||
}
|
||||
```
|
||||
|
||||
## Fuzzy Search Integration
|
||||
|
||||
The `FuzzyTaskSearch` class (`scripts/modules/utils/fuzzyTaskSearch.js`) provides intelligent task discovery:
|
||||
|
||||
### **Key Features**
|
||||
- **Semantic Matching**: Uses Fuse.js for similarity scoring
|
||||
- **Purpose Categories**: Pattern-based task categorization
|
||||
- **Relevance Scoring**: High/medium/low relevance thresholds
|
||||
- **Context-Aware**: Different search configurations for different use cases
|
||||
|
||||
### **Usage Pattern**
|
||||
```javascript
|
||||
import { FuzzyTaskSearch } from '../utils/fuzzyTaskSearch.js';
|
||||
|
||||
// Initialize with tasks data and context
|
||||
const fuzzySearch = new FuzzyTaskSearch(tasksData.tasks, 'research');
|
||||
|
||||
// Find relevant tasks
|
||||
const searchResults = fuzzySearch.findRelevantTasks(query, {
|
||||
maxResults: 8,
|
||||
includeRecent: true,
|
||||
includeCategoryMatches: true
|
||||
});
|
||||
|
||||
// Get task IDs for context gathering
|
||||
const taskIds = fuzzySearch.getTaskIds(searchResults);
|
||||
```
|
||||
|
||||
## Implementation Patterns for Commands
|
||||
|
||||
### **1. Context-Aware Command Structure**
|
||||
```javascript
|
||||
// In command action handler
|
||||
async function commandAction(prompt, options) {
|
||||
// 1. Parameter validation and parsing
|
||||
const taskIds = options.id ? parseTaskIds(options.id) : [];
|
||||
const filePaths = options.files ? parseFilePaths(options.files) : [];
|
||||
|
||||
// 2. Initialize context gatherer
|
||||
const projectRoot = findProjectRoot() || '.';
|
||||
const tasksPath = path.join(projectRoot, 'tasks', 'tasks.json');
|
||||
const gatherer = new ContextGatherer(projectRoot, tasksPath);
|
||||
|
||||
// 3. Auto-discover relevant tasks if none specified
|
||||
if (taskIds.length === 0) {
|
||||
const fuzzySearch = new FuzzyTaskSearch(tasksData.tasks, 'research');
|
||||
const discoveredIds = fuzzySearch.getTaskIds(
|
||||
fuzzySearch.findRelevantTasks(prompt)
|
||||
);
|
||||
taskIds.push(...discoveredIds);
|
||||
}
|
||||
|
||||
// 4. Gather context with token breakdown
|
||||
const contextResult = await gatherer.gather({
|
||||
tasks: taskIds,
|
||||
files: filePaths,
|
||||
customContext: options.context,
|
||||
includeProjectTree: options.projectTree,
|
||||
format: 'research',
|
||||
includeTokenCounts: true
|
||||
});
|
||||
|
||||
// 5. Display token breakdown (for CLI)
|
||||
if (outputFormat === 'text') {
|
||||
displayDetailedTokenBreakdown(contextResult.tokenBreakdown);
|
||||
}
|
||||
|
||||
// 6. Use context in AI call
|
||||
const aiResult = await generateTextService(role, session, systemPrompt, userPrompt);
|
||||
|
||||
// 7. Display results with enhanced formatting
|
||||
displayResults(aiResult, contextResult.tokenBreakdown);
|
||||
}
|
||||
```
|
||||
|
||||
### **2. Token Display Pattern**
|
||||
```javascript
|
||||
function displayDetailedTokenBreakdown(tokenBreakdown, systemTokens, userTokens) {
|
||||
const sections = [];
|
||||
|
||||
// Build context breakdown
|
||||
if (tokenBreakdown.tasks?.length > 0) {
|
||||
const taskDetails = tokenBreakdown.tasks.map(task =>
|
||||
`${task.type === 'subtask' ? ' ' : ''}${task.id}: ${task.tokens.toLocaleString()}`
|
||||
).join('\n');
|
||||
sections.push(`Tasks (${tokenBreakdown.tasks.reduce((sum, t) => sum + t.tokens, 0).toLocaleString()}):\n${taskDetails}`);
|
||||
}
|
||||
|
||||
if (tokenBreakdown.files?.length > 0) {
|
||||
const fileDetails = tokenBreakdown.files.map(file =>
|
||||
` ${file.path}: ${file.tokens.toLocaleString()} (${file.size})`
|
||||
).join('\n');
|
||||
sections.push(`Files (${tokenBreakdown.files.reduce((sum, f) => sum + f.tokens, 0).toLocaleString()}):\n${fileDetails}`);
|
||||
}
|
||||
|
||||
// Add prompts breakdown
|
||||
sections.push(`Prompts: system ${systemTokens.toLocaleString()}, user ${userTokens.toLocaleString()}`);
|
||||
|
||||
// Display in clean box
|
||||
const content = sections.join('\n\n');
|
||||
console.log(boxen(content, {
|
||||
title: chalk.cyan('Token Usage'),
|
||||
padding: { top: 1, bottom: 1, left: 2, right: 2 },
|
||||
borderStyle: 'round',
|
||||
borderColor: 'cyan'
|
||||
}));
|
||||
}
|
||||
```
|
||||
|
||||
### **3. Enhanced Result Display Pattern**
|
||||
```javascript
|
||||
function displayResults(result, query, detailLevel, tokenBreakdown) {
|
||||
// Header with query info
|
||||
const header = boxen(
|
||||
chalk.green.bold('Research Results') + '\n\n' +
|
||||
chalk.gray('Query: ') + chalk.white(query) + '\n' +
|
||||
chalk.gray('Detail Level: ') + chalk.cyan(detailLevel),
|
||||
{
|
||||
padding: { top: 1, bottom: 1, left: 2, right: 2 },
|
||||
margin: { top: 1, bottom: 0 },
|
||||
borderStyle: 'round',
|
||||
borderColor: 'green'
|
||||
}
|
||||
);
|
||||
console.log(header);
|
||||
|
||||
// Process and highlight code blocks
|
||||
const processedResult = processCodeBlocks(result);
|
||||
|
||||
// Main content in clean box
|
||||
const contentBox = boxen(processedResult, {
|
||||
padding: { top: 1, bottom: 1, left: 2, right: 2 },
|
||||
margin: { top: 0, bottom: 1 },
|
||||
borderStyle: 'single',
|
||||
borderColor: 'gray'
|
||||
});
|
||||
console.log(contentBox);
|
||||
|
||||
console.log(chalk.green('✓ Research complete'));
|
||||
}
|
||||
```
|
||||
|
||||
## Code Block Enhancement
|
||||
|
||||
### **Syntax Highlighting Pattern**
|
||||
```javascript
|
||||
import { highlight } from 'cli-highlight';
|
||||
|
||||
function processCodeBlocks(text) {
|
||||
return text.replace(/```(\w+)?\n([\s\S]*?)```/g, (match, language, code) => {
|
||||
try {
|
||||
const highlighted = highlight(code.trim(), {
|
||||
language: language || 'javascript',
|
||||
theme: 'default'
|
||||
});
|
||||
return `\n${highlighted}\n`;
|
||||
} catch (error) {
|
||||
return `\n${code.trim()}\n`;
|
||||
}
|
||||
});
|
||||
}
|
||||
```
|
||||
|
||||
## Integration Guidelines
|
||||
|
||||
### **When to Use Context Gathering**
|
||||
- ✅ **DO**: Use for AI-powered commands that benefit from project context
|
||||
- ✅ **DO**: Use when users might want to reference specific tasks or files
|
||||
- ✅ **DO**: Use for research, analysis, or generation commands
|
||||
- ❌ **DON'T**: Use for simple CRUD operations that don't need AI context
|
||||
|
||||
### **Performance Considerations**
|
||||
- ✅ **DO**: Set reasonable file size limits (50KB default)
|
||||
- ✅ **DO**: Limit project tree depth (3-5 levels)
|
||||
- ✅ **DO**: Provide token counts to help users understand context size
|
||||
- ✅ **DO**: Allow users to control what context is included
|
||||
|
||||
### **Error Handling**
|
||||
- ✅ **DO**: Gracefully handle missing files with warnings
|
||||
- ✅ **DO**: Validate task IDs and provide helpful error messages
|
||||
- ✅ **DO**: Continue processing even if some context sources fail
|
||||
- ✅ **DO**: Provide fallback behavior when context gathering fails
|
||||
|
||||
### **Future Command Integration**
|
||||
Commands that should consider adopting this pattern:
|
||||
- `analyze-complexity` - Could benefit from file context
|
||||
- `expand-task` - Could use related task context
|
||||
- `update-task` - Could reference similar tasks for consistency
|
||||
- `add-task` - Could use project context for better task generation
|
||||
|
||||
## Export Patterns
|
||||
|
||||
### **Context Gatherer Module**
|
||||
```javascript
|
||||
export {
|
||||
ContextGatherer,
|
||||
createContextGatherer // Factory function
|
||||
};
|
||||
```
|
||||
|
||||
### **Fuzzy Search Module**
|
||||
```javascript
|
||||
export {
|
||||
FuzzyTaskSearch,
|
||||
PURPOSE_CATEGORIES,
|
||||
RELEVANCE_THRESHOLDS
|
||||
};
|
||||
```
|
||||
|
||||
This context gathering system provides a foundation for building more intelligent, context-aware commands that can leverage project knowledge to provide better AI-powered assistance.
|
||||
Reference in New Issue
Block a user