/sac-optimize

name: sac-optimize
description: Analyze SAC scripts for performance optimization opportunities
allowed-tools:
  - Read
  - Grep
  - Glob
argument-hint: "[script-file-or-symptom]"

Shell Snippet Notes

• Shell snippets assume Bash on Linux/macOS, WSL2, or Git Bash.
• Install the command-specific tooling shown near each snippet before running it.
• Confirm before running commands that delete files, change ownership, deploy, or modify remote systems.

Output Contract

Return performance findings, backend-call risks, caching or batching suggestions, and validation steps. Default to read-only analysis and do not edit scripts.

Analyze SAC scripts for performance issues and provide optimization recommendations.

Performance Analysis Workflow

Step 1: Collect Script Information

Ask the user to provide: 1. **Scripts to analyze** - paste code or describe functionality 2. **Performance symptoms** - slow load, laggy interactions, timeouts? 3. **Number of data points** - how much data is being processed? 4. **User complaints** - specific slow operations?

Step 2: Performance Anti-Patterns Checklist

High-Impact Issues (Fix First)

**1. getMembers() without accessMode**

// SLOW: Hits backend every time
var members = ds.getMembers("Location");

// FAST: Uses cached booked values
var members = ds.getMembers("Location", {
    accessMode: MemberAccessMode.BookedValues
});

**Impact**: Each call = 1 backend roundtrip (~200-500ms)

**2. Missing setRefreshPaused() Batching**

// SLOW: 3 separate backend calls
ds.setDimensionFilter("Dim1", value1);
ds.setDimensionFilter("Dim2", value2);
ds.setDimensionFilter("Dim3", value3);

// FAST: 1 backend call
ds.setRefreshPaused(true);
ds.setDimensionFilter("Dim1", value1);
ds.setDimensionFilter("Dim2", value2);
ds.setDimensionFilter("Dim3", value3);
ds.setRefreshPaused(false);

**Impact**: Reduces N calls to 1 call

**3. Heavy onInitialization Code**

// BAD: Slow startup
// onInitialization
var members = Chart_1.getDataSource().getMembers("Location");
members.forEach(function(m) {
    // Heavy processing
});

// GOOD: Empty or minimal initialization
// onInitialization
// (keep empty - defer to lazy loading)

**Impact**: Blocks initial render

Medium-Impact Issues

**4. getData() Instead of getResultSet()**

// SLOWER: May hit backend
var data = ds.getData();

// FASTER: Uses cached result set
var resultSet = ds.getResultSet();

**5. Repeated getDataSource() Calls**

// INEFFICIENT: Multiple lookups
Chart_1.getDataSource().setDimensionFilter("A", "1");
Chart_1.getDataSource().setDimensionFilter("B", "2");

// BETTER: Cache reference
var ds = Chart_1.getDataSource();
ds.setDimensionFilter("A", "1");
ds.setDimensionFilter("B", "2");

**6. Backend Calls in Loops**

// TERRIBLE: N backend calls
for (var i = 0; i < items.length; i++) {
    ds.getMembers(items[i]); // Backend call each iteration!
}

// BETTER: Single call, process in memory
var resultSet = ds.getResultSet();
// Process resultSet in memory

Step 3: Performance Optimization Patterns

Pattern 1: Efficient Data Access

// For reading dimension values
var ds = Chart_1.getDataSource();

// Option A: From result set (fastest, cached)
var resultSet = ds.getResultSet();
var uniqueValues = {};
resultSet.forEach(function(row) {
    uniqueValues[row["Location"].id] = row["Location"].description;
});

// Option B: Booked values only (fast, cached)
var members = ds.getMembers("Location", {
    accessMode: MemberAccessMode.BookedValues
});

// Option C: All master data (slowest, hits backend)
// Only use when you need unbooked values
var allMembers = ds.getMembers("Location", {
    accessMode: MemberAccessMode.MasterData
});

Pattern 2: Batched Filter Updates

function applyMultipleFilters(ds, filters) {
    ds.setRefreshPaused(true);

    try {
        for (var dim in filters) {
            if (filters[dim] === null) {
                ds.removeDimensionFilter(dim);
            } else {
                ds.setDimensionFilter(dim, filters[dim]);
            }
        }
    } finally {
        ds.setRefreshPaused(false); // Always unpause
    }
}

// Usage
applyMultipleFilters(Chart_1.getDataSource(), {
    "Location": "US",
    "Product": ["A", "B", "C"],
    "Year": null  // Remove this filter
});

Pattern 3: Lazy Loading

// Script variable to track initialization
var initialized = false;

// In onResultChanged or first interaction
if (!initialized) {
    // Perform one-time heavy setup
    initialized = true;
}

Pattern 4: Debounced Updates

// For rapid user input (e.g., typing in search)
var debounceTimer = null;

function onSearchInput(value) {
    if (debounceTimer) {
        clearTimeout(debounceTimer);
    }

    debounceTimer = setTimeout(function() {
        // Actual filter operation
        ds.setDimensionFilter("Search", value);
    }, 300); // Wait 300ms after last input
}

Step 4: Performance Metrics

Enable Performance Logging

Add URL parameter: `?APP_PERFORMANCE_LOGGING=true`

This shows:

• Script execution times
• Backend call durations
• Widget render times

Manual Timing

console.time("operationName");
// ... operation ...
console.timeEnd("operationName"); // Logs duration

Step 5: Optimization Report Template

**Current State Analysis:**

• onInitialization: [Empty/Light/Heavy]
• Backend calls per interaction: [Count]
• Main bottlenecks: [List]

**Recommendations:**

| Priority | Issue | Current Impact | Fix | Expected Improvement | |----------|-------|---------------|-----|---------------------| | High | getMembers without accessMode | 5 calls/action | Add BookedValues | -80% calls | | High | No refresh batching | 3 calls/filter | Add setRefreshPaused | -66% calls | | Medium | Heavy onInit | 2s startup | Move to lazy load | -1.5s start