forked from LengKundee/MQL5-Google-Onedrive
Profiled scripts/web_dashboard.py and identified redundant disk I/O and markdown parsing as a primary latency bottleneck. Implemented a MarkdownCache class with: - Lazy loading to avoid I/O on cold starts. - mtime-based tracking to skip re-rendering unless the source file changes. - Pre-calculated module-level constants for file paths. Performance Impact: - Baseline Latency: 33.48ms - Optimized Latency: 7.58ms (~77% improvement) Verified via: - scripts/benchmark_dashboard.py (custom benchmark) - scripts/ci_validate_repo.py - scripts/test_automation.py - Playwright frontend verification script (dashboard.png) |
||
|---|---|---|
| .. | ||
| bolt.md | ||
| deploy.yml | ||