Major upgrade making the AI loop visible and the project ready for public release.
UI / UX
- Live AI Thinking Feed: streams reasoning, decisions, and outcomes per iteration
- Parameter Changes panel: prev → new + reason for every AI-driven edit
- Validation Activity panel: out-of-sample + sensitivity runs with live metrics
- Early Termination banner: surfaces why optimization stopped (targets met, no profit, budget, stuck, user stop)
- 3-phase tracker renamed Exploration / Iteration / Validation with live N/total
- Best Result modal exposes Evolution Path showing how the AI arrived at the winner
- Run-detail modal accessible from every recent run row
- Setup form validation (dates, walk-forward order, params selection, AI targets)
- Pause button removed; misleading sidebar nav consolidated to Dashboard / New Run / Reports / Source
Backend
- AIGuidedLoop streams ai_thinking, param_changes, ai_targets_met, ai_stuck
- Pipeline emits validation_start / validation_run_start / validation_run_complete / validation_done
- Pipeline emits early_termination on every early-stop path
- /api/best_result returns best run + full evolution chain
- /api/run/<id> + /api/runs sorted by ts
- AIReasoner falls back to ANTHROPIC_API_KEY env var when config is a placeholder
- Demo mode (APEX_DEMO_MODE=1) generates deterministic synthetic backtests so judges can run end-to-end without MT5
Open-source readiness
- README.md with pitch, demo flow, architecture diagram, quickstart, event reference
- LICENSE (MIT)
- config.example.yaml template (config.yaml now git-ignored)
- requirements.txt: added anthropic / requests / psutil / beautifulsoup4, capped majors
- .gitignore: secrets, *.set, scratch screenshots, ea_registry.yaml
- demo/run_demo.py: one-command offline demo runner
- 10 polished screenshots for README + judge review
Root cause: Removing ea: from config.yaml broke 4 call sites in
optimizer_loop.py (_execute_run, _build_components) and main.py that
still read cfg['ea']. Server started but MT5 never launched.
Fixes:
config.yaml - Restored ea: block as bridge for legacy callers
optimizer_loop.py - _build_components: loads EARegistry, creates
IniBuilder(schema=schema) instead of manifest path
_execute_run: uses self._profile.{name,symbol,tf}
instead of cfg['ea']; passes profile to runner.run()
mt5/runner.py - Validation + TradeLog now use EAProfile data when
profile is provided (legacy fallback preserved)
Tested:
7/7 pre-flight checks pass (config, registry, schema, IniBuilder, runner,
MutationEngine, OptimizerLoop._build_components)
Live browser test: Start clicked → MT5 launched → logs appeared → Reports
page loaded (9 cards) → Back to dashboard navigated in-tab
System fully operational
1. CRITICAL FIX - Reports 500 Internal Server Error:
Root cause: {{}} in onclick was Jinja2 template expression
Fix: Changed to single {} in reports_index.html onclick
2. Score history chart always empty:
Root cause: score_update only emitted inside wfv.passed block
Fix: Emit score_update after every hypothesis run with {promoted:false}
Fix: Baseline run also pushes to chart via run_complete handler
3. IS gate too strict - optimizer produces 0 candidates forever:
Root cause: min_calmar=0.35 but best EA has calmar=0.165
Fix: Two-tier IS check - passes if composite_score improves vs baseline
(absolute thresholds still apply as alternative pass condition)
4. Findings emitted twice (double UI display):
Root cause: baseline analyzed once after run, then re-analyzed in iter 1
Fix: Cache baseline findings, reuse in iteration 1 without re-emitting
5. Chart labels improved:
'Baseline' for first point, 'It1·h1' for hypothesis runs
Calmar normalized -0.5..2.0 -> 0..1 for chart display
6. Best score header only updates on actual promotion (not per-hypothesis)
Reports 500 (Internal Server Error):
- app.py: Replace NaN/Infinity->null before json.loads() in both
/reports and /api/runs routes. Old summary.json files with invalid
NaN are now handled transparently.
- reports/writer.py: Added _safe_val() helper using math.isfinite().
All summary.json values now use safe rounding (NaN->0, Inf->0).
No new hypotheses (optimizer stops after 1 iter):
- optimizer_loop.py: Changed dedup from DB-wide history to
session-scoped (self.session_tested_deltas). Previous runs from
old sessions no longer block new hypothesis proposals.
- Each tested hypothesis is tracked within the session only.
Result: optimizer now runs multiple iterations, scores are valid
numbers, reports page loads without 500.
- NaN composite score: dropna().mean() on empty series returns NaN,
nan is truthy so (nan or 0)=nan. Fixed: use None when no MFE data
so scorer uses 0.5 fallback instead of propagating NaN.
- Analysis stopping: gate WFV was importing from main.execute_run
(old CLI code that uses broken runner). Fixed: gate.run_walk_forward
now accepts an executor callable from optimizer_loop.
- Back to Dashboard opened blank tab: Reports button used
window.open(_blank). Fixed: same-tab navigation + history.back().
- NaN in reports card: downstream of composite_score NaN above.
LEGSTECH Optimizer