feat: containerize for mmd.i80.dk deployment via Gitea/Nomad

- Add Dockerfile (python:3.12-slim, HF_HOME=/app/data/hf-cache)
- Add mmd.nomad (multi-task: web=dashboard, worker=scheduler)
- Add .gitea/workflows/deploy.yml (build->Harbor, deploy->Nomad)
- Add scheduler.py (stdlib scheduler replaces cron in container)
- Add requirements.txt
- dashboard.py: LOG_DIR + PORT/HOST from env vars
- saxo_auth.py: TOKEN_FILE from SAXO_TOKEN_FILE env var
- .gitignore: proper ignores for container project

Volume moneymaker-data (/app/data) holds:
  - logs/ (shared between web+worker)
  - .saxo_token.json (pre-copy once after first deploy)
  - hf-cache/ (HuggingFace FinBERT cache)

Gitea secrets required: DATABASE_URL, ANTHROPIC_API_KEY,
SAXO_APP_KEY, SAXO_APP_SECRET_1, HARBOR_ROBOT_TOKEN

scheduler.py

@@ -0,0 +1,86 @@
+"""
+scheduler.py — MoneyMaker worker daemon.
+
+Runs the pipeline at scheduled times (Mon-Fri UTC):
+  04:00  analyze-only  (NLP before Copenhagen market open 09:00 CET)
+  07:30  trade window 1
+  10:00  trade window 2
+  12:30  trade window 3
+  14:30  trade window 4
+  17:00  daily P&L report
+
+Outputs appended to LOG_DIR/runner_YYYY-MM-DD.log so the dashboard can
+show a live log tail. Uses only stdlib — no extra scheduler dependency.
+"""
+import os
+import sys
+import time
+import traceback
+from datetime import datetime, timezone
+from pathlib import Path
+
+LOG_DIR = Path(os.getenv("LOG_DIR", str(Path(__file__).parent / "data" / "logs")))
+LOG_DIR.mkdir(parents=True, exist_ok=True)
+
+# hour, minute, weekdays (0=Mon…4=Fri), analyze_only, is_report
+SCHEDULE = [
+    (4,  0,  frozenset(range(5)), True,  False),
+    (7,  30, frozenset(range(5)), False, False),
+    (10, 0,  frozenset(range(5)), False, False),
+    (12, 30, frozenset(range(5)), False, False),
+    (14, 30, frozenset(range(5)), False, False),
+    (17, 0,  frozenset(range(5)), False, True),
+]
+
+
+def _log_path() -> Path:
+    today = datetime.now(timezone.utc).strftime("%Y-%m-%d")
+    return LOG_DIR / f"runner_{today}.log"
+
+
+def _run_job(analyze_only: bool, is_report: bool) -> None:
+    log = _log_path()
+    with open(log, "a", buffering=1) as fh:
+        old_out, old_err = sys.stdout, sys.stderr
+        sys.stdout = sys.stderr = fh
+        try:
+            if is_report:
+                from report import print_report
+                print_report()
+            else:
+                from runner import run_pipeline
+                run_pipeline(analyze_only=analyze_only)
+        except Exception as exc:
+            print(f"[scheduler] ERROR: {exc}")
+            traceback.print_exc()
+        finally:
+            sys.stdout = old_out
+            sys.stderr = old_err
+
+
+def main() -> None:
+    last_run: dict = {}
+    print(f"[scheduler] started — LOG_DIR={LOG_DIR}", flush=True)
+
+    while True:
+        now = datetime.now(timezone.utc)
+        dow = now.weekday()
+
+        for hour, minute, days, analyze_only, is_report in SCHEDULE:
+            if dow not in days:
+                continue
+            key = (hour, minute, now.date())
+            if key in last_run:
+                continue
+            if now.hour == hour and now.minute == minute:
+                last_run[key] = True
+                label = "report" if is_report else ("analyze-only" if analyze_only else "full")
+                print(f"[scheduler] {now.strftime('%H:%M UTC')} — {label}", flush=True)
+                _run_job(analyze_only, is_report)
+                print(f"[scheduler] {label} done", flush=True)
+
+        time.sleep(30)
+
+
+if __name__ == "__main__":
+    main()