PolStock

PolStock is a short-term trading signal assistant for Indonesian stocks.

Its job is to help answer:

Which tickers are bullish right now, backed by strong participation, and tradable over the next 1–14 days?

The project is now being refocused away from a news-first political-impact dashboard into a short-swing signal engine where market structure and participation lead, and news/policy context acts as an optional conviction boost.

Product direction

PolStock v1 should prioritize:

• actionable long setups over broad event summaries
• 1d / 3d / 7d / 14d holding windows
• bullish continuation or rebound setups over generic sentiment labels
• participation-backed signals using volume and traded-value proxies
• clear trade plans: entry, stop-loss, take-profit, invalidation, reasons

v1 scope

Supported signal types

• breakout continuation

  • price pushes through / above resistance
  • bullish momentum confirmation
  • participation expansion required

• support rebound

  • price near support
  • oversold-to-recovery behavior
  • momentum improving

• squeeze breakout watch

  • volatility compression / Bollinger squeeze
  • bullish pressure building
  • upgraded to BUY only when trigger + participation confirm

• news-accelerated breakout

  • a valid bullish technical setup gets extra conviction from relevant news/policy context
  • news alone must not create a BUY signal

v1 signal actions

• BUY — bullish setup complete, participation confirmed, risk-defined
• WATCH — setup forming, but trigger or participation is incomplete
• IGNORE — no actionable edge

v1 horizon targets

• 1d — urgent breakout / ignition setup
• 3d — fresh continuation setup
• 7d — standard short swing
• 14d — slower but still actionable bullish development

v1 non-goals

These are intentionally out of scope for the first usable version:

• production-grade short-selling / bearish strategy optimization
• news-only BUY signals without technical confirmation
• ranking stocks primarily by political relevance instead of trade actionability
• pretending transaction-count data exists when the data source only supports volume/value proxies

Current architecture focus

PolStock still includes:

• on-demand refresh with /api/refresh
• Indonesian market/news ingestion
• stock quote + technical indicator fetching
• event/news analysis and policy-context enrichment
• signal history, snapshots, backtest infrastructure, and dashboard/bot surfaces

But the primary decision layer is being shifted toward:

Bullish market structure
+ momentum confirmation
+ participation expansion
+ trigger quality
+ optional event/news boost
= short-term trade signal

Run locally

cd /opt/hermes/politics_stock_mapper
python3 app.py

Then open http://127.0.0.1/.

API

Core endpoints currently in the backend include:

• GET / – dashboard
• GET /healthz – health check
• GET /api/dashboard – consolidated dashboard payload
• POST /api/refresh – fetch + analyze + render-ready JSON
• GET /api/signals/daily-summary – grouped daily signal summary
• GET /api/signals/ticker/{ticker} – ticker-level signal explanation
• GET /api/signals/history – persisted signal history
• GET /api/backtest – backtest metrics
• GET /api/watchlist / PUT /api/watchlist – watchlist state

Documentation

• docs/short-term-signal-spec.md — current product spec for the short-term signal pivot
• docs/SPEC.md — broader legacy/refocus analysis and implementation inventory
• docs/ARCHITECTURE.md — current system architecture overview (will be updated further as the pivot lands)

Notes

• The app uses live public data and caches refresh results for a short TTL.
• If transaction-count data is unavailable from the upstream source, v1 uses volume and traded-value proxies and labels them honestly.
• Relevant news/policy context still matters, but it should boost a valid setup rather than create one on its own.
• The project should not claim a setup is “actionable” unless the signal has clear entry, risk, and holding-window logic.