Files
signal-platform/app/config.py
T
dennisthiessen 7060b9a019
Deploy / lint (push) Successful in 6s
Deploy / test (push) Successful in 38s
Deploy / deploy (push) Successful in 25s
parallelize the backtest across worker processes (true multi-core)
The replay was CPU-bound and single-core: the earlier asyncio.to_thread offload
kept the API responsive but, because of the GIL, ran on one core. Per-ticker
replay is independent, so fan it out across worker processes (which sidestep the
GIL) for real multi-core speedup.

- New `settings.backtest_workers` (default 4), capped to cpu_count-1 so a core
  stays free for the web server.
- Uses a `forkserver` context (workers forked from a clean single-threaded
  server — avoids the fork-with-threads deadlock); falls back to `fork`. On
  spawn-only platforms (Windows) and for 1-ticker runs it uses the thread path,
  so dev/tests are unaffected.
- Worker takes primitive column arrays (cheap to pickle), rebuilds bars, and
  returns (candidates, plain-dict signal series) — both picklable across the
  process boundary. Bars are still fetched in the event loop (ORM-safe).
- Pool creation is guarded: if the pool can't start, the job falls back to the
  sequential thread path instead of failing.

334 backend tests pass (parallel path is POSIX/server-only, so it's covered by
construction + the picklability/worker-count tests; the thread fallback is
exercised by the run_backtest smoke test).

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
2026-06-23 23:20:20 +02:00

91 lines
3.3 KiB
Python

from pydantic_settings import BaseSettings, SettingsConfigDict
class Settings(BaseSettings):
model_config = SettingsConfigDict(env_file=".env", env_file_encoding="utf-8")
# Database
database_url: str = "postgresql+asyncpg://stock_backend:changeme@localhost:5432/stock_data_backend"
# Auth
jwt_secret: str = "change-this-to-a-random-secret"
jwt_expiry_minutes: int = 60
# OHLCV Provider — Alpaca Markets
alpaca_api_key: str = ""
alpaca_api_secret: str = ""
# Sentiment Provider — Gemini with Search Grounding (legacy)
gemini_api_key: str = ""
gemini_model: str = "gemini-2.0-flash"
# Sentiment Provider — OpenAI
openai_api_key: str = ""
openai_model: str = "gpt-4o-mini"
openai_sentiment_batch_size: int = 5
# Sentiment Provider — DeepSeek / xAI (OpenAI-compatible; optional env fallback)
deepseek_api_key: str = ""
xai_api_key: str = ""
# Fundamentals Provider — Financial Modeling Prep
fmp_api_key: str = ""
# Fundamentals Provider — Finnhub (optional fallback)
finnhub_api_key: str = ""
# Fundamentals Provider — Alpha Vantage (optional fallback)
alpha_vantage_api_key: str = ""
# Alerts — Telegram (optional env fallback; can also be set in Admin)
telegram_bot_token: str = ""
telegram_chat_id: str = ""
# Scheduled Jobs
data_collector_frequency: str = "daily"
sentiment_poll_interval_minutes: int = 30
# Sentiment search-budget controls (Gemini grounding free tier = 5000/month).
# Only fetch sentiment for relevant tickers (watchlist + open trades + top-N by
# composite), skip ones refreshed within fresh_hours, and cap per run.
sentiment_fresh_hours: int = 72
sentiment_max_per_run: int = 25
sentiment_top_composite: int = 30
fundamental_fetch_frequency: str = "weekly" # quarterly-ish data; weekly conserves API quota
rr_scan_frequency: str = "daily"
alerts_frequency: str = "hourly"
fundamental_rate_limit_retries: int = 3
fundamental_rate_limit_backoff_seconds: int = 15
# Pause between tickers in the bulk fundamentals job. Free tiers throttle
# hard (Finnhub ~60 calls/min, ~3 calls/ticker → ~3s/ticker); without
# spacing the job bursts straight into 429s. 0 disables.
fundamental_request_spacing_seconds: float = 3.0
# Scoring Defaults
default_watchlist_auto_size: int = 10
default_rr_threshold: float = 1.5
# Outcome evaluation: trading days before an undecided setup expires
outcome_evaluation_max_bars: int = 30
# OHLCV history depth to fetch. New tickers backfill this far; the manual
# "data_backfill" job re-fetches the full window for everyone. ~5 years so
# long-lookback factors (12-month momentum, 52-week high) and multi-regime
# backtests become computable. ~252 trading days/year.
ohlcv_history_days: int = 1825
# Backtest parallelism: replay tickers across this many worker processes on
# POSIX (forkserver), capped to cpu_count-1 so a core stays free for the web
# server. 1 disables it (sequential). No effect on Windows / spawn-only
# platforms — those fall back to a single worker thread.
backtest_workers: int = 4
# Database Pool
db_pool_size: int = 5
db_pool_timeout: int = 30
# Logging
log_level: str = "INFO"
settings = Settings()