tkrmagid
|
2c42c1151c
|
fix(backend): accelerate 0.30.1 → 0.34.2 (chronos-forecasting 호환)
chronos-forecasting 1.4.1 의 의존성: `accelerate>=0.32,<1`.
기존 핀 `accelerate==0.30.1` 이 해당 범위 밖이라 pip 가
ResolutionImpossible 로 실패했음.
검증: python:3.11-slim 컨테이너에서
pip install --dry-run -r reqs.txt (torch==2.3.1 포함, 30개 deps 전체)
exit 0. resolve 결과:
chronos-forecasting 1.4.1
transformers 4.41.2
tokenizers 0.19.1
accelerate 0.34.2
torch 2.3.1
|
2026-05-21 01:48:13 +09:00 |
|
tkrmagid
|
4fb6cec383
|
feat(phase-6): Next.js UI + TypeScript strict + 백엔드 mypy 설정
UI:
- web/lib/api.ts: 백엔드 모든 엔드포인트의 클라이언트 + 타입 (Symbol,
ChartPayload, PredictResponse, LatestPredictionResponse, MetricsResponse,
NewsResponse). NEXT_PUBLIC_API_BASE 자동 정규화.
- web/components/SearchBox: 디바운스 검색, seed_only 토글, trigram + prefix.
- web/components/StockChart: lightweight-charts 캔들 + 예측 overlay
(median dashed + q10/q90 점선). base_date 에서 target_date 까지 이어 붙임.
- web/components/PredictionPanel: "예상차트 보기" 버튼 → POST /api/predict
→ user_triggered=TRUE 저장 → onResult 콜백으로 StockChart 에 반영.
표로 +1/+3/+5거래일 direction, prob_up/flat/down, expected_return,
ci_low~ci_high 표시.
- web/components/MetricsPanel: 최근 30일 hit_rate / mae.
- web/components/NewsList: 최근 뉴스 + 감성 라벨/점수.
- web/app/page.tsx: 검색 페이지.
- web/app/[code]/page.tsx: 종목 상세 (차트 + 패널 + 메트릭 + 뉴스).
TypeScript 보강 (사용자 요청 "typescript도 추가해서 나중에 수정하기 쉽게"):
- tsconfig.json: strict 외에 forceConsistentCasingInFileNames,
noFallthroughCasesInSwitch, noImplicitOverride 추가.
- package.json: typecheck (tsc --noEmit), check (typecheck + lint) 스크립트,
eslint + eslint-config-next 14.2.3.
- .eslintrc.json: next/core-web-vitals.
- package-lock.json 커밋 (재현 가능한 dep).
백엔드:
- pyproject.toml: [tool.mypy] 추가. strict_optional, no_implicit_optional,
check_untyped_defs. 3rd-party stub 없는 pykrx/chronos 등은 ignore.
검증: `npx tsc --noEmit` 통과 (exit=0).
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
2026-05-20 16:10:24 +09:00 |
|
tkrmagid
|
b1ca6ab5d3
|
feat(phase-3): Chronos zero-shot 예측 + 피처 빌더
- backend/app/models/chronos.py: amazon/chronos-t5-small (env CHRONOS_MODEL
override 가능). lazy singleton, cuda + bf16 자동, q10/median/q90 + raw
samples 반환 (앙상블 가중평균용).
- backend/app/models/features.py: 종목별 학습/추론 피처 DataFrame.
OHLCV + TA(rsi/macd/atr/bb/sma/ema/vol_z) + 외인기관거래대금 + macro
(kospi/kosdaq/usdkrw/us10y + r1) + sentiment(v_sentiment_daily, 3d rolling).
학습용은 with_targets=True 로 y_close_h{1,3,5}, y_ret_h*, y_dir_h*
(±0.3% flat band) 추가.
- pyproject.toml: chronos-forecasting 1.4.1, accelerate 0.30.1, joblib 1.4.2.
이 단계까지는 코드만. 실제 모델 다운로드는 첫 ping/predict 호출 시점에.
Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
|
2026-05-20 15:59:14 +09:00 |
|
tkrmagid
|
cacddf5adf
|
feat(phase-0): scaffold backend + web + docker + DB schema
- docker-compose.yml: timescaledb-ha (timescaledb 2.27 + vectorscale + pgvector + pgai)
+ backend (FastAPI, CUDA 12.1) + web (Next.js 14)
- docker-compose.gpu.yml: GPU profile overlay for RTX 3070 Ti
- build.bat: Windows bootstrap, auto-detects nvidia-smi and selects GPU/CPU compose
- backend: Dockerfile, pyproject.toml, FastAPI skeleton with /health and /health/db
- DB migration 001_init.sql: symbols (with trigram search), ohlcv_daily/1m (hypertables),
macro_daily, trading_value_daily, news (vector embedding), predictions
(with user_triggered flag for on-demand UX), prediction_outcomes, model_performance
- web: Next.js 14 + Tailwind + lightweight-charts placeholder
- README: KIS/DART/HuggingFace token issuance guides + 10 seed tickers + run instructions
|
2026-05-20 14:37:35 +09:00 |
|