2a4950d8a0
소형 건설업체(100억 미만)를 위한 AI 기반 토목공사 통합관리 플랫폼 Backend (FastAPI): - SQLAlchemy 모델 13개 (users, projects, wbs, tasks, daily_reports, reports, inspections, quality, weather, permits, rag, settings) - API 라우터 11개 (auth, projects, tasks, daily_reports, reports, inspections, weather, rag, kakao, permits, settings) - Services: Claude AI 래퍼, CPM Gantt 계산, 기상청 API, RAG(pgvector), 카카오 Skill API - Alembic 마이그레이션 (pgvector 포함) - pytest 테스트 (CPM, 날씨 경보) Frontend (Next.js 15): - 11개 페이지 (대시보드, 프로젝트, Gantt, 일보, 검측, 품질, 날씨, 인허가, RAG, 설정) - TanStack Query + Zustand + Tailwind CSS 인프라: - Docker Compose (PostgreSQL pgvector + backend + frontend) - 한국어 README 및 설치 가이드 Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
35 lines
982 B
Python
35 lines
982 B
Python
"""AI-powered inspection request generation."""
|
|
import json
|
|
from app.services.ai_engine import complete_json
|
|
from app.services.prompts.inspection import SYSTEM_PROMPT, build_prompt
|
|
|
|
|
|
async def generate_checklist(
|
|
project_name: str,
|
|
inspection_type: str,
|
|
location_detail: str | None,
|
|
requested_date: str,
|
|
wbs_name: str | None,
|
|
) -> list[dict]:
|
|
"""Generate inspection checklist items using Claude."""
|
|
prompt = build_prompt(
|
|
project_name=project_name,
|
|
inspection_type=inspection_type,
|
|
location_detail=location_detail,
|
|
requested_date=requested_date,
|
|
wbs_name=wbs_name,
|
|
)
|
|
|
|
raw = await complete_json(
|
|
messages=[{"role": "user", "content": prompt}],
|
|
system=SYSTEM_PROMPT,
|
|
temperature=0.2,
|
|
)
|
|
|
|
try:
|
|
data = json.loads(raw)
|
|
return data.get("checklist_items", [])
|
|
except (json.JSONDecodeError, KeyError):
|
|
# Fallback: return empty checklist
|
|
return []
|