"""Daily digest 빌더 (Markdown / HTML). 규칙: 변화 없는 항목은 표시하지 않는다. """ from __future__ import annotations import html from typing import Any, Dict, List from .classifier import classify_bucket from .summarizer import summarize DISCLAIMER_KO = ( "본 도구는 연구·교육 목적의 참고용이며, 임상시험 등록·MASH 처방·임상 의사결정 " "근거로 사용할 수 없습니다. 모든 데이터는 합성(synthetic)입니다." ) SECTIONS = ( ("ctg", "ClinicalTrials.gov diff (synthetic)"), ("fda", "FDA news diff (synthetic)"), ("ema", "EMA EPAR (synthetic)"), ("pmda", "PMDA news (synthetic)"), ("ir", "Sponsor IR / SEC 8-K (synthetic)"), ("aasld", "AASLD The Liver Meeting 2026 abstracts (synthetic)"), ) def _enrich(buckets: Dict[str, List[Dict[str, Any]]]) -> Dict[str, List[Dict[str, Any]]]: return {key: classify_bucket(events) for key, events in buckets.items()} def _affected_drugs(event: Dict[str, Any]) -> List[str]: if event.get("drug"): return [event["drug"]] return list(event.get("drugs") or []) def _has_any_event(buckets: Dict[str, List[Dict[str, Any]]]) -> bool: return any(len(v) > 0 for v in buckets.values()) def build_markdown(buckets: Dict[str, List[Dict[str, Any]]], snapshot_date: str) -> str: enriched = _enrich(buckets) lines: List[str] = [] lines.append(f"# MASHTrialPulse Daily Digest — {snapshot_date}") lines.append("") lines.append(f"> Disclaimer: {DISCLAIMER_KO}") lines.append("") if not _has_any_event(enriched): lines.append("_오늘은 변화가 감지되지 않았습니다 (no diff events)._") lines.append("") return "\n".join(lines) for key, title in SECTIONS: events = enriched.get(key, []) if not events: continue lines.append(f"## {title}") lines.append("") for ev in events: tags = ", ".join(ev.get("tags", [])) drugs = ", ".join(_affected_drugs(ev)) or "(약물 미상)" ko = summarize(ev, snapshot_date) id_str = ev.get("nct_id") or ev.get("id") or "" lines.append(f"- **[{tags}]** {ko}") lines.append(f" - 영향 약물: {drugs}") if id_str: lines.append(f" - id: {id_str}") if ev.get("detail"): lines.append(f" - detail: {ev['detail']}") lines.append("") lines.append("---") lines.append("") lines.append("### 출처 (sources, all synthetic)") lines.append("") lines.append("- ClinicalTrials.gov API v2 (synthetic snapshot)") lines.append("- FDA NewsRoom (synthetic)") lines.append("- EMA EPAR (synthetic)") lines.append("- PMDA news (synthetic)") lines.append("- Sponsor IR pages + SEC EDGAR 8-K (synthetic)") lines.append("- AASLD The Liver Meeting 2026 abstracts (synthetic)") lines.append("") return "\n".join(lines) def build_html(buckets: Dict[str, List[Dict[str, Any]]], snapshot_date: str) -> str: enriched = _enrich(buckets) parts: List[str] = [] parts.append("") parts.append(f"MASHTrialPulse {html.escape(snapshot_date)}") parts.append("") parts.append(f"

MASHTrialPulse Daily Digest — {html.escape(snapshot_date)}

") parts.append(f"

Disclaimer: {html.escape(DISCLAIMER_KO)}

") if not _has_any_event(enriched): parts.append("

오늘은 변화가 감지되지 않았습니다 (no diff events).

") parts.append("") return "".join(parts) for key, title in SECTIONS: events = enriched.get(key, []) if not events: continue parts.append(f"

{html.escape(title)}

") for ev in events: tags = ev.get("tags", []) drugs = ", ".join(_affected_drugs(ev)) or "(약물 미상)" ko = summarize(ev, snapshot_date) id_str = ev.get("nct_id") or ev.get("id") or "" parts.append("
") for t in tags: parts.append(f"{html.escape(t)}") parts.append(f" {html.escape(ko)}") parts.append( f"
영향 약물: {html.escape(drugs)}" + (f" · id: {html.escape(id_str)}" if id_str else "") + (f" · detail: {html.escape(ev.get('detail', ''))}" if ev.get("detail") else "") + "
" ) parts.append("
") parts.append("
") parts.append("

출처 (sources, all synthetic)

") return "".join(parts)