chore: 문서를 작성할 때 한국어의 표현 작성 스킬 추가 및 1인칭 관점의 글 작성 검증 테스트 추가

2026-07-29 16:48:03 +09:00
parent c39406bbdd
commit 41501b5d06
520 changed files with 95494 additions and 2231 deletions
@@ -0,0 +1,120 @@
+from __future__ import annotations
+
+from typing import Any
+
+from claridoc.models import Brief, Outline, Source, SourcePack
+
+
+def build_evidence_map(brief: Brief, outline: Outline, sources: SourcePack) -> dict[str, Any]:
+    source_by_id = {source.id: source for source in sources.sources}
+    sections: list[dict[str, Any]] = []
+    for section in outline.sections:
+        evidence = []
+        for source_id in section.evidence_ids:
+            source = source_by_id.get(source_id)
+            if source is None:
+                continue
+            evidence.append(_source_record(source))
+        sections.append(
+            {
+                "section_id": section.id,
+                "intent": section.intent,
+                "title": section.title,
+                "reader_question": section.reader_question,
+                "decision_requirements": section.decision_requirements,
+                "evidence": evidence,
+                "evidence_gap": bool(section.decision_requirements and not evidence),
+            }
+        )
+    return {
+        "schema_version": 2,
+        "document": brief.title,
+        "citation_style": brief.constraints.citation_style,
+        "reader_document_contains_internal_source_ids": brief.constraints.citation_style == "source_id",
+        "sections": sections,
+        "sources": [_source_record(source) for source in sources.sources],
+    }
+
+
+def render_provenance(brief: Brief, outline: Outline, sources: SourcePack) -> str:
+    source_by_id = {source.id: source for source in sources.sources}
+    lines = [
+        "# Evidence and decision provenance",
+        "",
+        "> This is an internal sidecar. It is not reader-facing article content.",
+        "> Source IDs, repository paths, line ranges, status labels, and access dates belong here—not in `document.md`.",
+        "",
+        f"- Document: **{brief.title}**",
+        f"- Citation rendering: `{brief.constraints.citation_style}`",
+        f"- Evidence sources: **{len(sources.sources)}**",
+        "",
+        "## Section evidence map",
+        "",
+        "| Section | Decision contract | Evidence | Status / location |",
+        "|---|---|---|---|",
+    ]
+    for section in outline.sections:
+        decision = ", ".join(section.decision_requirements) if section.decision_requirements else "—"
+        if not section.evidence_ids:
+            lines.append(f"| {escape(section.title)} | {escape(decision)} | **GAP** | No allocated evidence |")
+            continue
+        for position, source_id in enumerate(section.evidence_ids):
+            source = source_by_id.get(source_id)
+            if source is None:
+                lines.append(f"| {escape(section.title)} | {escape(decision)} | `{source_id}` | Unknown source |")
+                continue
+            section_name = section.title if position == 0 else "↳"
+            location = _location(source)
+            status = source.status or "unspecified"
+            lines.append(
+                f"| {escape(section_name)} | {escape(decision if position == 0 else '—')} | "
+                f"`{source.id}` {escape(source.title)} | `{escape(status)}` · {escape(location)} |"
+            )
+    lines.extend(["", "## Source details", ""])
+    for source in sources.sources:
+        lines.extend(
+            [
+                f"### `{source.id}` {source.title}",
+                "",
+                f"- Type: `{source.source_type}`",
+                f"- Status: `{source.status or 'unspecified'}`",
+                f"- Location: `{_location(source)}`",
+                f"- Public/reference URL: `{source.url}`",
+                f"- Claim IDs: {', '.join(f'`{item}`' for item in source.claim_ids) or '—'}",
+                f"- Decision IDs: {', '.join(f'`{item}`' for item in source.decision_ids) or '—'}",
+                f"- Retrieval priority: `{source.priority:.4f}`",
+                "",
+            ]
+        )
+    return "\n".join(lines).rstrip() + "\n"
+
+
+def _source_record(source: Source) -> dict[str, Any]:
+    return {
+        "id": source.id,
+        "title": source.title,
+        "source_type": source.source_type,
+        "status": source.status,
+        "path": source.path,
+        "heading": source.heading,
+        "line_start": source.line_start,
+        "line_end": source.line_end,
+        "url": source.url,
+        "accessed": source.accessed,
+        "claim_ids": list(source.claim_ids),
+        "decision_ids": list(source.decision_ids),
+        "priority": source.priority,
+    }
+
+
+def _location(source: Source) -> str:
+    location = source.path or source.url
+    if source.heading:
+        location += f" — {source.heading}"
+    if source.line_start is not None:
+        location += f" (lines {source.line_start}-{source.line_end or source.line_start})"
+    return location
+
+
+def escape(value: str) -> str:
+    return value.replace("|", "\\|").replace("\n", " ")