feat: 结题报告初稿助手首版 v1.0.1（draft-closing-report-default）

scripts/service/draft_service.py

@@ -0,0 +1,354 @@
+"""结题报告初稿生成内核（唯一业务内核）。
+
+五步组装（对齐《核心建设内容：各部门"Skill"深度需求规格书》科研处 Skill 002）：
+  C1 指标对比：按成果类型汇总实际数量 vs 预期指标 → 超额完成/持平/未达标/数据待补
+  C2 完成情况章节：逐指标成文 + 成果明细（来源标注）
+  C3 决算转换：按科目汇总金额与占比，生成规范描述句（金额不改写）
+  C4 素材要点：代表性成果与年度报告摘句逐条提取（标注来源）
+  C5 骨架组装：通用结题结构，叙事章节只放素材要点 + [待补充] 指引
+  W1 数据完整性：缺数据生成占位，不阻塞整稿
+
+红线：初稿中每个数字都来自台账；严禁编造数据与虚构论文结论。
+"""
+
+from __future__ import annotations
+
+import json
+import uuid
+from datetime import date
+from typing import Any, Optional
+
+from db.connection import get_conn, init_db
+from util.logging_config import get_skill_logger
+from util.timeutil import now_unix
+
+_PLACEHOLDER = "[待补充：请结合素材要点与原始报告撰写，注意引用本人成果]"
+
+_EXPENSE_SENTENCE_TEMPLATES = {
+    "设备费": "购置/使用科研设备",
+    "材料费": "采购实验材料",
+    "测试化验加工费": "开展测试化验与加工",
+    "燃料动力费": "支付燃料动力费用",
+    "差旅费": "赴外开展调研与学术交流",
+    "会议费": "组织/参加学术会议",
+    "国际合作与交流费": "开展国际合作与交流",
+    "出版/文献/信息传播/知识产权事务费": "支付出版与知识产权事务费用",
+    "劳务费": "支付参与项目的研究生与临时聘用人员劳务",
+    "专家咨询费": "支付专家咨询费用",
+    "其他": "其他与项目研究直接相关的支出",
+}
+
+
+def _log():
+    try:
+        return get_skill_logger()
+    except RuntimeError:
+        return None
+
+
+def _wan(value: Optional[float]) -> str:
+    if value is None:
+        return "—"
+    return f"{value / 10000:.2f} 万元"
+
+
+def _camel(name: str) -> str:
+    parts = str(name).split("_")
+    if len(parts) <= 1:
+        return str(name)
+    head, *rest = parts
+    return head + "".join(p[:1].upper() + p[1:] for p in rest if p)
+
+
+def _snake(name: str) -> str:
+    return "".join("_" + ch.lower() if ch.isupper() else ch for ch in str(name)).lstrip("_")
+
+
+def _norm_project(raw: dict[str, Any]) -> dict[str, Any]:
+    aliases = {
+        "projectNo": ("project_no", "项目编号"),
+        "projectName": ("project_name", "项目名称"),
+        "piName": ("pi_name", "负责人"),
+        "department": ("department", "学院", "所属学院"),
+        "program": ("program", "项目类型", "资助渠道"),
+        "startDate": ("start_date", "执行期开始", "开始日期"),
+        "endDate": ("end_date", "执行期结束", "结束日期"),
+        "expected": ("expected_json", "预期成果指标", "预期指标"),
+        "annualReportExcerpt": ("annual_report_excerpt", "年度报告摘句", "年度报告"),
+        "status": ("status", "状态"),
+    }
+    out: dict[str, Any] = {}
+    for key, value in raw.items():
+        out.setdefault(str(key), value)
+        for canonical, names in aliases.items():
+            if str(key) in names or _camel(str(key)) == canonical:
+                out.setdefault(canonical, value)
+    return out
+
+
+def _norm_achievement(raw: dict[str, Any]) -> dict[str, Any]:
+    aliases = {
+        "projectNo": ("project_no", "项目编号"),
+        "achievementType": ("achievement_type", "成果类型", "类型"),
+        "title": ("title", "成果名称", "标题"),
+        "venueOrNo": ("venue_or_no", "期刊/专利号", "期刊", "专利号"),
+        "achDate": ("ach_date", "日期", "发表日期"),
+        "isRepresentative": ("is_representative", "代表性成果"),
+        "note": ("note", "备注"),
+    }
+    out: dict[str, Any] = {}
+    for key, value in raw.items():
+        out.setdefault(str(key), value)
+        for canonical, names in aliases.items():
+            if str(key) in names or _camel(str(key)) == canonical:
+                out.setdefault(canonical, value)
+    return out
+
+
+def _norm_expense(raw: dict[str, Any]) -> dict[str, Any]:
+    aliases = {
+        "projectNo": ("project_no", "项目编号"),
+        "category": ("category", "科目"),
+        "amount": ("amount", "金额"),
+        "occurredDate": ("occurred_date", "发生日期"),
+        "description": ("description", "原始摘要", "描述"),
+    }
+    out: dict[str, Any] = {}
+    for key, value in raw.items():
+        out.setdefault(str(key), value)
+        for canonical, names in aliases.items():
+            if str(key) in names or _camel(str(key)) == canonical:
+                out.setdefault(canonical, value)
+    return out
+
+
+def _c1_compare(expected: list[dict[str, Any]], achievements: list[dict[str, Any]]) -> list[dict[str, Any]]:
+    """C1：按类型对比预期指标与实际成果。"""
+    counts: dict[str, int] = {}
+    for ach in achievements:
+        ach_type = str(ach.get("achievement_type") or "").strip()
+        counts[ach_type] = counts.get(ach_type, 0) + 1
+    rows = []
+    for item in expected:
+        item_type = str(item.get("type") or "").strip()
+        target = item.get("target")
+        try:
+            target_v = int(float(target))
+        except (TypeError, ValueError):
+            target_v = None
+        unit = str(item.get("unit") or "项")
+        actual = counts.get(item_type)
+        if target_v is None:
+            verdict, detail = "数据待补", f"预期指标未配置数值（类型：{item_type}）"
+        elif actual is None:
+            verdict, detail = "未达标", f"预期 {target_v} {unit}，台账中未录入该类成果"
+        else:
+            if actual > target_v:
+                verdict = "超额完成"
+            elif actual == target_v:
+                verdict = "持平"
+            else:
+                verdict = "未达标"
+            detail = f"预期 {target_v} {unit}，实际 {actual} {unit}，{verdict}"
+        rows.append({"type": item_type, "target": target_v, "unit": unit, "actual": actual, "verdict": verdict, "detail": detail})
+    return rows
+
+
+def _c3_expense_sentences(expenses: list[dict[str, Any]], total_approved: Optional[float]) -> tuple[list[dict[str, Any]], str]:
+    """C3：按科目汇总并生成规范描述句。返回 (rows, 汇总段落)。"""
+    by_category: dict[str, float] = {}
+    for item in expenses:
+        try:
+            amount = float(item.get("amount") or 0)
+        except (TypeError, ValueError):
+            amount = 0.0
+        category = str(item.get("category") or "其他").strip() or "其他"
+        by_category[category] = by_category.get(category, 0.0) + amount
+    total = round(sum(by_category.values()), 2)
+    rows = []
+    sentences = []
+    for category, amount in sorted(by_category.items(), key=lambda kv: -kv[1]):
+        ratio = round(amount / total, 4) if total > 0 else 0.0
+        action = _EXPENSE_SENTENCE_TEMPLATES.get(category, f"开展{category}相关支出")
+        sentence = f"{action}，支出 {_wan(amount)}，占比 {ratio * 100:.1f}%"
+        rows.append({"category": category, "amount": round(amount, 2), "ratio": ratio, "sentence": sentence})
+        sentences.append(sentence)
+    if not rows:
+        return [], "经费决算明细尚未录入，请在补录后重新生成初稿（W1）。"
+    tail = f"。项目决算合计支出 {_wan(total)}。"
+    if total_approved is not None:
+        tail += "支出总额控制在立项经费范围内。" if total <= total_approved else "支出总额超出立项经费，请补充说明。"
+    return rows, "本项目经费支出按科目如下：" + "；".join(sentences) + "。" + tail
+
+
+def _c4_narrative_points(achievements: list[dict[str, Any]], excerpt: str) -> list[dict[str, str]]:
+    """C4：从代表性成果与年度报告摘句提取叙事素材要点。"""
+    points = []
+    for ach in achievements:
+        if ach.get("is_representative"):
+            title = str(ach.get("title") or "").strip()
+            venue = str(ach.get("venue_or_no") or "").strip()
+            points.append({"source": "成果清单", "point": f"代表性成果：{title}" + (f"（{venue}）" if venue else "")})
+    for sentence in str(excerpt or "").replace("；", "。").split("。"):
+        sentence = sentence.strip()
+        if sentence:
+            points.append({"source": "年度报告", "point": sentence})
+    return points
+
+
+def build_draft_sections(
+    project: dict[str, Any],
+    achievements: list[dict[str, Any]],
+    expenses: list[dict[str, Any]],
+) -> dict[str, Any]:
+    """组装初稿各章节数据（纯计算，不落库不出文件）。"""
+    import json as _json
+
+    project = _norm_project(project)
+    raw_expected = project.get("expected_json")
+    if raw_expected in (None, ""):
+        raw_expected = project.get("expected") or project.get("expectedJson") or []
+    if isinstance(raw_expected, str):
+        try:
+            expected = _json.loads(raw_expected)
+        except (TypeError, ValueError):
+            expected = []
+    elif isinstance(raw_expected, list):
+        expected = raw_expected
+    else:
+        expected = []
+    if isinstance(expected, dict):
+        expected = expected.get("indicators") or []
+
+    achievements = [{_snake(k): v for k, v in _norm_achievement(a).items()} for a in achievements]
+    expenses = [{_snake(k): v for k, v in _norm_expense(e).items()} for e in expenses]
+
+    completion = _c1_compare(expected, achievements)
+    expense_rows, expense_paragraph = _c3_expense_sentences(expenses, None)
+    excerpt = project.get("annual_report_excerpt") or project.get("annualReportExcerpt") or ""
+    narrative_points = _c4_narrative_points(achievements, str(excerpt))
+
+    achievement_lines = []
+    for ach in achievements:
+        title = str(ach.get("title") or "").strip()
+        venue = str(ach.get("venue_or_no") or "").strip()
+        ach_date = str(ach.get("ach_date") or "").strip()
+        line = f"{ach.get('achievement_type')}：{title}" + (f"（{venue}，{ach_date}）" if venue or ach_date else "")
+        achievement_lines.append(line + "。来源：成果清单")
+
+    sections = {
+        "1. 研究背景与目标": [_PLACEHOLDER],
+        "2. 研究内容与进展": (
+            [f"· {p['point']}（来源：{p['source']}）" for p in narrative_points] or [f"· {_PLACEHOLDER}"]
+        ) + [_PLACEHOLDER],
+        "3. 预期成果完成情况": [f"{row['type']}：{row['detail']}。来源：预期指标与成果清单" for row in completion]
+        or ["预期成果指标未配置，请在项目台账补录后重新生成（W1）。"],
+        "4. 经费决算说明": [expense_paragraph, "明细："] + [f"· {row['sentence']}" for row in expense_rows],
+        "5. 成果清单": achievement_lines or ["成果清单尚未录入，请在补录后重新生成初稿（W1）。"],
+        "6. 存在问题与展望": [_PLACEHOLDER],
+    }
+    return {
+        "completion": completion,
+        "expense_rows": expense_rows,
+        "expense_paragraph": expense_paragraph,
+        "narrative_points": narrative_points,
+        "achievement_lines": achievement_lines,
+        "sections": sections,
+    }
+
+
+def run_draft(project_raw: dict[str, Any], achievements: list[dict[str, Any]], expenses: list[dict[str, Any]], *, batch_no: str = "") -> dict[str, Any]:
+    """初稿内核：三步生成 → 写 docx + HTML 预览 → 落库。"""
+    from service.docx_service import write_docx
+    from service.report_service import write_preview_for_kernel
+
+    log = _log()
+    init_db()
+    project = _norm_project(project_raw)
+    project_no = str(project.get("project_no") or "").strip()
+    if not project_no:
+        raise ValueError("项目编号（project_no）不能为空")
+    batch_no = batch_no or now_unix().to_bytes(5, "big").hex() + uuid.uuid4().hex[:6]
+
+    if log:
+        log.info("draft_kernel_start batch_no=%s project_no=%s stage=kernel", batch_no, project_no)
+
+    sections_data = build_draft_sections(project, achievements, expenses)
+    sections = sections_data["sections"]
+
+    draft_file = write_docx(project, sections, batch_no)
+
+    draft = {
+        "project_no": project_no,
+        "project_name": str(project.get("project_name") or ""),
+        "batch_no": batch_no,
+        "sections": sections,
+        "completion": sections_data["completion"],
+        "narrative_points": sections_data["narrative_points"],
+    }
+    preview_file = write_preview_for_kernel(draft=draft)
+
+    conn = get_conn()
+    try:
+        cur = conn.cursor()
+        cur.execute(
+            """
+            INSERT INTO report_drafts
+                (batch_no, project_no, project_name, completion_json, narrative_points_json, draft_file, preview_file)
+            VALUES (?, ?, ?, ?, ?, ?, ?)
+            """,
+            (
+                batch_no,
+                project_no,
+                str(project.get("project_name") or ""),
+                json.dumps(sections_data["completion"], ensure_ascii=False),
+                json.dumps(sections_data["narrative_points"], ensure_ascii=False),
+                draft_file,
+                preview_file,
+            ),
+        )
+        draft_id = int(cur.lastrowid)
+        cur.execute(
+            "UPDATE closing_projects SET status = '初稿已生成', updated_at = ? WHERE project_no = ?",
+            (now_unix(), project_no),
+        )
+        conn.commit()
+    finally:
+        conn.close()
+
+    payload = {
+        "ok": True,
+        "batch_no": batch_no,
+        "draft_id": draft_id,
+        "project_no": project_no,
+        "draft_file": draft_file,
+        "preview_file": preview_file,
+        "completion": sections_data["completion"],
+        "narrative_points_count": len(sections_data["narrative_points"]),
+        "disclaimer": "初稿由台账数据与模板生成；数据章节请核对，叙事章节须本人终审，严禁引入台账之外的成果。",
+    }
+    if log:
+        log.info("draft_kernel_done batch_no=%s project_no=%s draft_id=%s stage=kernel", batch_no, project_no, draft_id)
+    return payload
+
+
+def query_pending_projects(project_no: str = "") -> list[dict[str, Any]]:
+    """读取待结题项目（可按项目编号筛选），附成果与决算数量。"""
+    conn = get_conn()
+    try:
+        cur = conn.cursor()
+        if project_no:
+            cur.execute("SELECT * FROM closing_projects WHERE project_no = ?", (project_no,))
+        else:
+            cur.execute("SELECT * FROM closing_projects WHERE status = '待结题' ORDER BY id")
+        columns = [d[0] for d in cur.description]
+        projects = [dict(zip(columns, row)) for row in cur.fetchall()]
+        for project in projects:
+            pno = project.get("project_no")
+            cur.execute("SELECT COUNT(*) FROM achievements WHERE project_no = ?", (pno,))
+            project["achievement_count"] = int(cur.fetchone()[0])
+            cur.execute("SELECT COUNT(*) FROM expense_items WHERE project_no = ?", (pno,))
+            project["expense_count"] = int(cur.fetchone()[0])
+    finally:
+        conn.close()
+    return projects