refactor(config): 重构配置类以支持动态元数据字段

配置类 ThesisConfig 现在使用 metadata 字典直接透传 TOML 配置，
无需为每个变量单独声明字段。新增模板变量只需修改 TOML 文件，
无需修改 Python 代码。

BREAKING CHANGE: 配置文件结构发生改变，从单独字段改为统一的
metadata 节点。

.gitignore

@@ -20,3 +20,7 @@ wheels/
 *.docx
 *.doc
 *.txt
+
+.vscode/
+.tmp/
+*.toml

transit/__init__.py

@@ -4,6 +4,7 @@ from .parser import parse_markdown
 from .body import body_to_paragraphs, replace_placeholder
 from .renderer import generate_thesis
 from .config import load_config, ThesisConfig
+from .references import references_to_paragraphs, format_gb7714
 
 __all__ = [
     "parse_markdown",
@@ -12,4 +13,6 @@ __all__ = [
     "generate_thesis",
     "load_config",
     "ThesisConfig",
+    "references_to_paragraphs",
+    "format_gb7714",
 ]

transit/body.py

@@ -6,7 +6,9 @@ Markdown 正文 → Word 段落转换。
 """
 
 import re
+from copy import deepcopy
 from docx import Document
+from docx.oxml.ns import qn
 
 _PAT_HEADING = re.compile(r"^(#{1,6})\s+(.+)$", re.MULTILINE)
 
@@ -84,6 +86,11 @@ def replace_placeholder(
             placeholder_style = para.style.name if para.style else None
             parent = para._element.getparent()
             idx = list(parent).index(para._element)
+
+            # 保存原段落的编号属性（numPr），用于继承自动编号
+            orig_pPr = para._element.find(qn("w:pPr"))
+            numPr = orig_pPr.find(qn("w:numPr")) if orig_pPr is not None else None
+
             parent.remove(para._element)
 
             for pd_data in reversed(paragraphs):
@@ -93,15 +100,24 @@ def replace_placeholder(
                 # 尝试应用样式，逐步降级
                 applied = _apply_style(new_p, doc, style_name)
                 if not applied and style_name.startswith("Heading"):
-                    # 标题样式找不到
                     new_p.style = doc.styles["Normal"]
                 elif not applied:
-                    # 正文样式找不到 → 尝试占位符自身的样式
                     if placeholder_style:
                         _apply_style(new_p, doc, placeholder_style)
                     if new_p.style.name == "Normal" and placeholder_style:
                         new_p.style = doc.styles[placeholder_style]
 
+                # 继承原段落的编号属性（自动编号）
+                if numPr is not None:
+                    new_pPr = new_p._element.find(qn("w:pPr"))
+                    if new_pPr is None:
+                        new_pPr = new_p._element.makeelement(qn("w:pPr"), {})
+                        new_p._element.insert(0, new_pPr)
+                    existing = new_pPr.find(qn("w:numPr"))
+                    if existing is not None:
+                        new_pPr.remove(existing)
+                    new_pPr.append(deepcopy(numPr))
+
                 parent.insert(idx, new_p._element)
             break
 

transit/config.py

@@ -6,17 +6,15 @@ import tomllib
 
 @dataclass
 class ThesisConfig:
-    """论文配置数据（学生信息、元数据等，不包含正文内容）。"""
+    """论文配置数据。
 
-    student_name: str = "<None>"
-    student_id: str = "<None>"
-    college: str = "<None>"
-    major: str = "<None>"
-    class_: str = "<None>"
-    advisor: str = "<None>"
-    advisor_title: str = "<None>"
-    title: str = "<None>"
+    ``metadata`` 直接透传 TOML 的 ``[metadata]`` 节，不再为每个变量声明字段。
+    新增模板变量只需改 TOML，无需修改 Python。
+    """
 
+    metadata: dict = field(default_factory=dict)
+
+    # 以下字段仍有业务逻辑，保留为显式属性
     title_from_md: bool = True
     body_start_keywords: list[str] = field(default_factory=lambda: ["绪论", "引言"])
     body_end_keywords: list[str] = field(
@@ -24,19 +22,11 @@ class ThesisConfig:
     )
     body_style: str = "Body Text Indent"
     level_offset: int = -1
+    reference_style: str = "列出段落1"
 
     def to_dict(self) -> dict:
-        """转成模板渲染用的扁平字典，排除 options 命名空间。"""
-        return {
-            "student_name": self.student_name,
-            "student_id": self.student_id,
-            "college": self.college,
-            "major": self.major,
-            "class": self.class_,
-            "advisor": self.advisor,
-            "advisor_title": self.advisor_title,
-            "title": self.title,
-        }
+        """透传 metadata（模板变量来源）。"""
+        return self.metadata
 
 
 def load_config(path: str | Path) -> ThesisConfig:
@@ -49,14 +39,7 @@ def load_config(path: str | Path) -> ThesisConfig:
     opts = raw.get("options", {})
 
     return ThesisConfig(
-        student_name=meta.get("student_name", "<None>"),
-        student_id=meta.get("student_id", "<None>"),
-        college=meta.get("college", "<None>"),
-        major=meta.get("major", "<None>"),
-        class_=meta.get("class", "<None>"),
-        advisor=meta.get("advisor", "<None>"),
-        advisor_title=meta.get("advisor_title", "<None>"),
-        title=meta.get("title", "<None>"),
+        metadata=meta,
         title_from_md=opts.get("title_from_md", True),
         body_start_keywords=opts.get("body_start_keywords", ["绪论", "引言"]),
         body_end_keywords=opts.get(
@@ -64,4 +47,5 @@ def load_config(path: str | Path) -> ThesisConfig:
         ),
         body_style=opts.get("body_style", "Body Text Indent"),
         level_offset=opts.get("level_offset", -1),
+        reference_style=opts.get("reference_style", "列出段落1"),
     )

transit/references.py

@@ -0,0 +1,114 @@
+"""
+GB 7714 《文后参考文献著录规则》顺序编码制 — 参考文献解析与格式化。
+
+将 Markdown 中 ``[N] ... [TYPE] ...`` 格式的参考文献逐条解析，
+按文献类型（书 M、期刊 J、会议 C、学位论文 D、标准 S、电子资源 EB/OL 等）
+重新编排为符合 GB 7714 规范的格式，并输出为独立段落。
+"""
+
+import re
+from typing import Optional
+
+# 单行匹配: [N] 开头
+_RE_LINE = re.compile(r"^\[(\d+)\]\s*(.*)$")
+# 文献类型标记: [M] [J] [C] [D] [S] [EB/OL] [P] [N]
+_RE_TYPE = re.compile(r"\[(\w+(?:/\w+)?)\]")
+
+
+# GB 7714 中各文献类型的标准格式模板（仅用于说明，实际格式化直接拼接）
+TYPE_LABELS: dict[str, str] = {
+    "M": "专著",
+    "J": "期刊文章",
+    "C": "会议论文",
+    "D": "学位论文",
+    "S": "标准",
+    "EB/OL": "电子资源",
+    "P": "专利",
+    "N": "报纸文章",
+}
+
+
+def parse_reference_line(line: str) -> Optional[dict]:
+    """解析单行参考文献，提取序号、作者+标题、文献类型、来源信息。
+
+    期望格式::
+
+        [N] Authors. Title[TYPE]. Source info.
+    """
+    line = line.strip()
+    m = _RE_LINE.match(line)
+    if not m:
+        return None
+
+    number = int(m.group(1))
+    rest = m.group(2).strip()
+
+    # 定位文献类型标记 [TYPE]
+    tm = _RE_TYPE.search(rest)
+    if not tm:
+        return None
+
+    before_type = rest[: tm.start()].strip().rstrip(".")
+    doc_type = tm.group(1)
+    after_type = rest[tm.end() :].strip().lstrip(".").strip()
+
+    return {
+        "number": number,
+        "before_type": before_type,  # "Authors. Title"
+        "doc_type": doc_type,  # "M", "J", "EB/OL", …
+        "after_type": after_type,  # 来源信息
+    }
+
+
+def _normalize_period(text: str) -> str:
+    """确保文本以英文句点结尾（GB 7714 要求）。"""
+    text = text.rstrip()
+    if text and not text.endswith("."):
+        text += "."
+    return text
+
+
+def format_gb7714(ref: dict) -> str:
+    """按 GB 7714 重新编排一条参考文献（不含序号前缀，由 Word 样式自动编号）。
+
+    格式::
+
+        Authors. Title[TYPE]. Source.
+    """
+    bt = ref["before_type"]
+    dt = ref["doc_type"]
+    at = ref["after_type"]
+    formatted = f"{bt}[{dt}]. {at}"
+    return _normalize_period(formatted)
+
+
+def references_to_paragraphs(
+    ref_text: str,
+    ref_style: str = "列出段落1",
+) -> list[dict]:
+    """将参考文献原始文本转换为格式化段落列表。
+
+    返回的每个元素::
+
+        {"text": str, "level": 0, "style": ref_style}
+
+    每条参考文献为一个独立段落。
+    """
+    if not ref_text or ref_text == "<None>":
+        return [{"text": "<None>", "level": 0, "style": ref_style}]
+
+    lines = [l.strip() for l in ref_text.strip().split("\n") if l.strip()]
+    paragraphs: list[dict] = []
+
+    for line in lines:
+        ref = parse_reference_line(line)
+        if ref:
+            formatted = format_gb7714(ref)
+        else:
+            # 无法解析时，至少去掉 [N] 前缀
+            fallback = re.sub(r"^\[\d+\]\s*", "", line)
+            formatted = _normalize_period(fallback)
+
+        paragraphs.append({"text": formatted, "level": 0, "style": ref_style})
+
+    return paragraphs

transit/renderer.py

@@ -12,9 +12,11 @@ from docx import Document
 from .config import load_config, ThesisConfig
 from .parser import parse_markdown
 from .body import body_to_paragraphs, replace_placeholder
+from .references import references_to_paragraphs
 
 
-_TEXT_FIELDS = [
+# 解析器可能产生的字段（用于填充报告）
+_PARSER_FIELDS = [
     "title",
     "abstact_cn_context",
     "abstract_cn_keywords",
@@ -23,13 +25,7 @@ _TEXT_FIELDS = [
     "acknowledgement",
     "reference",
     "appendix",
-    "student_name",
-    "student_id",
-    "college",
-    "major",
-    "class",
-    "advisor",
-    "advisor_title",
+    "body_md",
 ]
 
 
@@ -80,12 +76,11 @@ def generate_thesis(
         body_end_kw=config.body_end_keywords,
     )
 
-    # 3. 合并配置 → 上下文（配置优先）
+    # 3. 合并配置 → 上下文（配置填充解析器未产生的空白）
     for k, v in config.to_dict().items():
         if k == "title" and config.title_from_md and context.get("title"):
             continue  # 以 markdown 标题为准
-        if v != "<None>":
-            context[k] = v
+        context.setdefault(k, v)
 
     # 4. 用 defaultdict 兜底缺失键
     ctx = defaultdict(lambda: "<None>", context)
@@ -97,8 +92,13 @@ def generate_thesis(
         if body_md else []
     )
 
-    # 6. 占位符
+    # 6. 解析参考文献为段落列表
+    ref_text = ctx.get("reference", "")
+    ref_paragraphs = references_to_paragraphs(ref_text, ref_style=config.reference_style)
+
+    # 7. 占位符（替代模板变量，后处理时替换）
     ctx["body_placeholder"] = "__CONTEXT_PLACEHOLDER__"
+    ctx["reference"] = "__REFERENCE_PLACEHOLDER__"
 
     # 7. 渲染模板
     doc = DocxTemplate(str(template_path))
@@ -108,28 +108,33 @@ def generate_thesis(
     temp_path = Path(output_path).with_suffix(".tmp")
     doc.save(str(temp_path))
 
-    # 9. 正文注入
+    # 9. 正文注入＋参考文献注入
     final_doc = Document(str(temp_path))
     replace_placeholder(
         final_doc, "__CONTEXT_PLACEHOLDER__", body_paragraphs,
         default_body_style=config.body_style,
     )
+    replace_placeholder(
+        final_doc, "__REFERENCE_PLACEHOLDER__", ref_paragraphs,
+        default_body_style=config.reference_style,
+    )
     final_doc.save(str(output_path))
     temp_path.unlink(missing_ok=True)
 
     print(f"[完成] 论文生成完成: {output_path}")
 
-    # 10. 字段填充报告
+    # 10. 字段填充报告（动态收集所有模板与解析字段）
+    report_fields = list(dict.fromkeys([*config.metadata.keys(), *_PARSER_FIELDS]))
     print("\n--- 字段填充情况 ---")
-    for key in _TEXT_FIELDS:
-        val = ctx[key]
+    for key in report_fields:
+        val = ctx.get(key, "<None>")
         if val == "<None>":
             print(f"  [缺失] {key}")
         else:
             preview = str(val)[:60].replace("\n", " ")
             print(f"  [OK] {key}: {preview}...")
 
-    missing = [k for k in _TEXT_FIELDS if ctx[k] == "<None>"]
+    missing = [k for k in report_fields if ctx.get(k, "<None>") == "<None>"]
     if missing:
         print("\n[警告] 以下字段缺失，已填充 '<None>'：")
         for f in missing: