cscec-gwdoc-mcp 0.1.0__tar.gz

cscec_gwdoc_mcp-0.1.0/PKG-INFO

@@ -0,0 +1,57 @@
+Metadata-Version: 2.5
+Name: cscec-gwdoc-mcp
+Version: 0.1.0
+Summary: 中建企业公文 Word（.docx）生成 stdio MCP 服务，供百炼函数计算 uvx 托管接入。
+Author: CSCEC
+License: Proprietary
+Keywords: cscec,docx,mcp,word,公文
+Requires-Python: >=3.10
+Requires-Dist: lxml>=4.9
+Requires-Dist: mcp[cli]<2,>=1.0
+Requires-Dist: pydantic>=2.0
+Requires-Dist: python-docx>=1.1
+Description-Content-Type: text/markdown
+
+# cscec-gwdoc-mcp
+
+中建企业公文 Word（.docx）生成 **stdio MCP** 服务。可被 `uvx` 安装运行，适配阿里云百炼函数计算（FC）代码包（uvx）托管模式，再由中建 AI 等外部 MCP 客户端经百炼接入。
+
+## 工具
+
+`generate_official_doc`：根据结构化公文内容生成 `.docx`。
+
+| 参数 | 类型 | 说明 |
+| --- | --- | --- |
+| `title` | str | 公文标题（必填） |
+| `sections` | list[dict] | 章节列表，每项 `{heading, paragraphs, subheadings?}` |
+| `document_number` | str | 文号 |
+| `recipient` | str | 主送单位 |
+| `introduction` | str | 导语 |
+| `attachments` | list[str] | 附件名称 |
+| `issuer` | str | 发文单位 |
+| `department` | str | 发文部门（代章） |
+| `issue_date` | str | 发文日期 |
+| `cc` | str | 抄送 |
+| `document_type` | str | 公文类型，默认「通报」 |
+
+## 运行
+
+```bash
+uvx cscec-gwdoc-mcp
+```
+
+服务以 stdio 方式启动；生成的 `.docx` 作为 MCP 内嵌二进制资源（base64）随工具结果回传，不提供公网下载链接。
+
+## 本地开发自测
+
+```bash
+uv venv
+uv pip install -e .
+python -m cscec_gwdoc_mcp.server
+```
+
+## 数据与排版说明
+
+- 生成器优先套用包内范文的页面设置、Word 样式及页眉页脚，再重建正文。
+- 未提供的文号、日期、单位、统计数字、整改时限等应标记为待补充，人工核对后再正式发文。
+- 「方正小标宋简体」「仿宋_GB2312」「楷体_GB2312」等字体需目标环境安装，缺字体会被自动替换。

cscec_gwdoc_mcp-0.1.0/README.md

@@ -0,0 +1,43 @@
+# cscec-gwdoc-mcp
+
+中建企业公文 Word（.docx）生成 **stdio MCP** 服务。可被 `uvx` 安装运行，适配阿里云百炼函数计算（FC）代码包（uvx）托管模式，再由中建 AI 等外部 MCP 客户端经百炼接入。
+
+## 工具
+
+`generate_official_doc`：根据结构化公文内容生成 `.docx`。
+
+| 参数 | 类型 | 说明 |
+| --- | --- | --- |
+| `title` | str | 公文标题（必填） |
+| `sections` | list[dict] | 章节列表，每项 `{heading, paragraphs, subheadings?}` |
+| `document_number` | str | 文号 |
+| `recipient` | str | 主送单位 |
+| `introduction` | str | 导语 |
+| `attachments` | list[str] | 附件名称 |
+| `issuer` | str | 发文单位 |
+| `department` | str | 发文部门（代章） |
+| `issue_date` | str | 发文日期 |
+| `cc` | str | 抄送 |
+| `document_type` | str | 公文类型，默认「通报」 |
+
+## 运行
+
+```bash
+uvx cscec-gwdoc-mcp
+```
+
+服务以 stdio 方式启动；生成的 `.docx` 作为 MCP 内嵌二进制资源（base64）随工具结果回传，不提供公网下载链接。
+
+## 本地开发自测
+
+```bash
+uv venv
+uv pip install -e .
+python -m cscec_gwdoc_mcp.server
+```
+
+## 数据与排版说明
+
+- 生成器优先套用包内范文的页面设置、Word 样式及页眉页脚，再重建正文。
+- 未提供的文号、日期、单位、统计数字、整改时限等应标记为待补充，人工核对后再正式发文。
+- 「方正小标宋简体」「仿宋_GB2312」「楷体_GB2312」等字体需目标环境安装，缺字体会被自动替换。

cscec_gwdoc_mcp-0.1.0/cscec_gwdoc_mcp/__init__.py

@@ -0,0 +1,4 @@
+# -*- coding: utf-8 -*-
+"""中建企业公文 Word 生成 stdio MCP 服务。"""
+
+__version__ = "0.1.0"

cscec_gwdoc_mcp-0.1.0/cscec_gwdoc_mcp/server.py

@@ -0,0 +1,166 @@
+# -*- coding: utf-8 -*-
+"""stdio MCP 服务：中建企业公文 Word 生成。
+
+通过 FastMCP 以 stdio 方式运行（供百炼函数计算 uvx 托管）。
+生成的 .docx 不再走公网下载链接，而是作为 MCP 内嵌二进制资源（base64）
+随工具结果回传。
+"""
+
+from __future__ import annotations
+
+import base64
+import re
+import tempfile
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+from urllib.parse import quote
+
+from importlib.resources import files
+
+from mcp.server.fastmcp import FastMCP
+from mcp.types import BlobResourceContents, EmbeddedResource, TextContent
+
+from .service import OfficialDocument, SectionItem, build_document
+
+DOCX_MIME = "application/vnd.openxmlformats-officedocument.wordprocessingml.document"
+ASSET_NAME = "公文格式范文.docx"
+
+mcp = FastMCP("cscec-gwdoc-mcp")
+
+
+def _template_bytes() -> bytes:
+    """读取包内范文模板二进制。"""
+    return files("cscec_gwdoc_mcp.assets").joinpath(ASSET_NAME).read_bytes()
+
+
+def _safe_filename(title: str) -> str:
+    """清理文件名中的非法字符。"""
+    name = re.sub(r'[\\/:*?"<>|\r\n]+', "_", title).strip()
+    return name or "公文"
+
+
+def _build_document_payload(
+    title: str,
+    sections: Optional[List[Dict[str, Any]]] = None,
+    document_number: str = "",
+    recipient: str = "",
+    introduction: str = "",
+    attachments: Optional[List[str]] = None,
+    issuer: str = "",
+    department: str = "",
+    issue_date: str = "",
+    cc: str = "",
+    document_type: str = "通报",
+) -> OfficialDocument:
+    section_items: List[SectionItem] = []
+    for raw in sections or []:
+        if not isinstance(raw, dict):
+            continue
+        section_items.append(
+            SectionItem(
+                heading=str(raw.get("heading", "")),
+                paragraphs=[str(p) for p in raw.get("paragraphs", []) or []],
+                subheadings=[str(s) for s in raw.get("subheadings", []) or []],
+            )
+        )
+    return OfficialDocument(
+        title=title,
+        sections=section_items,
+        document_number=document_number,
+        recipient=recipient,
+        introduction=introduction,
+        attachments=list(attachments or []),
+        issuer=issuer,
+        department=department,
+        issue_date=issue_date,
+        cc=cc,
+        document_type=document_type,
+    )
+
+
+def render_docx(payload: OfficialDocument) -> tuple[str, bytes]:
+    """在临时目录生成 docx，返回 (文件名, 文件二进制)。"""
+    template_bytes = _template_bytes()
+    with tempfile.TemporaryDirectory() as tmp:
+        out_path = Path(tmp) / f"{_safe_filename(payload.title)}.docx"
+        build_document(payload, out_path, template_bytes)
+        return out_path.name, out_path.read_bytes()
+
+
+@mcp.tool()
+def generate_official_doc(
+    title: str,
+    sections: Optional[List[Dict[str, Any]]] = None,
+    document_number: str = "",
+    recipient: str = "",
+    introduction: str = "",
+    attachments: Optional[List[str]] = None,
+    issuer: str = "",
+    department: str = "",
+    issue_date: str = "",
+    cc: str = "",
+    document_type: str = "通报",
+) -> List[Any]:
+    """根据结构化公文内容生成 Word（.docx）公文。
+
+    所有数据必须来自用户资料；未提供的文号、日期、单位、统计数字、
+    整改时限等应在调用前标记为待补充，不要自行编造。
+
+    :param title: 公文标题（必填）
+    :param sections: 正文章节列表，每项含 heading(章节标题)、
+        paragraphs(正文段落列表)、subheadings(可选，括号序号子标题列表)
+    :param document_number: 文号，如「中建四华建安监函字〔2026〕60号」
+    :param recipient: 主送单位，如「司属各项目：」
+    :param introduction: 导语 / 开头说明
+    :param attachments: 附件名称列表
+    :param issuer: 发文单位
+    :param department: 发文部门（代章）
+    :param issue_date: 发文日期
+    :param cc: 抄送范围
+    :param document_type: 公文类型，默认「通报」
+    :return: 生成的 .docx 文件（内嵌二进制资源）及文字说明
+    """
+    payload = _build_document_payload(
+        title=title,
+        sections=sections,
+        document_number=document_number,
+        recipient=recipient,
+        introduction=introduction,
+        attachments=attachments,
+        issuer=issuer,
+        department=department,
+        issue_date=issue_date,
+        cc=cc,
+        document_type=document_type,
+    )
+
+    filename, data = render_docx(payload)
+    blob = base64.b64encode(data).decode()
+
+    text = (
+        f"公文已生成：{filename}\n"
+        f"类型：{payload.document_type}；标题：{payload.title}\n"
+        f"文件大小：{len(data)} 字节；格式：.docx（Word）。\n"
+        f"正文共 {len(payload.sections)} 个章节；未提供的文号、日期、单位、"
+        f"数据与时限等需人工核对后再正式发文。"
+    )
+
+    return [
+        TextContent(type="text", text=text),
+        EmbeddedResource(
+            type="resource",
+            resource=BlobResourceContents(
+                uri=f"file:///{quote(filename)}",
+                mimeType=DOCX_MIME,
+                blob=blob,
+            ),
+        ),
+    ]
+
+
+def main() -> None:
+    mcp.run(transport="stdio")
+
+
+if __name__ == "__main__":
+    main()

cscec_gwdoc_mcp-0.1.0/cscec_gwdoc_mcp/service.py

@@ -0,0 +1,244 @@
+# -*- coding: utf-8 -*-
+"""公文构建核心逻辑（纯函数，不含任何 HTTP / 服务层）。
+
+由原 word_service.py 的 FastAPI 版本剥离而来：保留 OfficialDocument /
+SectionItem 模型与 build_document 排版逻辑；模板范文从包内资源读取，
+产物写入调用方给定的输出路径。
+"""
+
+from __future__ import annotations
+
+import io
+import re
+from pathlib import Path
+from typing import List, Optional
+
+from docx import Document
+from docx.document import Document as DocumentType
+from docx.oxml import OxmlElement
+from docx.oxml.ns import qn
+from docx.shared import Pt
+from docx.text.paragraph import Paragraph
+from lxml import etree
+from pydantic import BaseModel, Field
+
+
+# ---------------------------------------------------------------------------
+# 数据模型
+# ---------------------------------------------------------------------------
+
+class SectionItem(BaseModel):
+    """正文章节。
+
+    heading：章节标题；paragraphs：正文段落列表；
+    subheadings：子标题（括号序号标题）列表，顺序排在正文段落之前。
+    """
+
+    heading: str
+    paragraphs: List[str] = Field(default_factory=list)
+    subheadings: List[str] = Field(default_factory=list)
+
+
+class OfficialDocument(BaseModel):
+    """结构化公文内容。所有未提供的字段应在提示词层标记为待补充。"""
+
+    title: str
+    sections: List[SectionItem] = Field(default_factory=list)
+    document_number: str = ""
+    recipient: str = ""
+    introduction: str = ""
+    attachments: List[str] = Field(default_factory=list)
+    issuer: str = ""
+    department: str = ""
+    issue_date: str = ""
+    cc: str = ""
+    document_type: str = "通报"
+
+
+# ---------------------------------------------------------------------------
+# 字号 / 字体工具
+# ---------------------------------------------------------------------------
+
+def _pt(size: Optional[str]) -> Optional[Pt]:
+    """将配置里的字号字符串转为 Pt；已是 pt 数值时直接使用。"""
+    if not size:
+        return None
+    match = re.search(r"(\d+(?:\.\d+)?)", str(size))
+    if not match:
+        return None
+    return Pt(float(match.group(1)))
+
+
+def _set_run_font(run, font_name: str, size: Optional[Pt], bold: Optional[bool] = None) -> None:
+    run.font.name = font_name
+    if size is not None:
+        run.font.size = size
+    if bold is not None:
+        run.font.bold = bold
+    rpr = run._element.get_or_add_rPr()
+    rfonts = rpr.rFonts
+    if rfonts is None:
+        rfonts = OxmlElement("w:rFonts")
+        rpr.append(rfonts)
+    rfonts.set(qn("w:eastAsia"), font_name)
+    rfonts.set(qn("w:ascii"), font_name)
+    rfonts.set(qn("w:hAnsi"), font_name)
+
+
+def _set_paragraph_font(paragraph: Paragraph, font_name: str, size: Optional[Pt], bold: Optional[bool] = None) -> None:
+    """段落已有文本时，统一设置整段字体。"""
+    for run in paragraph.runs:
+        _set_run_font(run, font_name, size, bold)
+
+
+def _set_doc_default_font(doc: DocumentType, font_name: str, size: Optional[Pt]) -> None:
+    styles = doc.styles
+    if "Normal" in styles:
+        normal = styles["Normal"]
+        normal.font.name = font_name
+        normal.font.size = size
+        rpr = normal.element.get_or_add_rPr()
+        rfonts = rpr.rFonts
+        if rfonts is None:
+            rfonts = OxmlElement("w:rFonts")
+            rpr.append(rfonts)
+        rfonts.set(qn("w:eastAsia"), font_name)
+        rfonts.set(qn("w:ascii"), font_name)
+        rfonts.set(qn("w:hAnsi"), font_name)
+
+
+# ---------------------------------------------------------------------------
+# 文本处理
+# ---------------------------------------------------------------------------
+
+def _split_inline_bold(text: str) -> List[tuple]:
+    """切分 ``**加粗**`` 片段，返回 [(文本, 是否加粗), ...]。"""
+    parts: List[tuple] = []
+    pattern = re.compile(r"\*\*(.+?)\*\*")
+    last = 0
+    for match in pattern.finditer(text):
+        if match.start() > last:
+            parts.append((text[last:match.start()], False))
+        parts.append((match.group(1), True))
+        last = match.end()
+    if last < len(text):
+        parts.append((text[last:], False))
+    if not parts:
+        parts.append((text, False))
+    return parts
+
+
+def _add_body_paragraph(doc: DocumentType, text: str, font_name: str, size: Optional[Pt]) -> None:
+    paragraph = doc.add_paragraph()
+    paragraph.paragraph_format.first_line_indent = Pt(28)
+    paragraph.paragraph_format.line_spacing = 1.5
+    parts = _split_inline_bold(text)
+    for part_text, bold in parts:
+        run = paragraph.add_run(part_text)
+        _set_run_font(run, font_name, size, bold if bold else None)
+
+
+# ---------------------------------------------------------------------------
+# 主构建逻辑
+# ---------------------------------------------------------------------------
+
+def build_document(data: OfficialDocument, output_path: Path, template_bytes: bytes) -> Path:
+    """根据结构化内容生成 Word 公文。
+
+    :param data: 结构化公文内容
+    :param output_path: 生成的 docx 写入路径
+    :param template_bytes: 包内范文的二进制内容（页面设置 / 样式 / 页眉页脚参考）
+    :return: 生成文件路径
+    """
+    output_path = Path(output_path)
+    output_path.parent.mkdir(parents=True, exist_ok=True)
+
+    doc = Document(io.BytesIO(template_bytes))
+
+    # 清空模板正文（保留页面设置、样式与页眉页脚）
+    body = doc._element.body
+    for child in list(body):
+        if child.tag != qn("w:sectPr"):
+            body.remove(child)
+
+    _set_doc_default_font(doc, "仿宋_GB2312", _pt("16pt"))
+
+    # 文号
+    if data.document_number:
+        num_paragraph = doc.add_paragraph()
+        num_paragraph.alignment = 1  # 居中
+        num_paragraph.paragraph_format.space_before = Pt(6)
+        num_paragraph.paragraph_format.space_after = Pt(12)
+        num_run = num_paragraph.add_run(data.document_number)
+        _set_run_font(num_run, "仿宋_GB2312", _pt("16pt"))
+
+    # 标题
+    title_paragraph = doc.add_paragraph()
+    title_paragraph.alignment = 1  # 居中
+    title_paragraph.paragraph_format.space_before = Pt(12)
+    title_paragraph.paragraph_format.space_after = Pt(18)
+    title_run = title_paragraph.add_run(data.title)
+    _set_run_font(title_run, "方正小标宋简体", _pt("22pt"), True)
+
+    # 主送单位
+    if data.recipient:
+        recipient_paragraph = doc.add_paragraph()
+        recipient_paragraph.paragraph_format.line_spacing = 1.5
+        recipient_run = recipient_paragraph.add_run(data.recipient)
+        _set_run_font(recipient_run, "仿宋_GB2312", _pt("16pt"))
+
+    # 导语
+    if data.introduction:
+        _add_body_paragraph(doc, data.introduction, "仿宋_GB2312", _pt("16pt"))
+
+    # 章节
+    for section in data.sections:
+        if section.heading:
+            head_paragraph = doc.add_paragraph()
+            head_paragraph.paragraph_format.space_before = Pt(12)
+            head_paragraph.paragraph_format.space_after = Pt(6)
+            head_run = head_paragraph.add_run(section.heading)
+            _set_run_font(head_run, "黑体", _pt("16pt"), True)
+
+        for subheading in section.subheadings:
+            sub_paragraph = doc.add_paragraph()
+            sub_paragraph.paragraph_format.first_line_indent = Pt(28)
+            sub_paragraph.paragraph_format.line_spacing = 1.5
+            sub_run = sub_paragraph.add_run(subheading)
+            _set_run_font(sub_run, "楷体_GB2312", _pt("16pt"), True)
+
+        for paragraph_text in section.paragraphs:
+            _add_body_paragraph(doc, paragraph_text, "仿宋_GB2312", _pt("16pt"))
+
+    # 附件
+    if data.attachments:
+        attach_paragraph = doc.add_paragraph()
+        attach_paragraph.paragraph_format.space_before = Pt(12)
+        attach_text = "附件：" + "、".join(data.attachments)
+        attach_run = attach_paragraph.add_run(attach_text)
+        _set_run_font(attach_run, "仿宋_GB2312", _pt("16pt"))
+
+    # 落款
+    if data.issuer or data.department:
+        sign_paragraph = doc.add_paragraph()
+        sign_paragraph.alignment = 2  # 右对齐
+        sign_paragraph.paragraph_format.space_before = Pt(24)
+        sign_text = "　".join([part for part in [data.issuer, data.department] if part])
+        sign_run = sign_paragraph.add_run(sign_text)
+        _set_run_font(sign_run, "仿宋_GB2312", _pt("16pt"))
+
+    if data.issue_date:
+        date_paragraph = doc.add_paragraph()
+        date_paragraph.alignment = 2
+        date_run = date_paragraph.add_run(data.issue_date)
+        _set_run_font(date_run, "仿宋_GB2312", _pt("16pt"))
+
+    # 抄送
+    if data.cc:
+        cc_paragraph = doc.add_paragraph()
+        cc_paragraph.paragraph_format.space_before = Pt(18)
+        cc_run = cc_paragraph.add_run(f"抄送：{data.cc}")
+        _set_run_font(cc_run, "仿宋_GB2312", _pt("14pt"))
+
+    doc.save(str(output_path))
+    return output_path

cscec_gwdoc_mcp-0.1.0/pyproject.toml

@@ -0,0 +1,25 @@
+[build-system]
+requires = ["hatchling>=1.21"]
+build-backend = "hatchling.build"
+
+[project]
+name = "cscec-gwdoc-mcp"
+version = "0.1.0"
+description = "中建企业公文 Word（.docx）生成 stdio MCP 服务，供百炼函数计算 uvx 托管接入。"
+readme = "README.md"
+requires-python = ">=3.10"
+license = { text = "Proprietary" }
+authors = [{ name = "CSCEC" }]
+keywords = ["mcp", "word", "docx", "公文", "cscec"]
+dependencies = [
+    "mcp[cli]>=1.0,<2",
+    "python-docx>=1.1",
+    "pydantic>=2.0",
+    "lxml>=4.9",
+]
+
+[project.scripts]
+cscec-gwdoc-mcp = "cscec_gwdoc_mcp.server:main"
+
+[tool.hatch.build.targets.wheel]
+packages = ["cscec_gwdoc_mcp"]