aria-code 0.52.0__py3-none-any.whl

aria_code/markdown_pdf.py

@@ -0,0 +1,289 @@
+"""
+markdown_pdf.py — Markdown 报告 → 印刷级排版 HTML/PDF（中英双语模板）
+=====================================================================
+与 report_generator.py（HTML 研报引擎）互补：本模块面向**Markdown 交付物**
+（/report --format md 的投研报告、分析纪要、任何模型生成的 md 文本），
+渲染成专业浅色印刷模板后走 report_generator.export_pdf 引擎链出 PDF。
+
+模板特性（与 Arthera research 渲染管线同源）:
+  · 真表格：深色列头、斑马纹、跨页重复表头（thead）
+  · 信号徽章着色：买入/BUY 绿 · 减仓/REDUCE 琥珀 · 卖出/SELL 红 · 持有/HOLD 灰
+  · 首个引用块升级为元信息条（生成时间 | 数据截止 …）
+  · 打印分页规则：@page A4、表格行不跨页断裂、章节标题不孤行
+  · lang="zh" / "en" 切换模板文案（页脚、免责声明）；"auto" 按内容检测
+
+PDF 分页取决于引擎：headless Chrome（推荐，完整 @page 支持）；
+macOS WebKit 兜底为单长页（内容完整，无分页）。
+
+对外:
+  render_markdown_html(md_text, lang="auto", title=None) -> str
+  markdown_to_pdf(md_text_or_path, pdf_path, lang="auto") -> Optional[Path]
+  tool_export_markdown_pdf(params) + register_markdown_pdf_tools(...)
+"""
+
+from __future__ import annotations
+
+import logging
+import re
+from datetime import datetime
+from pathlib import Path
+from typing import Any, Dict, List, Optional
+
+logger = logging.getLogger(__name__)
+
+try:
+    import markdown as _markdown
+    HAS_MARKDOWN = True
+except ImportError:
+    HAS_MARKDOWN = False
+
+_LABELS = {
+    "zh": {
+        "footer": "本报告由 Aria 确定性渲染管线生成",
+        "disclaimer": "仅供研究参考，不构成投资建议",
+    },
+    "en": {
+        "footer": "Generated by the Aria deterministic render pipeline",
+        "disclaimer": "For research reference only — not investment advice",
+    },
+}
+
+_BADGES = [
+    (r"买入|BUY|STRONG_BUY", "#0a7a43", "#e6f6ec"),
+    (r"减仓|REDUCE", "#8a5b00", "#fdf3d7"),
+    (r"卖出|SELL|离场|EXIT", "#b91c1c", "#fdeaea"),
+    (r"持有|HOLD|观望|WATCH", "#4b5563", "#eef1f4"),
+    (r"共振|CONFIRMS?", "#0a6e8a", "#e3f4f9"),
+    (r"分歧|DIVERGES?", "#a16207", "#fdf3d7"),
+]
+
+_CSS = """
+@page { size: A4; margin: 14mm 12mm 16mm 12mm; }
+* { box-sizing: border-box; }
+body { font-family: -apple-system, "PingFang SC", "Hiragino Sans GB",
+       "Microsoft YaHei", "Noto Sans CJK SC", "Segoe UI", sans-serif;
+       color: #1a202c; background: #fff; font-size: 10.5pt;
+       line-height: 1.55; margin: 0; }
+.page { max-width: 186mm; margin: 0 auto; }
+h1 { font-size: 19pt; margin: 0 0 4px; padding-bottom: 8px;
+     border-bottom: 3px solid #1f2937; }
+h2 { font-size: 13pt; margin: 22px 0 8px; padding: 6px 10px;
+     background: #f3f5f8; border-left: 4px solid #334155;
+     border-radius: 3px; page-break-after: avoid; }
+h3 { font-size: 11.5pt; margin: 14px 0 6px; page-break-after: avoid; }
+p { margin: 6px 0; }
+hr { border: none; border-top: 1px solid #e2e8f0; margin: 14px 0; }
+.meta { display: flex; flex-wrap: wrap; gap: 6px 14px; align-items: center;
+        background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 6px;
+        padding: 8px 12px; margin: 10px 0 4px; font-size: 9pt; color: #334155; }
+.meta b, .meta strong { color: #0f172a; }
+blockquote { margin: 8px 0; padding: 6px 12px; font-size: 9.5pt;
+             color: #57534e; background: #fffbeb;
+             border-left: 3px solid #d97706; border-radius: 3px; }
+blockquote p { margin: 2px 0; }
+table { border-collapse: collapse; width: 100%; margin: 10px 0 14px;
+        font-size: 9pt; page-break-inside: auto; }
+thead { display: table-header-group; }
+tr { page-break-inside: avoid; }
+th { background: #1f2937; color: #fff; font-weight: 600; padding: 6px 8px;
+     text-align: left; white-space: nowrap; }
+td { border: 1px solid #e2e8f0; padding: 5px 8px; vertical-align: top; }
+tbody tr:nth-child(even) { background: #f8fafc; }
+td:first-child { white-space: nowrap; }
+/* Wide-table auto-fit: shrink font/padding by column count so content
+   never crosses the page boundary (classes added by _fit_wide_tables). */
+table.t-wide  { font-size: 7.5pt; }
+table.t-wide  th, table.t-wide  td { padding: 4px 5px; }
+table.t-xwide { font-size: 6.8pt; }
+table.t-xwide th, table.t-xwide td { padding: 3px 4px; }
+table.t-wide th, table.t-xwide th { white-space: normal; }
+table.t-wide td, table.t-xwide td { word-break: keep-all; overflow-wrap: break-word; }
+table.t-wide  td:nth-child(-n+2),
+table.t-xwide td:nth-child(-n+2) { white-space: nowrap; }
+table.t-wide  .badge { font-size: 7pt;  padding: 1px 5px; }
+table.t-xwide .badge { font-size: 6.3pt; padding: 0 4px; }
+.badge { display: inline-block; padding: 1px 7px; border-radius: 10px;
+         font-size: 8.5pt; font-weight: 600; white-space: nowrap; }
+ul, ol { margin: 6px 0; padding-left: 22px; }
+li { margin: 3px 0; }
+code { font-family: "SF Mono", Menlo, Consolas, monospace;
+       background: #f1f5f9; padding: 1px 5px; border-radius: 3px;
+       font-size: 9pt; }
+pre { background: #f8fafc; border: 1px solid #e2e8f0; border-radius: 6px;
+      padding: 10px 12px; overflow-x: auto; page-break-inside: avoid; }
+pre code { background: none; padding: 0; }
+strong { color: #0f172a; }
+.footer { margin-top: 26px; padding-top: 8px;
+          border-top: 1px solid #e2e8f0; font-size: 8pt;
+          color: #94a3b8; text-align: center; }
+"""
+
+
+def detect_lang(md_text: str) -> str:
+    """内容含 CJK 字符占比 > 5% 判为中文。"""
+    if not md_text:
+        return "en"
+    cjk = sum(1 for ch in md_text[:4000] if "一" <= ch <= "鿿")
+    return "zh" if cjk > len(md_text[:4000]) * 0.05 else "en"
+
+
+def _apply_badges(html: str) -> str:
+    def _decorate_cell(m: re.Match) -> str:
+        cell = m.group(0)
+        for pat, fg, bg in _BADGES:
+            cell = re.sub(
+                rf"((?:🟢|🟡|🔴|⬜|✅|⚠️|➖)?\s*(?:<strong>)?(?:{pat})(?:</strong>)?)",
+                rf'<span class="badge" style="color:{fg};background:{bg}">\1</span>',
+                cell, count=1,
+            )
+        return cell
+    return re.sub(r"<td>.*?</td>", _decorate_cell, html, flags=re.S)
+
+
+def _fit_wide_tables(html: str) -> str:
+    """≥12 列 → t-xwide (6.8pt)；≥9 列 → t-wide (7.5pt)：按边界自适应字号。"""
+    def _tag(m: re.Match) -> str:
+        table = m.group(0)
+        head = table.split("</tr>", 1)[0]
+        n = len(re.findall(r"<th[ >]", head))
+        if n >= 12:
+            cls = "t-xwide"
+        elif n >= 9:
+            cls = "t-wide"
+        else:
+            return table
+        return re.sub(r"<table>", f'<table class="{cls}">', table, count=1)
+    return re.sub(r"<table>.*?</table>", _tag, html, flags=re.S)
+
+
+def _promote_meta_bar(html: str) -> str:
+    return re.sub(
+        r"(</h1>\s*)<blockquote>(.*?)</blockquote>",
+        lambda m: m.group(1) + '<div class="meta">'
+                  + re.sub(r"</?p>", "", m.group(2)).replace("|", "·")
+                  + "</div>",
+        html, count=1, flags=re.S,
+    )
+
+
+def render_markdown_html(md_text: str, lang: str = "auto",
+                         title: Optional[str] = None) -> str:
+    """Markdown 文本 → 完整可打印 HTML 文档（浅色印刷模板）。"""
+    if not HAS_MARKDOWN:
+        raise RuntimeError("python-markdown 未安装: pip install markdown")
+    if lang == "auto":
+        lang = detect_lang(md_text)
+    labels = _LABELS.get(lang, _LABELS["en"])
+
+    body = _markdown.markdown(
+        md_text, extensions=["tables", "sane_lists", "nl2br", "fenced_code"],
+        output_format="html5",
+    )
+    body = _promote_meta_bar(body)
+    body = _apply_badges(body)
+    body = _fit_wide_tables(body)
+
+    if not title:
+        m = re.search(r"<h1[^>]*>(.*?)</h1>", body, re.S)
+        title = re.sub(r"<[^>]+>", "", m.group(1)) if m else "Report"
+    footer = (f"{labels['footer']} · {datetime.now():%Y-%m-%d %H:%M} · "
+              f"{labels['disclaimer']}")
+    return (
+        f'<!doctype html><html lang="{lang}"><head><meta charset="utf-8">'
+        f"<title>{title}</title><style>{_CSS}</style></head>"
+        f'<body><div class="page">{body}'
+        f'<div class="footer">{footer}</div></div></body></html>'
+    )
+
+
+def markdown_to_pdf(md: str | Path, pdf_path: Path,
+                    lang: str = "auto") -> Optional[Path]:
+    """Markdown（文本或 .md 路径）→ 排版 PDF。返回 PDF 路径，失败 None。"""
+    md_text = (Path(md).read_text(encoding="utf-8")
+               if isinstance(md, Path) or (isinstance(md, str)
+                   and len(md) < 512 and Path(md).expanduser().is_file())
+               else str(md))
+    html = render_markdown_html(md_text, lang=lang)
+    from report_generator import html_string_to_pdf
+    return html_string_to_pdf(html, Path(pdf_path))
+
+
+# ── Agent 工具 ───────────────────────────────────────────────────────────────
+
+def tool_export_markdown_pdf(params: Dict[str, Any]) -> Dict[str, Any]:
+    """LOCAL_TOOLS 入口。params: {markdown|md_path, filename?, lang?}"""
+    try:
+        md = params.get("markdown") or params.get("md_path")
+        if not md:
+            return {"success": False, "error": "缺少 markdown 或 md_path"}
+        lang = params.get("lang", "auto")
+        stem = str(params.get("filename") or "report").removesuffix(".pdf")
+        try:
+            from artifacts import create_user_artifact
+            record = create_user_artifact("reports", params.get("topic"),
+                                          stem, ".pdf")
+            pdf_path = record.path
+        except Exception:
+            pdf_path = Path.cwd() / f"{stem}.pdf"
+        source = Path(md) if (isinstance(md, str) and len(md) < 512
+                              and Path(md).expanduser().is_file()) else md
+        out = markdown_to_pdf(source, pdf_path, lang=lang)
+        if out is None:
+            return {"success": False,
+                    "error": "无可用 PDF 引擎（安装 Chrome/Edge，"
+                             "或 macOS: pip install pyobjc-framework-WebKit）"}
+        return {"success": True, "path": str(out),
+                "size_kb": out.stat().st_size // 1024,
+                "lang": lang if lang != "auto" else detect_lang(
+                    md if isinstance(md, str) else Path(md).read_text(encoding="utf-8")[:4000])}
+    except Exception as exc:
+        logger.debug("[md-pdf] export error: %s", exc, exc_info=True)
+        return {"success": False, "error": str(exc)}
+
+
+MARKDOWN_PDF_TOOL_SCHEMAS: List[Dict[str, Any]] = [
+    {
+        "type": "function",
+        "function": {
+            "name": "export_markdown_pdf",
+            "description": (
+                "把 Markdown 报告渲染成专业排版 PDF（印刷级模板：真表格、"
+                "信号徽章着色、元信息条、A4 分页）。支持中英文模板"
+                "（lang: zh/en/auto，auto 按内容检测）。传 markdown 原文"
+                "或 md_path 文件路径均可；返回 PDF 路径。"
+            ),
+            "parameters": {
+                "type": "object",
+                "properties": {
+                    "markdown": {"type": "string",
+                                 "description": "Markdown 原文（与 md_path 二选一）"},
+                    "md_path": {"type": "string",
+                                "description": "Markdown 文件路径（与 markdown 二选一）"},
+                    "filename": {"type": "string",
+                                 "description": "输出文件名（不含扩展名）"},
+                    "topic": {"type": "string", "description": "产物子目录主题"},
+                    "lang": {"type": "string", "enum": ["auto", "zh", "en"],
+                             "description": "模板语言，默认 auto"},
+                },
+                "required": [],
+            },
+        },
+    },
+]
+
+
+def register_markdown_pdf_tools(tool_registry: Dict, schema_registry: List) -> int:
+    if not HAS_MARKDOWN:
+        return 0
+    added = 0
+    if "export_markdown_pdf" not in tool_registry:
+        tool_registry["export_markdown_pdf"] = (
+            tool_export_markdown_pdf,
+            "Markdown 报告 → 专业排版 PDF（中英双语模板）",
+        )
+        added += 1
+    existing = {s.get("function", {}).get("name") for s in schema_registry}
+    for schema in MARKDOWN_PDF_TOOL_SCHEMAS:
+        if schema["function"]["name"] not in existing:
+            schema_registry.append(schema)
+    return added

aria_code/market_data_client.py

@@ -0,0 +1,4 @@
+"""Compatibility import for :mod:`clients.market_data_client`."""
+from importlib import import_module as _import_module
+import sys as _sys
+_sys.modules[__name__] = _import_module("clients.market_data_client")