@hongmaple0820/med-scale-research-os 0.43.0

package/target-research-platform/src/output/evidence_package.py

@@ -0,0 +1,230 @@
+"""
+证据包生成器
+将所有输出打包为完整的研究交付物
+"""
+import json
+import shutil
+from pathlib import Path
+from datetime import datetime
+from typing import Optional
+
+from .chart_generator import ChartGenerator
+from .data_exporter import DataExporter
+from .report_generator import ReportGenerator
+
+
+class EvidencePackage:
+    """
+    生成完整的研究证据包
+
+    结构:
+    output/{gene}_{disease}/
+    ├── charts/              # PNG/SVG 图表
+    ├── data/                # CSV/TSV 原始数据
+    ├── report.html          # HTML 报告
+    ├── report.pdf           # PDF 报告
+    ├── report.docx          # DOCX 报告
+    ├── pipeline_output.json # 完整 JSON 结果
+    ├── manifest.json        # 包清单（含溯源）
+    └── README.md            # 使用说明
+    """
+
+    def __init__(self, output_dir: str, dpi: int = 150):
+        self.output_dir = Path(output_dir)
+        self.output_dir.mkdir(parents=True, exist_ok=True)
+        self.dpi = dpi
+
+        self.charts_dir = self.output_dir / "charts"
+        self.data_dir = self.output_dir / "data"
+
+        self.chart_gen = ChartGenerator(str(self.charts_dir), dpi=dpi)
+        self.data_exp = DataExporter(str(self.data_dir))
+        self.report_gen = ReportGenerator(str(self.output_dir))
+
+    def generate(
+        self,
+        pipeline_results: dict,
+        gene_symbol: str,
+        disease: str,
+        title: str = None,
+    ) -> dict:
+        """
+        生成完整证据包
+
+        Args:
+            pipeline_results: pipeline 输出的完整结果字典
+            gene_symbol: 基因符号
+            disease: 疾病名称
+            title: 报告标题
+
+        Returns:
+            包含所有生成文件路径的字典
+        """
+        manifest = {
+            "gene": gene_symbol,
+            "disease": disease,
+            "generated_at": datetime.now().isoformat(),
+            "pipeline_version": "1.0",
+            "files": {},
+        }
+
+        stages = pipeline_results.get("stages", pipeline_results)
+
+        # === 1. 导出 JSON ===
+        json_path = self.data_exp.export_pipeline_json(pipeline_results, "pipeline_output.json")
+        manifest["files"]["json"] = json_path
+
+        # === 2. 导出 CSV 数据 ===
+        csv_exports = {}
+        s1 = stages.get("stage1", {})
+        s1_summary = s1.get("summary", s1)
+        if s1_summary:
+            csv_exports.update(self.data_exp.export_all_stage1(s1_summary, gene_symbol))
+
+        s2 = stages.get("stage2", {})
+        s2_summary = s2.get("summary", s2)
+        if s2_summary:
+            csv_exports.update(self.data_exp.export_all_stage2(s2_summary, disease))
+
+        s4 = stages.get("stage4", {}).get("summary", {})
+        milestones = s4.get("milestones", {})
+        if milestones and "phases" in milestones:
+            csv_exports["milestones"] = self.data_exp.export_milestones(
+                milestones["phases"], gene_symbol
+            )
+
+        manifest["files"]["csv"] = {k: v.get("csv", "") for k, v in csv_exports.items() if isinstance(v, dict)}
+
+        # === 3. 生成图表 ===
+        chart_files = {}
+
+        # Stage1: 组织表达
+        tissues = s1_summary.get("tissue_expression", {}).get("gtex", {}).get("tissues", {})
+        if tissues:
+            chart_files["tissue_expression"] = self.chart_gen.tissue_expression_bar(
+                tissues, gene_symbol
+            )
+
+        # Stage1: 免疫相关性
+        immune = s1_summary.get("immune", {})
+        for cancer_type, corr_data in immune.items():
+            if isinstance(corr_data, dict) and "correlations" in corr_data:
+                chart_files[f"immune_{cancer_type}"] = self.chart_gen.immune_correlation_heatmap(
+                    corr_data["correlations"], gene_symbol, cancer_type
+                )
+
+        # Stage2: 综合排名
+        targets = s2_summary.get("top_targets", [])
+        if targets:
+            chart_files["composite_ranking"] = self.chart_gen.composite_ranking_bar(targets)
+
+        # Stage2: 火山图
+        de_results = s2_summary.get("differential_expression", {}).get("results", [])
+        if de_results:
+            chart_files["volcano"] = self.chart_gen.volcano_plot(de_results, gene_symbol)
+
+        # Stage3: 成药性雷达
+        if targets:
+            chart_files["druggability_radar"] = self.chart_gen.druggability_radar(targets)
+
+        # Stage4: 里程碑
+        phases = s4.get("phases", milestones.get("phases", []))
+        if phases:
+            chart_files["milestones"] = self.chart_gen.milestone_gantt(phases, gene_symbol)
+
+        manifest["files"]["charts"] = {k: v.get("png", "") for k, v in chart_files.items() if isinstance(v, dict)}
+
+        # === 4. 生成 HTML 报告 ===
+        html_path = self.report_gen.generate_html(
+            pipeline_results, gene_symbol, disease, title=title,
+        )
+        manifest["files"]["html_report"] = html_path
+
+        # === 5. 生成 PDF ===
+        try:
+            pdf_path = self.report_gen.generate_pdf(html_path)
+            manifest["files"]["pdf_report"] = pdf_path
+        except Exception as e:
+            manifest["files"]["pdf_report"] = f"Error: {e}"
+
+        # === 6. 生成 DOCX ===
+        try:
+            docx_path = self.report_gen.generate_docx(
+                pipeline_results, gene_symbol, disease, title=title
+            )
+            manifest["files"]["docx_report"] = docx_path
+        except Exception as e:
+            manifest["files"]["docx_report"] = f"Error: {e}"
+
+        # === 7. 写入 manifest ===
+        manifest_path = self.output_dir / "manifest.json"
+        with open(manifest_path, "w", encoding="utf-8") as f:
+            json.dump(manifest, f, ensure_ascii=False, indent=2, default=str)
+
+        # === 8. 生成 README ===
+        self._generate_readme(gene_symbol, disease, manifest)
+
+        return manifest
+
+    def _generate_readme(self, gene_symbol: str, disease: str, manifest: dict):
+        """生成包说明文件"""
+        lines = [
+            f"# {gene_symbol} — {disease} 研究证据包",
+            "",
+            f"生成时间: {manifest['generated_at']}",
+            f"Pipeline 版本: {manifest['pipeline_version']}",
+            "",
+            "## 文件清单",
+            "",
+            "| 类型 | 文件 | 说明 |",
+            "|------|------|------|",
+        ]
+
+        files = manifest.get("files", {})
+        if files.get("html_report"):
+            lines.append(f"| HTML 报告 | `{Path(files['html_report']).name}` | 可交互网页报告 |")
+        if files.get("pdf_report") and not str(files.get("pdf_report", "")).startswith("Error"):
+            lines.append(f"| PDF 报告 | `{Path(files['pdf_report']).name}` | 可打印/存档 PDF |")
+        if files.get("docx_report") and not str(files.get("docx_report", "")).startswith("Error"):
+            lines.append(f"| DOCX 报告 | `{Path(files['docx_report']).name}` | 可编辑 Word 文档 |")
+        if files.get("json"):
+            lines.append(f"| JSON 数据 | `{Path(files['json']).name}` | 完整结构化结果 |")
+
+        csv_files = files.get("csv", {})
+        for name, path in csv_files.items():
+            if path:
+                lines.append(f"| CSV 数据 | `data/{Path(path).name}` | {name} 原始数据 |")
+
+        chart_files = files.get("charts", {})
+        for name, path in chart_files.items():
+            if path:
+                lines.append(f"| 图表 | `charts/{Path(path).name}` | {name} 图表 (PNG) |")
+
+        lines.extend([
+            "",
+            "## 数据溯源",
+            "",
+            "| 数据源 | 版本 | 用途 |",
+            "|--------|------|------|",
+            "| GTEx | v8 | 正常组织表达 |",
+            "| TCGA/GDC | 2024 | 肿瘤表达与生存分析 |",
+            "| GEO | 2024 | 差异表达分析 |",
+            "| OpenTargets | 2024.02 | 靶点-疾病关联 |",
+            "",
+            "## 使用说明",
+            "",
+            "1. 用浏览器打开 `report.html` 查看可视化报告",
+            "2. `report.pdf` 可直接打印或提交",
+            "3. `report.docx` 可编辑后纳入正式文档",
+            "4. `data/` 目录包含所有原始数据 (CSV)，可导入 Excel/R/Python",
+            "5. `charts/` 目录包含出版级图表 (PNG/SVG)",
+            "6. `pipeline_output.json` 包含完整结构化结果，可供 agent 解析",
+            "",
+            "## 引用",
+            "",
+            "本报告由药物靶点智能研究平台自动生成。",
+            "数据来源: GTEx, TCGA/GDC, GEO, OpenTargets, PubMed, ClinicalTrials.gov",
+        ])
+
+        readme_path = self.output_dir / "README.md"
+        readme_path.write_text("\n".join(lines), encoding="utf-8")

package/target-research-platform/src/output/journal_templates.py

@@ -0,0 +1,193 @@
+"""
+期刊投稿模板 — 《中华肿瘤杂志》等中文核心期刊格式。
+"""
+from __future__ import annotations
+
+from datetime import datetime
+from typing import Any
+
+
+def render_zhonghua_oncology(
+    context: dict[str, Any],
+    body_md: str,
+    *,
+    figures: list[dict[str, str]] | None = None,
+) -> str:
+    """
+    将 IMRaD 正文包装为《中华肿瘤杂志》风格稿件。
+
+    格式要点：
+    - 中英文题目、结构化摘要、关键词
+    - 正文层级：前言 / 材料与方法 / 结果 / 讨论 / 参考文献
+    - 图表中英文双语题注
+    - 文末利益冲突、伦理声明、基金占位
+    """
+    target = context.get("target", "")
+    disease_zh = context.get("disease_zh", context.get("disease", ""))
+    title_zh = context.get("title_zh", "")
+    title_en = context.get("title_en", "")
+    abstract = _extract_section(body_md, "摘要") or _build_structured_abstract(context)
+    keywords = f"{target}；{disease_zh}；治疗靶点；成药性；临床试验；生物信息学"
+    keywords_en = f"{target}; {context.get('disease', '')}; therapeutic target; druggability; clinical trials"
+
+    figure_block = _render_figures(figures or context.get("figures") or [])
+
+    # 去掉 body 中已有的重复摘要/标题，保留正文核心
+    core_body = body_md
+    for marker in ("# ", "## 摘要", "## 1 引言", "## 1 前言"):
+        if marker in core_body:
+            break
+    if "## 1 引言" in body_md:
+        core_body = body_md.split("## 1 引言", 1)[-1]
+        core_body = "## 1 前言\n\n" + core_body.lstrip()
+    elif "## 1 前言" in body_md:
+        idx = body_md.index("## 1 前言")
+        core_body = body_md[idx:]
+
+    today = datetime.now().strftime("%Y-%m-%d")
+
+    manuscript = f"""---
+journal: 中华肿瘤杂志
+template_version: "1.0"
+generated_at: {context.get("generated_at", today)}
+---
+
+<div align="center">
+
+# {title_zh}
+
+**{title_en}**
+
+[作者姓名]<sup>1</sup>　[共同作者]<sup>1,2</sup>
+
+<sup>1</sup> [单位名称，城市 邮编]　<sup>2</sup> [合作单位]
+
+**通信作者：** [姓名]，E-mail: [email@institution.edu.cn]
+
+</div>
+
+---
+
+**【摘要】**　{_strip_md_headers(abstract)}
+
+**【关键词】**　{keywords}
+
+**【Abstract】**　{_english_abstract_placeholder(context)}
+
+**【Key words】**　{keywords_en}
+
+---
+
+**基金项目：** 国家自然科学基金（编号：待填）；省部级课题（待填）
+
+**收稿日期：** {today}　　**修回日期：** 待填
+
+---
+
+{core_body}
+
+---
+
+{figure_block}
+
+---
+
+## 利益冲突声明
+
+所有作者声明不存在利益冲突。
+
+## 伦理审查
+
+本研究使用公开去标识化数据库，未涉及人体或动物实验，免予伦理审查。
+
+## 作者贡献声明
+
+[待填：研究设计 / 数据分析 / 论文撰写 / 经费支持 — CRediT 格式]
+
+---
+
+> **排版说明（投稿前删除本段）：**
+> 1. 图表需按在正文中首次出现顺序编号（图1、图2…；表1、表2…）
+> 2. 《中华肿瘤杂志》要求参考文献著录格式为温哥华样式序号制
+> 3. 英文摘要需由作者或专业翻译润色
+> 4. 配置 `OPENAI_API_KEY` 可启用 LLM 深度文献评述
+
+*模板：中华肿瘤杂志 v1.0 · 靶点研究平台*
+"""
+    return _sanitize_submission_placeholders(manuscript)
+
+
+def _sanitize_submission_placeholders(text: str) -> str:
+    """Keep journal fields present without making unverifiable submission claims."""
+    replacements = {
+        "国家自然科学基金（编号：待填）；省部级课题（待填）": "待填写（投稿前由作者确认）",
+        "所有作者声明不存在利益冲突。": "利益冲突声明：待全体作者确认后填写。",
+        "本研究使用公开去标识化数据库，未涉及人体或动物实验，免予伦理审查。": (
+            "本稿基于公开数据库与去标识化数据；是否需要伦理审查或豁免应由作者所在机构确认。"
+        ),
+    }
+    for old, new in replacements.items():
+        text = text.replace(old, new)
+    return text
+
+
+def _extract_section(md: str, heading: str) -> str | None:
+    token = f"## {heading}"
+    if token not in md:
+        return None
+    part = md.split(token, 1)[1]
+    for stop in ("\n## ", "\n---"):
+        if stop in part:
+            part = part.split(stop, 1)[0]
+    return part.strip()
+
+
+def _strip_md_headers(text: str) -> str:
+    lines = []
+    for line in text.splitlines():
+        if line.startswith("#"):
+            continue
+        if line.strip() == "---":
+            continue
+        lines.append(line)
+    return "\n".join(lines).strip()
+
+
+def _build_structured_abstract(context: dict[str, Any]) -> str:
+    clin = context.get("clinical_summary") or {}
+    dev = context.get("developability") or {}
+    ms = context.get("milestones") or {}
+    return (
+        f"**目的** 评估 {context.get('target')} 在{context.get('disease_zh')}中的治疗靶点价值。"
+        f"**方法** 整合 GTEx、TCGA、PubMed 与 ClinicalTrials.gov 四阶段计算分析。"
+        f"**结果** 临床试验 {clin.get('total_trials', '—')} 项，可开发性 {dev.get('score', '—')}/100，"
+        f"研发周期约 {ms.get('total_duration', '—')}。"
+        f"**结论** 具备研发基础，需明确差异化策略。"
+    )
+
+
+def _english_abstract_placeholder(context: dict[str, Any]) -> str:
+    return (
+        f"Objective: To assess {context.get('target')} as a therapeutic target in "
+        f"{context.get('disease', '')} using integrative public data. "
+        f"Methods: Multi-omics and clinical registry analysis. "
+        f"Results: See Chinese abstract for quantitative findings. "
+        f"Conclusions: Supports further development with differentiation required. "
+        f"[English abstract to be polished by authors before submission.]"
+    )
+
+
+def _render_figures(figures: list[dict[str, str]]) -> str:
+    if not figures:
+        return ""
+    lines = ["## 图表（嵌入稿）\n"]
+    for i, fig in enumerate(figures, 1):
+        path = fig.get("relative") or fig.get("path", "")
+        title_zh = fig.get("title_zh", f"图{i}")
+        title_en = fig.get("title_en", f"Figure {i}")
+        lines.append(f"### 图{i}　{title_zh}\n")
+        lines.append(f"**Figure {i}.** {title_en}\n")
+        if path:
+            lines.append(f"![图{i}]({path})\n")
+        lines.append("")
+    return "\n".join(lines)