@bill_nee/dwg-reader 1.2.0 → 1.4.0

package/lib/index.js

@@ -1,5 +1,6 @@
 import { fileURLToPath } from "node:url";
 import { spawn } from "node:child_process";
+import { PDFParse } from "pdf-parse";
 import { LibreDwg, Dwg_File_Type, DwgCodePage } from "../vendor/libredwg-web/lib/index.js";
 
 const WASM_DIR = fileURLToPath(new URL("../vendor/libredwg-web/wasm", import.meta.url));
@@ -7,6 +8,8 @@ const MAX_DWG_BYTES = 512 * 1024 * 1024; // 512 MB 上限，超大图纸会先
 const MAX_ENTITY_DETAIL = 2000;
 const PYTHON_EXE = "D:\\dsh-home\\dsh-runtimes\\dsh-primary-runtime\\dependencies\\python\\python.exe";
 const DXF_SCRIPT = fileURLToPath(new URL("../scripts/read_dxf.py", import.meta.url));
+const OCR_SCRIPT = fileURLToPath(new URL("../scripts/ocr_image.py", import.meta.url));
+const MAX_PDF_BYTES = 100 * 1024 * 1024; // 100 MB PDF 上限
 
 let libreDwgPromise;
 
@@ -568,8 +571,329 @@ export const inject = ["tools", "fs", "workingDirectory"];
 export function apply(ctx, config) {
   const disposeDwg = ctx.tools.register(buildReadDwgTool(ctx));
   const disposeDxf = ctx.tools.register(buildReadDxfTool(ctx));
+  const disposePdf = ctx.tools.register(buildReadPdfTool(ctx));
+  const disposeImage = ctx.tools.register(buildReadImageTool(ctx));
   return () => {
     disposeDwg();
     disposeDxf();
+    disposePdf();
+    disposeImage();
+  };
+}
+
+function buildReadImageTool(ctx) {
+  return {
+    name: "read_image",
+    description:
+      "读取图片/截屏文件（尺寸、格式、元数据 + Tesseract OCR 文字提取，无需安装图片查看软件）。" +
+      "支持 PNG/JPEG/WebP/BMP/TIFF；OCR 语言为简体中文+英文。",
+    parameters: {
+      type: "object",
+      additionalProperties: true,
+      properties: {
+        file_path: {
+          type: "string",
+          description: "图片文件路径（绝对路径或相对当前工作目录）。",
+        },
+        ocr: {
+          type: "boolean",
+          description: "是否执行 OCR 文字提取（默认 true）。",
+        },
+        tessdata_prefix: {
+          type: "string",
+          description: "Tesseract tessdata 目录（可选；缺省使用系统默认）。",
+        },
+      },
+      required: ["file_path"],
+    },
+    output: {
+      schema: {
+        type: "object",
+        additionalProperties: true,
+        properties: {
+          path: { type: "string", description: "解析后的文件路径。" },
+          fileName: { type: "string", description: "文件名。" },
+          fileSizeBytes: { type: "integer", description: "文件字节数。" },
+          format: { type: "string", description: "图片格式（如 PNG/JPEG）。" },
+          width: { type: "integer", description: "图片宽度（像素）。" },
+          height: { type: "integer", description: "图片高度（像素）。" },
+          mode: { type: "string", description: "颜色模式（如 RGB/RGBA）。" },
+          ocrText: { type: "string", description: "OCR 提取的文字（可能为空）。" },
+          ocrError: { type: "string", description: "OCR 错误信息；成功时为 null。" },
+          error: { type: "string", description: "错误信息；成功时为 null。" },
+        },
+      },
+      render: (args, value) => {
+        const lines = [];
+        lines.push(`<path>${value.path}</path>`);
+        lines.push(`<type>image</type>`);
+        if (value.error) {
+          lines.push(`<content>读取图片失败：${value.error}</content>`);
+          return [{ type: "text", text: lines.join("\n") }];
+        }
+        const sizeKb = (value.fileSizeBytes / 1024).toFixed(1);
+        lines.push(`<content>已解析图片：${value.fileName}（${sizeKb} KB）`);
+        lines.push(`格式：${value.format}；尺寸：${value.width}×${value.height}；模式：${value.mode}`);
+        if (value.ocrError) lines.push(`OCR 错误：${value.ocrError}`);
+        if (value.ocrText) {
+          const preview = value.ocrText.slice(0, 300);
+          lines.push(`OCR 文字（${value.ocrText.length} 字符）：${preview}${value.ocrText.length > 300 ? "…" : ""}`);
+        } else if (!value.ocrError) {
+          lines.push("OCR 文字：无（未检测到可识别文字或 OCR 未启用）");
+        }
+        lines.push(`</content>`);
+        return [{ type: "text", text: lines.join("\n") }];
+      },
+      presentationMeta: () => null,
+    },
+    isConcurrencySafe: () => false,
+    presentCall: (args) => ({
+      card: "generic",
+      title: `读取图片 ${args.file_path || "(未指定路径)"}`,
+      kind: "read",
+      locations: [{ path: args.file_path || "" }],
+    }),
+    presentResult: (args, result) => {
+      if (result && result.isError) return undefined;
+      return {
+        card: "generic",
+        title: `图片解析完成：${args.file_path || ""}`,
+        kind: "read",
+        locations: [{ path: args.file_path || "" }],
+      };
+    },
+    async execute(args, exec) {
+      if (typeof args !== "object" || args === null) {
+        throw new Error('invalid arguments: expected an object with a "file_path" string');
+      }
+      if (typeof args.file_path !== "string" || !args.file_path.trim()) {
+        throw new Error('invalid arguments: "file_path" must be a non-empty string');
+      }
+      if (args.ocr !== undefined && typeof args.ocr !== "boolean") {
+        throw new Error('invalid arguments: "ocr" must be a boolean');
+      }
+      if (args.tessdata_prefix !== undefined && typeof args.tessdata_prefix !== "string") {
+        throw new Error('invalid arguments: "tessdata_prefix" must be a string');
+      }
+
+      const resolveOptions = await sessionResolveOptions(ctx, exec);
+      const target = await ctx.fs.resolve(args.file_path.trim(), resolveOptions);
+      const info = await ctx.fs.stat(target, exec.signal);
+      const displayPath = ctx.fs.processPath(target);
+      if (!info) {
+        ctx.emit("fs/observed", target, { kind: "absent" }, exec);
+        throw new Error(`cannot read "${displayPath}": not found`);
+      }
+      if (info.type !== "file") {
+        throw new Error(`cannot read "${displayPath}": not a regular file`);
+      }
+
+      const pythonArgs = [
+        OCR_SCRIPT,
+        args.file_path.trim(),
+        "D:\\Tools\\Tesseract-OCR\\tesseract.exe",
+      ];
+
+      const result = await new Promise((resolve, reject) => {
+        const proc = spawn(PYTHON_EXE, pythonArgs, { stdio: ["pipe", "pipe", "pipe"] });
+        let stdout = "";
+        let stderr = "";
+        proc.stdout.on("data", (d) => (stdout += d.toString()));
+        proc.stderr.on("data", (d) => (stderr += d.toString()));
+        proc.on("close", (code) => {
+          if (code !== 0) {
+            reject(new Error(`Python 进程退出码 ${code}: ${stderr || stdout}`));
+            return;
+          }
+          try {
+            resolve(JSON.parse(stdout));
+          } catch (err) {
+            reject(new Error(`OCR 结果解析失败: ${err.message}; stdout=${stdout.slice(0, 200)}`));
+          }
+        });
+        proc.on("error", (err) => reject(new Error(`无法启动 Python 进程: ${err.message}`)));
+      });
+
+      if (result.error) {
+        throw new Error(`cannot read "${displayPath}": ${result.error}`);
+      }
+
+      ctx.emit("fs/observed", target, { kind: "present", version: info.version }, exec);
+
+      return toJson({
+        path: displayPath,
+        fileName: displayPath.split(/[\\/]/).pop(),
+        fileSizeBytes: result.fileSizeBytes ?? info.size !== undefined ? info.size : 0,
+        format: result.format ?? "",
+        width: result.width ?? null,
+        height: result.height ?? null,
+        mode: result.mode ?? null,
+        ocrText: result.ocrText ?? "",
+        ocrError: result.ocrError ?? null,
+      });
+    },
+  };
+}
+
+function buildReadPdfTool(ctx) {
+  return {
+    name: "read_pdf",
+    description:
+      "读取 PDF 文件（基于 pdf-parse，无需安装 Adobe Acrobat）。" +
+      "返回 PDF 页数、元数据（标题/作者/创建日期等）与可选全文文本。",
+    parameters: {
+      type: "object",
+      additionalProperties: true,
+      properties: {
+        file_path: {
+          type: "string",
+          description: "PDF 文件路径（绝对路径或相对当前工作目录）。",
+        },
+        extract_text: {
+          type: "boolean",
+          description: "是否提取全文文本（默认 true）。",
+        },
+        max_text_chars: {
+          type: "integer",
+          description: "提取文本的最大字符数（默认 50000，0 表示不限制）。",
+        },
+      },
+      required: ["file_path"],
+    },
+    output: {
+      schema: {
+        type: "object",
+        additionalProperties: true,
+        properties: {
+          path: { type: "string", description: "解析后的文件路径。" },
+          fileName: { type: "string", description: "文件名。" },
+          fileSizeBytes: { type: "integer", description: "文件字节数。" },
+          format: { type: "string", description: "文件格式（固定为 pdf）。" },
+          pageCount: { type: "integer", description: "PDF 页数。" },
+          metadata: { type: "object", additionalProperties: true, description: "PDF 元数据（标题、作者、创建日期等）。" },
+          text: { type: "string", description: "提取的全文文本（可能被截断）。" },
+          textLength: { type: "integer", description: "提取文本的字符数。" },
+          textTruncated: { type: "boolean", description: "文本是否被截断。" },
+          error: { type: "string", description: "错误信息；成功时为 null。" },
+        },
+      },
+      render: (args, value) => {
+        const lines = [];
+        lines.push(`<path>${value.path}</path>`);
+        lines.push(`<type>pdf</type>`);
+        if (value.error) {
+          lines.push(`<content>读取 PDF 失败：${value.error}</content>`);
+          return [{ type: "text", text: lines.join("\n") }];
+        }
+        const sizeMb = (value.fileSizeBytes / 1024 / 1024).toFixed(2);
+        lines.push(`<content>已解析 PDF：${value.fileName}（${sizeMb} MB）`);
+        lines.push(`页数：${value.pageCount}`);
+        if (value.metadata && Object.keys(value.metadata).length) {
+          const metaLines = Object.entries(value.metadata)
+            .filter(([, v]) => v != null && String(v).trim())
+            .map(([k, v]) => `  ${k}: ${v}`)
+            .join("\n");
+          if (metaLines) lines.push(`元数据：\n${metaLines}`);
+        }
+        lines.push(`文本长度：${value.textLength} 字符${value.textTruncated ? "（已截断）" : ""}`);
+        if (value.text) {
+          const preview = value.text.slice(0, 300);
+          lines.push(`文本预览：${preview}${value.text.length > 300 ? "…" : ""}`);
+        }
+        lines.push(`</content>`);
+        return [{ type: "text", text: lines.join("\n") }];
+      },
+      presentationMeta: () => null,
+    },
+    isConcurrencySafe: () => false,
+    presentCall: (args) => ({
+      card: "generic",
+      title: `读取 PDF ${args.file_path || "(未指定路径)"}`,
+      kind: "read",
+      locations: [{ path: args.file_path || "" }],
+    }),
+    presentResult: (args, result) => {
+      if (result && result.isError) return undefined;
+      return {
+        card: "generic",
+        title: `PDF 解析完成：${args.file_path || ""}`,
+        kind: "read",
+        locations: [{ path: args.file_path || "" }],
+      };
+    },
+    async execute(args, exec) {
+      if (typeof args !== "object" || args === null) {
+        throw new Error('invalid arguments: expected an object with a "file_path" string');
+      }
+      if (typeof args.file_path !== "string" || !args.file_path.trim()) {
+        throw new Error('invalid arguments: "file_path" must be a non-empty string');
+      }
+      if (args.extract_text !== undefined && typeof args.extract_text !== "boolean") {
+        throw new Error('invalid arguments: "extract_text" must be a boolean');
+      }
+      if (args.max_text_chars !== undefined && (typeof args.max_text_chars !== "number" || !Number.isInteger(args.max_text_chars))) {
+        throw new Error('invalid arguments: "max_text_chars" must be an integer');
+      }
+
+      const resolveOptions = await sessionResolveOptions(ctx, exec);
+      const target = await ctx.fs.resolve(args.file_path.trim(), resolveOptions);
+      const info = await ctx.fs.stat(target, exec.signal);
+      const displayPath = ctx.fs.processPath(target);
+      if (!info) {
+        ctx.emit("fs/observed", target, { kind: "absent" }, exec);
+        throw new Error(`cannot read "${displayPath}": not found`);
+      }
+      if (info.type !== "file") {
+        throw new Error(`cannot read "${displayPath}": not a regular file`);
+      }
+      if (info.size !== undefined && info.size > MAX_PDF_BYTES) {
+        throw new Error(
+          `cannot read "${displayPath}": file is ${(info.size / 1024 / 1024).toFixed(1)} MB, ` +
+            `超过 ${MAX_PDF_BYTES / 1024 / 1024} MB 上限`
+        );
+      }
+      const bytes = await ctx.fs.readBytes(target, exec.signal, MAX_PDF_BYTES);
+      ctx.emit("fs/observed", target, { kind: "present", version: info.version }, exec);
+
+      let parser;
+      try {
+        parser = new PDFParse({ data: bytes });
+      } catch (error) {
+        throw new Error(`cannot parse "${displayPath}": 初始化 PDF 解析器失败: ${error.message}`);
+      }
+
+      let text = "";
+      let textTruncated = false;
+      let pageCount = 0;
+      let metadata = {};
+      try {
+        const infoResult = await parser.getInfo();
+        pageCount = infoResult.info?.Pages != null ? Number(infoResult.info.Pages) : 0;
+        metadata = infoResult.metadata ?? {};
+        if (args.extract_text !== false) {
+          const textResult = await parser.getText();
+          text = textResult?.text ?? "";
+          const maxChars = args.max_text_chars ?? 50000;
+          if (maxChars > 0 && text.length > maxChars) {
+            text = text.slice(0, maxChars);
+            textTruncated = true;
+          }
+        }
+      } catch (error) {
+        throw new Error(`cannot parse "${displayPath}": PDF 内容提取失败: ${error.message}`);
+      }
+
+      return toJson({
+        path: displayPath,
+        fileName: displayPath.split(/[\\/]/).pop(),
+        fileSizeBytes: info.size !== undefined ? info.size : bytes.byteLength,
+        format: "pdf",
+        pageCount,
+        metadata,
+        text,
+        textLength: text.length,
+        textTruncated,
+      });
+    },
   };
 }

package/package.json

@@ -1,14 +1,22 @@
 {
   "name": "@bill_nee/dwg-reader",
-  "version": "1.2.0",
-  "description": "直接读取 AutoCAD DWG/DXF 图纸的 DSH 插件（DWG 基于 libredwg-web 0.8.2 WASM，DXF 基于 Python ezdxf 桥接，GPL-3.0）。",
+  "version": "1.4.0",
+  "description": "直接读取 AutoCAD DWG/DXF、PDF 与图片/截屏的 DSH 插件（DWG 基于 libredwg-web 0.8.2 WASM，DXF 基于 Python ezdxf 桥接，PDF 基于 pdf-parse，图片基于 Pillow+Tesseract OCR，GPL-3.0）。",
   "license": "GPL-3.0-only",
   "author": "MEB 工程造价项目助手 <meb@example.com>",
   "repository": {
     "type": "git",
     "url": "https://github.com/NIDEXIANG/dwg-reader-plugin-20261029.git"
   },
-  "keywords": ["dwg", "dwg-reader", "autocad", "cad", "dsh-plugin", "libredwg", "wasm"],
+  "keywords": [
+    "dwg",
+    "dwg-reader",
+    "autocad",
+    "cad",
+    "dsh-plugin",
+    "libredwg",
+    "wasm"
+  ],
   "type": "module",
   "engines": {
     "node": ">=20"
@@ -32,5 +40,8 @@
     "bundle": {
       "patch": "./cordis.patch.yml"
     }
+  },
+  "dependencies": {
+    "pdf-parse": "^2.4.5"
   }
 }

package/scripts/ocr_image.py

@@ -0,0 +1,70 @@
+#!/usr/bin/env python3
+"""对图片进行 OCR 文字提取（供 DSH dwg-reader 插件调用，直接调用 tesseract 命令行）。"""
+import sys
+import json
+import os
+import subprocess
+import tempfile
+
+IMAGE_PATH = sys.argv[1] if len(sys.argv) > 1 else None
+TESSERACT_CMD = sys.argv[2] if len(sys.argv) > 2 else r"D:\Tools\Tesseract-OCR\tesseract.exe"
+
+
+def main():
+    if not IMAGE_PATH or not os.path.exists(IMAGE_PATH):
+        print(json.dumps({"error": f"文件不存在: {IMAGE_PATH}"}))
+        sys.exit(1)
+
+    try:
+        from PIL import Image
+    except ImportError as exc:
+        print(json.dumps({"error": f"Pillow 未安装: {exc}"}))
+        sys.exit(1)
+
+    try:
+        image = Image.open(IMAGE_PATH)
+    except Exception as exc:
+        print(json.dumps({"error": f"无法打开图片: {exc}"}))
+        sys.exit(1)
+
+    width, height = image.size
+    fmt = image.format
+    mode = image.mode
+
+    # 调用 tesseract 命令行做 OCR
+    ocr_text = ""
+    ocr_error = None
+    try:
+        with tempfile.TemporaryDirectory() as tmpdir:
+            out_base = os.path.join(tmpdir, "ocr_out")
+            cmd = [TESSERACT_CMD, IMAGE_PATH, out_base, "-l", "chi_sim+eng"]
+            proc = subprocess.run(cmd, capture_output=True, text=True, timeout=120)
+            if proc.returncode != 0:
+                ocr_error = proc.stderr.strip() or f"tesseract 退出码 {proc.returncode}"
+            else:
+                out_txt = out_base + ".txt"
+                if os.path.exists(out_txt):
+                    with open(out_txt, "r", encoding="utf-8", errors="replace") as f:
+                        ocr_text = f.read().strip()
+    except Exception as exc:
+        ocr_error = str(exc)
+
+    result = {
+        "path": os.path.abspath(IMAGE_PATH),
+        "fileName": os.path.basename(IMAGE_PATH),
+        "fileSizeBytes": os.path.getsize(IMAGE_PATH),
+        "format": fmt,
+        "width": width,
+        "height": height,
+        "mode": mode,
+        "ocrText": ocr_text,
+        "ocrError": ocr_error,
+    }
+    # Windows 控制台默认 GBK，改为 UTF-8 输出到 stdout.buffer 避免编码错误
+    sys.stdout.buffer.write(json.dumps(result, ensure_ascii=False).encode("utf-8"))
+    sys.stdout.buffer.write(b"\n")
+    sys.stdout.buffer.flush()
+
+
+if __name__ == "__main__":
+    main()