bugfix 爱企查 法人 地址等

demo/Untitled-1.ini

@@ -0,0 +1,6 @@
+1760/1年
+500 /2年
+
+
+
+AI智能建档APP

script-spider-all-group.js

@@ -0,0 +1,111 @@
+// ==UserScript==
+// @name         集团客户数据抓取工具
+// @namespace    http://tampermonkey.net/
+// @version      1.0
+// @description  自动抓取集团客户表格数据并导出为CSV文件
+// @author       You
+// @match        file:///*D:/yd-other/爱企查/demo/group_all.html*
+// @grant        none
+// ==/UserScript==
+
+(function() {
+    'use strict';
+
+    // 创建黄色圆点按钮
+    function createYellowDotButton() {
+        const button = document.createElement('div');
+        button.style.position = 'fixed';
+        button.style.right = '20px';
+        button.style.top = '20px';
+        button.style.width = '30px';
+        button.style.height = '30px';
+        button.style.backgroundColor = 'yellow';
+        button.style.borderRadius = '50%';
+        button.style.cursor = 'pointer';
+        button.style.zIndex = '9999';
+        button.style.boxShadow = '0 0 5px rgba(0,0,0,0.5)';
+        button.title = '点击抓取表格数据';
+        
+        button.addEventListener('click', scrapeTableData);
+        document.body.appendChild(button);
+    }
+
+    // 抓取表格数据并导出为CSV
+    async function scrapeTableData() {
+        try {
+            // 获取所有表格行数据
+            const tableRows = document.querySelectorAll('.arco-table-tr:not(.arco-table-empty-row)');
+            
+            if (tableRows.length <= 1) { // 只有表头或者没有数据
+                alert('未找到表格数据');
+                return;
+            }
+
+            // 提取表头
+            const headers = [];
+            const headerRow = tableRows[0];
+            const headerCells = headerRow.querySelectorAll('.arco-table-th .arco-table-th-item-title');
+            
+            headerCells.forEach(cell => {
+                headers.push(cell.textContent.trim());
+            });
+
+            // 提取数据行
+            const dataRows = [];
+            for (let i = 1; i < tableRows.length; i++) {
+                const row = tableRows[i];
+                const cells = row.querySelectorAll('.arco-table-td .arco-table-cell-wrap-value');
+                const rowData = [];
+                
+                cells.forEach(cell => {
+                    // 处理嵌套的typography元素
+                    const typographyElement = cell.querySelector('.arco-typography');
+                    const text = typographyElement ? typographyElement.textContent.trim() : cell.textContent.trim();
+                    rowData.push(text);
+                });
+                
+                // 如果行中有数据，则添加到结果中
+                if (rowData.some(cell => cell !== '')) {
+                    dataRows.push(rowData);
+                }
+            }
+
+            // 构建CSV内容
+            let csvContent = headers.join(',') + '\n';
+            dataRows.forEach(row => {
+                csvContent += row.map(cell => `"${cell.replace(/"/g, '""')}"`).join(',') + '\n';
+            });
+
+            // 下载CSV文件
+            downloadCSV(csvContent, '集团客户数据.csv');
+            
+            alert(`成功抓取 ${dataRows.length} 条记录`);
+            
+        } catch (error) {
+            console.error('抓取数据时发生错误:', error);
+            alert('抓取数据时发生错误，请检查控制台');
+        }
+    }
+
+    // 下载CSV文件
+    function downloadCSV(csvContent, filename) {
+        const blob = new Blob(['\uFEFF' + csvContent], { type: 'text/csv;charset=utf-8;' });
+        const url = URL.createObjectURL(blob);
+        
+        const link = document.createElement('a');
+        link.setAttribute('href', url);
+        link.setAttribute('download', filename);
+        link.style.visibility = 'hidden';
+        
+        document.body.appendChild(link);
+        link.click();
+        document.body.removeChild(link);
+    }
+
+    // 页面加载完成后创建按钮
+    if (document.readyState === 'loading') {
+        document.addEventListener('DOMContentLoaded', createYellowDotButton);
+    } else {
+        createYellowDotButton();
+    }
+})();

script2-1-qcc-acq.js

@@ -213,21 +213,47 @@
 
     // 获取法定代表人
     getLegalRepresentative() {
-      const legalElements = Array.from(
-        this.table.querySelectorAll("td")
-      ).filter((td) => ToolUtils.cleanText(td.textContent) === "法定代表人");
 
-      if (legalElements.length > 0) {
+      // First check for 经营者 (operator) which is used for individual businesses
-        const valueCell = legalElements[0].nextElementSibling;
+      const operatorElements = Array.from(
-        if (valueCell && valueCell.classList.contains("image-text-content")) {
+        this.table.querySelectorAll("td")
-          const nameElement = valueCell.querySelector(".person-name-warp a");
+      ).filter((td) => (ToolUtils.cleanText(td.textContent) === "经营者") || 
-          if (nameElement) {
+      ToolUtils.cleanText(td.textContent) === "执行事务合伙人" ||
-            return ToolUtils.cleanText(nameElement.textContent);
+      ToolUtils.cleanText(td.textContent) === "负责人" ||
+      ToolUtils.cleanText(td.textContent) === "法定代表人"
+    );
+
+      if (operatorElements.length > 0) {
+        const valueCell = operatorElements[0].nextElementSibling;
+        if (valueCell) {
+          // Check if it has the image-text-content structure
+          if (valueCell.classList.contains("image-text-content")) {
+            const nameElement = valueCell.querySelector(".person-name-warp a");
+            if (nameElement) {
+              return ToolUtils.cleanText(nameElement.textContent);
+            }
+            return ToolUtils.cleanText(valueCell.textContent);
           }
+          // Regular structure
           return ToolUtils.cleanText(valueCell.textContent);
         }
       }
+      // const legalElements = Array.from(
+      //   this.table.querySelectorAll("td")
+      // ).filter((td) => ToolUtils.cleanText(td.textContent) === "法定代表人");
 
+      // if (legalElements.length > 0) {
+      //   const valueCell = legalElements[0].nextElementSibling;
+      //   if (valueCell && valueCell.classList.contains("image-text-content")) {
+      //     const nameElement = valueCell.querySelector(".person-name-warp a");
+      //     if (nameElement) {
+      //       return ToolUtils.cleanText(nameElement.textContent);
+      //     }
+      //     return ToolUtils.cleanText(valueCell.textContent);
+      //   }
+      // }
+
+      // 备用查找方式
       const titleElements = Array.from(
         this.table.querySelectorAll("td")
       ).filter((td) => td.textContent.includes("法定代表人"));
@@ -388,6 +414,20 @@
         return match ? match[0] : null;
       }
 
+      // 尝试查找其他可能包含员工数的字段
+      const employeeElements = Array.from(
+        this.table.querySelectorAll("td")
+      ).filter((td) => 
+        td.textContent.includes("人员规模") || 
+        td.textContent.includes("员工人数")
+      );
+
+      if (employeeElements.length > 0 && employeeElements[0].nextElementSibling) {
+        const valueCell = employeeElements[0].nextElementSibling;
+        const rawText = valueCell.textContent.replace(/[\r\n\t]/g, "").trim();
+        return rawText;
+      }
+
       return null;
     }
 
@@ -421,6 +461,12 @@
         }
       }
 
+      // 如果没有核准日期，尝试使用成立日期作为替代
+      const establishmentDate = this.getOptimizedValue("成立日期");
+      if (establishmentDate && /^\d{4}-\d{2}-\d{2}$/.test(establishmentDate)) {
+        return establishmentDate;
+      }
+
       return null;
     }
 
@@ -761,7 +807,36 @@ class NationalCreditParser {
 
   // 获取住所/注册地址
   getRegisteredAddress() {
-      return this.getOptimizedValue("住所");
+      // return this.getOptimizedValue("住所");
+      // 主要查找方式
+      const address = this.getOptimizedValue("注册地址");
+      if (address) {
+        return address;
+      }
+
+      // 备用查找方式
+      const alternativeElements = Array.from(
+        this.table.querySelectorAll("td")
+      ).filter((td) => 
+        ToolUtils.cleanText(td.textContent).includes("住所") ||
+        ToolUtils.cleanText(td.textContent).includes("营业地址") ||
+        ToolUtils.cleanText(td.textContent).includes("办公地址")
+      );
+
+      if (alternativeElements.length > 0 && alternativeElements[0].nextElementSibling) {
+        const valueCell = alternativeElements[0].nextElementSibling;
+        const valueElement =
+          valueCell.querySelector(".enter-bg-ele") ||
+          valueCell.querySelector(".addr-enter-bg-ele") ||
+          valueCell;
+          
+        const addressText = ToolUtils.cleanText(valueElement.textContent);
+        if (addressText) {
+          return addressText;
+        }
+      }
+
+      return null;
   }
 
   // 获取经营范围