dsh-data-cleaning-agent 0.8.5 → 0.8.6

package/CHANGELOG.md

@@ -2,6 +2,14 @@
 
 本文件记录 `dsh-data-cleaning-agent` 的版本变更。格式遵循 [Keep a Changelog](https://keepachangelog.com/zh-CN/1.1.0/)，版本号遵循 [SemVer](https://semver.org/lang/zh-CN/)。
 
+## [0.8.6] - 2026-09-06
+
+### Fixed
+- 上传/OCR 核验页使用完整原始清单，每页 20 行可翻页，显示序号与全列；不再把前 5 行摘要当成核验视图。
+- 匹配结果共用全量分页和中文字段展示，不再截断为前 8 行/列。
+- 核验阶段支持全量 CSV 和原值 JSON 下载，不触发 QCC 调用；CSV 对公式及长编号作文本保护。明细未完整载入时禁用核验下载。
+- OCR 完成后的引导明确真实核验路径、规则/体检/匹配步骤及自动回填，无需人工复制 dcq- 凭证，不将识别清单误报为匹配结果。
+
 ## [0.8.5] - 2026-09-06
 
 ### Fixed

package/README.en.md

@@ -2,7 +2,7 @@
 
 > A data cleaning & completion agent plugin for DeepSeek Harness: local CSV/XLSX/JSON engine plus optional Qichacha (QCC) MCP enterprise-data enrichment. Initiated and maintained by the Qichacha (QCC) team.
 >
-> Current source version / 当前源码版本: **0.8.5** (stable release)
+> Current source version / 当前源码版本: **0.8.6** (stable release)
 
 [![CI](https://github.com/duhu2000/dsh-data-cleaning-agent/actions/workflows/ci.yml/badge.svg)](https://github.com/duhu2000/dsh-data-cleaning-agent/actions/workflows/ci.yml)
 [![npm](https://img.shields.io/npm/v/dsh-data-cleaning-agent)](https://www.npmjs.com/package/dsh-data-cleaning-agent)

package/README.md

@@ -2,7 +2,7 @@
 
 > 在 DeepSeek Harness 中清洗、补全、画像企业名单数据的智能体插件：本地 CSV/XLSX/JSON 引擎 + 可选企查查（Qichacha/QCC）MCP 企业数据补全，由企查查（Qichacha/QCC）团队发起并维护。
 >
-> 当前源码版本 / Current source version: **0.8.5**（正式版本）
+> 当前源码版本 / Current source version: **0.8.6**（正式版本）
 
 [![CI](https://github.com/duhu2000/dsh-data-cleaning-agent/actions/workflows/ci.yml/badge.svg)](https://github.com/duhu2000/dsh-data-cleaning-agent/actions/workflows/ci.yml)
 [![npm](https://img.shields.io/npm/v/dsh-data-cleaning-agent)](https://www.npmjs.com/package/dsh-data-cleaning-agent)

package/lib/client.js

@@ -887,6 +887,7 @@ window.__ModuleLoader__.load({
       if (!from || !toId || String(fromId) === String(toId)) return runtimeFor(toId);
       const target = runtimeFor(toId);
       target.rows = from.rows;
+      target.sourceRows = from.sourceRows;
       target.headers = from.headers;
       target.source = from.source;
       target.lastCsv = from.lastCsv;
@@ -2681,6 +2682,7 @@ window.__ModuleLoader__.load({
     function applyParsed(result, actions, taskId = 'unassigned', source = {}) {
       const runtime = runtimeFor(taskId);
       runtime.rows = Array.isArray(result.rows) ? result.rows : [];
+      runtime.sourceRows = runtime.rows;
       runtime.headers = Array.isArray(result.headers) ? result.headers : [];
       runtime.source = {
         type: source.type || result.fmt || 'csv',
@@ -2712,6 +2714,7 @@ window.__ModuleLoader__.load({
       // 都以本次解析结果覆盖目标 runtime，避免复用已完成任务的一行旧数据。
       const target = runtimeFor(current.id);
       target.rows = Array.isArray(result.rows) ? result.rows : [];
+      target.sourceRows = target.rows;
       target.headers = Array.isArray(result.headers) ? result.headers : [];
       target.source = {
         type: source.type || result.fmt || 'csv',
@@ -2733,6 +2736,67 @@ window.__ModuleLoader__.load({
       return current;
     }
 
+    function reviewColumns(rows, headers = []) {
+      return [...new Set([...headers, ...rows.flatMap((row) => Object.keys(row || {}))])];
+    }
+
+    function reviewCsv(rows, headers) {
+      const cell = (value) => {
+        let text = value == null ? '' : typeof value === 'object' ? JSON.stringify(value) : String(value);
+        // 保留长编号的文本形式，阻止外部 OCR/企业字段成为 Excel 公式。
+        if (/^[\s]*[=+@-]/.test(text) || /^\d{15,}$/.test(text) || /^0\d+/.test(text)) text = "'" + text;
+        return '"' + text.replace(/"/g, '""') + '"';
+      };
+      return '\uFEFF' + [headers.map((key) => cell(resultFieldLabel(key))).join(','),
+        ...rows.map((row) => headers.map((key) => cell(row?.[key])).join(','))].join('\r\n');
+    }
+
+    /** 页内查看和核验下载均使用完整 rows，不使用 store 的五行摘要。 */
+    function DatasetReview({ rows, headers, title, expectedCount, onError }) {
+      const [page, setPage] = react.useState(0);
+      const pageSize = 20;
+      const total = rows.length;
+      const columns = reviewColumns(rows, headers);
+      const currentPage = Math.min(page, Math.max(0, Math.ceil(total / pageSize) - 1));
+      const start = currentPage * pageSize;
+      const download = (format) => {
+        try {
+          const content = format === 'csv' ? reviewCsv(rows, columns) : JSON.stringify(rows, null, 2);
+          const url = URL.createObjectURL(new Blob([content], { type: format === 'csv' ? 'text/csv;charset=utf-8' : 'application/json;charset=utf-8' }));
+          const anchor = document.createElement('a');
+          anchor.href = url;
+          anchor.download = title + '.' + format;
+          try { anchor.click(); } finally { setTimeout(() => URL.revokeObjectURL(url), 1000); }
+        } catch (error) { onError?.(error instanceof Error ? error.message : String(error)); }
+      };
+      const complete = expectedCount == null || expectedCount === total;
+      return h('section', { className: 'dcAgentSection', 'aria-label': title },
+        h('h3', null, title),
+        h('p', { className: 'dcAgentHint', role: 'status' }, complete
+          ? `共 ${total} 条，当前显示 ${total ? start + 1 : 0}–${Math.min(start + pageSize, total)} 条；共 ${columns.length} 列，可横向滚动。`
+          : `完整清单尚未就绪：预期 ${expectedCount} 条，当前载入 ${total} 条。请等待同步；刷新后若仍缺失，请重新上传源文件。`),
+        h('div', { className: 'dcAgentPreviewTable', style: { overflow: 'auto', maxHeight: '420px' } },
+          h('table', { className: 'dcAgentTable' },
+            h('thead', null, h('tr', null, h('th', null, '序号'), columns.map((key) => h('th', { key }, resultFieldLabel(key))))),
+            h('tbody', null, rows.slice(start, start + pageSize).map((row, index) => h('tr', { key: start + index },
+              h('td', null, String(start + index + 1)),
+              columns.map((key) => h('td', { key }, row?.[key] == null ? '' : typeof row[key] === 'object' ? JSON.stringify(row[key]) : String(row[key]))),
+            ))),
+          ),
+        ),
+        h('div', { className: 'dcAgentRow' },
+          h('button', { type: 'button', className: 'dcAgentButton', disabled: currentPage === 0, onClick: () => setPage(currentPage - 1) }, '上一页'),
+          h('span', null, `第 ${currentPage + 1} / ${Math.max(1, Math.ceil(total / pageSize))} 页`),
+          h('button', { type: 'button', className: 'dcAgentButton', disabled: start + pageSize >= total, onClick: () => setPage(currentPage + 1) }, '下一页'),
+        ),
+        h('div', { className: 'dcAgentRow' },
+          h('button', { type: 'button', className: 'dcAgentButton', disabled: !complete || !total, onClick: () => download('csv') }, `下载全部 ${total} 条 CSV`),
+          h('button', { type: 'button', className: 'dcAgentButton', disabled: !complete || !total, onClick: () => download('json') }, '下载原值 JSON'),
+        ),
+        h('p', { className: 'dcAgentHint' }, '下载包含全部行，不受分页影响。CSV 对公式及长编号作文本保护；JSON 保留原值。仅导出当前清单，不触发企查查调用。'),
+      );
+    }
+
     /** 右侧非模态工作台：中央区域始终保留 DSH 原生会话。 */
     function WorkbenchDrawer(props) {
       const { useStore, actions } = props;
@@ -2888,10 +2952,9 @@ window.__ModuleLoader__.load({
       const cachedTask = workflowTaskBySession.get(String(activeSessionId || 'unassigned')) ?? workflowTask;
       const runtimeKey = cachedTask?.id ?? `session:${activeSessionId || 'unassigned'}`;
       const runtime = runtimeFor(runtimeKey);
+      const stagedSource = runtimeFor(`session:${activeSessionId || 'unassigned'}`, false);
+      const sourceRows = stagedSource?.sourceRows ?? runtime.sourceRows ?? runtime.rows;
       const lastCsv = runtime.lastCsv;
-      const qccPreviewColumns = Array.isArray(qccRun?.rows) && qccRun.rows.length
-        ? Object.keys(qccRun.rows[0] || {}).slice(0, 8)
-        : [];
       const fieldByPattern = (pattern) => runtime.headers.find((field) => pattern.test(field)) ?? null;
       const phoneField = fieldByPattern(/^(phone|mobile|tel|telephone|联系电话|手机号码|手机号)$/i);
       const amountField = fieldByPattern(/^(amount|price|金额|注册资本)$/i);
@@ -3604,17 +3667,11 @@ window.__ModuleLoader__.load({
               stat('未匹配', qccRun.summary?.unresolved ?? 0, (qccRun.summary?.unresolved ?? 0) > 0 ? 'warn' : null),
               stat('失败', qccRun.summary?.failed ?? 0, (qccRun.summary?.failed ?? 0) > 0 ? 'bad' : null),
             ),
-            Array.isArray(qccRun.rows) && qccRun.rows.length ? h('div', { className: 'dcAgentTableWrap' },
-              h('table', { className: 'dcAgentTable', 'aria-label': '匹配补全结果预览' },
-                h('thead', null, h('tr', null,
-                  qccPreviewColumns.map((key) => h('th', { key }, resultFieldLabel(key))),
-                )),
-                h('tbody', null, qccRun.rows.slice(0, 8).map((row, index) => h('tr', { key: `${index}-${row?.qcc_credit_no || row?.credit_no || ''}` },
-                  qccPreviewColumns.map((key) => h('td', { key }, String(row?.[key] ?? ''))),
-                ))),
-              ),
-              h('p', { className: 'dcAgentHint' }, `显示前 ${Math.min(8, qccRun.rows.length)} 行；完整结果在下载阶段生成耐久制品。`),
-            ) : null,
+            Array.isArray(qccRun.rows) && qccRun.rows.length ? h(DatasetReview, {
+              key: qccRun.runId, rows: qccRun.rows, headers: runtime.headers,
+              title: '企查查匹配补全结果', expectedCount: qccRun.summary?.totalRows,
+              onError: actions.setError,
+            }) : null,
             (qccRun.reviewQueue || []).map((item) => h('div', { key: item.companyName, className: 'dcAgentSection' },
               h('h3', null, `待核验：${item.companyName}`),
               item.candidates.map((candidate) => h('div', { key: candidate.creditNo, className: 'dcAgentCandidate' },
@@ -3716,15 +3773,16 @@ window.__ModuleLoader__.load({
             (dataset.headers || []).slice(0, 12).map((name) => h('span', { key: name, className: 'dcAgentChip' }, name)),
             (dataset.headers || []).length > 12 ? h('span', { className: 'dcAgentChip' }, `+${dataset.headers.length - 12} 列`) : null,
           ) : null,
-          dataset ? h('div', { className: 'dcAgentPreviewTable', 'aria-label': '数据预览' },
-            h('table', { className: 'dcAgentTable' },
-              h('thead', null, h('tr', null, (dataset.headers || []).slice(0, 8).map((name) => h('th', { key: name }, name)))),
-              h('tbody', null, (dataset.preview || []).slice(0, 5).map((row, index) => h('tr', { key: index }, (dataset.headers || []).slice(0, 8).map((name) => h('td', { key: name }, String(row?.[name] ?? '')))))),
-            ),
-          ) : null,
+          dataset ? h(DatasetReview, {
+            key: runtimeKey, rows: sourceRows, headers: dataset.headers || [],
+            expectedCount: dataset.rowCount,
+            title: runtime.source?.type === 'image' ? '图片识别原始名单（尚未匹配）' : '导入原始清单',
+            onError: actions.setError,
+          }) : null,
+          hasData ? h('p', { className: 'dcAgentHint' }, '请核对完整名单。识别清单不代表已完成企查查匹配。核对后进入「字段映射与规则」，确认规则生成质量体检；再进入「数据匹配」估算调用量并确认账号额度，点击「生成可编辑任务说明」自动回填中央对话框，无需手工复制任务凭证。') : null,
           h('div', { className: 'dcAgentRow' },
             h('button', { type: 'button', className: 'dcAgentButton is-primary', disabled: busy, 'aria-label': '解析数据', onClick: handleParse }, busy ? '解析中…' : '解析数据'),
-            hasData ? h('button', { type: 'button', className: 'dcAgentButton', onClick: () => actions.setStep('rules') }, '下一步：字段映射与规则') : null,
+            hasData ? h('button', { type: 'button', className: 'dcAgentButton', disabled: busy || sourceRows.length !== dataset.rowCount, onClick: () => actions.setStep('rules') }, '已核对清单，下一步：字段映射与规则') : null,
           ),
         );
       }
@@ -3897,6 +3955,9 @@ window.__ModuleLoader__.load({
     // 测试用纯函数，不构成 Host / DSH 稳定 API。
     exports.__testing = {
       buildTaskPrompt,
+      DatasetReview,
+      reviewColumns,
+      reviewCsv,
       clearCleaningDraft,
       entriesToDataset,
       extractPromptEntries,

package/lib/image-intake.js

@@ -580,7 +580,7 @@ export function registerImageIntakeTool(tools, store) {
       },
       render: (_args, value) => [{
         type: 'text',
-        text: `企查查智能文档解析已识别图片企业名单：${value.entryCount} 条，已同步回数据清洗补全工作台等待核验。`,
+        text: `企查查智能文档解析已识别图片企业名单：${value.entryCount} 条，已同步回数据清洗补全工作台等待核验。请在「上传数据」查看原始清单全部行或下载全量 CSV/JSON，核对后进入「字段映射与规则」确认规则并生成质量体检，再到「数据匹配」估算调用量并确认账号额度，点击「生成可编辑任务说明」自动回填中央对话框。无需手工复制 dcq- 凭证或重新粘贴名单；识别名单不代表已匹配，尚未查询时不要编造匹配摘要。`,
       }],
     },
     async execute(args, exec) {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "dsh-data-cleaning-agent",
-  "version": "0.8.5",
+  "version": "0.8.6",
   "description": "Clean, complete, and profile enterprise name lists in DeepSeek Harness — a data cleaning & completion agent plugin with local CSV/XLSX/JSON engine and optional Qichacha (QCC) MCP enrichment. Maintained by Qichacha/QCC.",
   "type": "module",
   "main": "lib/index.js",