@liuhange/dsh-data-masking 2.0.0 → 2.0.2

package/lib/index.js

@@ -1,131 +1,115 @@
-import { defineTool } from "@deepseek-ai/dsh-tools";
-import * as fs from "fs";
-import * as path from "path";
-import { AuditLogger, BusinessRulesLoader, FileFormatAdapter, PathValidator, ReportGenerator } from "@deepseek-ai/dsh-data-asset-shared";
-//#region lib/types/sensitiveFieldScanner.js
-var SensitiveFieldScanner = class {
-	scan(lines, sensitivePatterns) {
-		const fields = [];
-		const typeCounts = {};
-		const patternEntries = Object.entries(sensitivePatterns);
-		for (let lineIndex = 0; lineIndex < lines.length; lineIndex++) {
-			const line = lines[lineIndex] ?? "";
-			for (const [type, config] of patternEntries) {
-				const regex = new RegExp(config.pattern, "g");
-				let match;
-				while ((match = regex.exec(line)) !== null) {
-					fields.push({
-						type,
-						value: match[0],
-						line: lineIndex + 1,
-						column: match.index + 1
-					});
-					typeCounts[type] = (typeCounts[type] ?? 0) + 1;
-				}
-			}
-		}
-		return {
-			fields,
-			typeCounts
-		};
-	}
-};
-//#endregion
-//#region lib/types/maskingStrategyExecutor.js
-var MaskingStrategyExecutor = class {
-	execute(value, _type, strategy) {
-		switch (this.resolveStrategy(strategy)) {
-			case "FULL": return "***";
-			case "PARTIAL": return this.applyPartial(value);
-			case "GENERALIZE": return "<脱敏数据>";
-		}
-	}
-	resolveStrategy(strategy) {
-		if (strategy === "FULL" || strategy === "PARTIAL" || strategy === "GENERALIZE") return strategy;
-		return "PARTIAL";
-	}
-	applyPartial(value, keepPrefix = 3, keepSuffix = 4) {
-		if (value.length <= keepPrefix + keepSuffix) return "***";
-		return value.slice(0, keepPrefix) + "***" + value.slice(-keepSuffix);
-	}
-};
-//#endregion
-//#region lib/types/index.js
-const name = "data-masking";
-const inject = ["tools"];
-function apply(ctx) {
-	const loader = new BusinessRulesLoader();
-	const formatAdapter = new FileFormatAdapter();
-	const reportGenerator = new ReportGenerator();
-	const pathValidator = new PathValidator();
-	const auditLogger = new AuditLogger();
-	const scanner = new SensitiveFieldScanner();
-	const strategyExecutor = new MaskingStrategyExecutor();
-	ctx.tools.register(defineTool({
-		name: "mask_sensitive_data",
-		description: "识别并脱敏数据中的敏感字段（身份证、手机号、银行卡、邮箱）",
-		parameters: {
-			filePath: {
-				type: "string",
-				required: true,
-				description: "待处理文件的路径"
-			},
-			strategy: {
-				type: "string",
-				description: "脱敏策略: FULL | PARTIAL | GENERALIZE"
-			}
-		},
-		output: {
-			schema: { type: "string" },
-			render: (_args, value) => [{
-				type: "text",
-				text: value
-			}]
-		},
-		async execute(args) {
-			const { config } = loader.load();
-			const workingDir = process.cwd();
-			const filePath = args.filePath;
-			const strategy = args.strategy ?? "PARTIAL";
-			try {
-				pathValidator.validate(filePath, workingDir);
-			} catch {
-				return `错误：路径不合法 - ${filePath}`;
-			}
-			const fullPath = path.resolve(filePath);
-			if (!fs.existsSync(fullPath)) return `错误：文件不存在 - ${fullPath}`;
-			const readResult = await formatAdapter.read(fullPath);
-			const lines = readResult.lines;
-			const scanResult = scanner.scan(lines, config.sensitivePatterns);
-			let maskedLines = [...lines];
-			for (const field of scanResult.fields) {
-				const fieldType = field.type;
-				const fieldStrategy = config.sensitivePatterns[fieldType].level ?? strategy;
-				const maskedValue = strategyExecutor.execute(field.value, fieldType, fieldStrategy);
-				maskedLines = maskedLines.map((line) => line.replace(field.value, maskedValue));
-			}
-			const ext = path.extname(fullPath);
-			const baseName = path.basename(fullPath, ext);
-			const outputPath = path.join(path.dirname(fullPath), `${baseName}_masked${ext}`);
-			await formatAdapter.write(outputPath, maskedLines, readResult.format);
-			const report = reportGenerator.generateMaskingReport({
-				inputPath: fullPath,
-				outputPath,
-				strategy: strategyExecutor.resolveStrategy(strategy),
-				findings: scanResult.fields,
-				fieldTypeCounts: scanResult.typeCounts
-			});
-			auditLogger.log({
-				pluginName: "data-masking",
-				operation: "mask_sensitive_data",
-				inputPath: fullPath,
-				outputPath,
-				result: "SUCCESS"
-			});
-			return report;
-		}
-	}));
-	console.log("[data-masking] 脱敏插件已加载");
+import { defineTool } from '@deepseek-ai/dsh-tools';
+import * as fs from 'fs';
+import * as path from 'path';
+import { BusinessRulesLoader, FileFormatAdapter, ReportGenerator, PathValidator, AuditLogger, } from '@liuhange/dsh-data-asset-shared';
+import { SensitiveFieldScanner } from './sensitiveFieldScanner.js';
+import { MaskingStrategyExecutor } from './maskingStrategyExecutor.js';
+import { AdvancedMaskingExecutor } from './advancedMaskingExecutor.js';
+import { defaultAdvancedMaskingConfig } from './defaultAdvancedMaskingConfig.js';
+export const name = 'data-masking';
+export const inject = ['tools'];
+export function apply(ctx) {
+    const loader = new BusinessRulesLoader();
+    const formatAdapter = new FileFormatAdapter();
+    const reportGenerator = new ReportGenerator();
+    const pathValidator = new PathValidator();
+    const auditLogger = new AuditLogger();
+    const scanner = new SensitiveFieldScanner();
+    const strategyExecutor = new MaskingStrategyExecutor();
+    const advancedExecutor = new AdvancedMaskingExecutor();
+    ctx.tools.register(defineTool({
+        name: 'mask_sensitive_data',
+        description: '识别并脱敏数据中的敏感字段（身份证、手机号、银行卡、邮箱），支持高级算法（FPE/k-匿名/差分隐私/哈希）',
+        parameters: {
+            filePath: { type: 'string', required: true, description: '待处理文件的路径' },
+            strategy: {
+                type: 'string',
+                description: '脱敏策略: FULL | PARTIAL | GENERALIZE',
+            },
+            algorithm: {
+                type: 'string',
+                description: '高级脱敏算法: FPE | k-anonymity | differential-privacy | hash',
+            },
+            fieldName: {
+                type: 'string',
+                description: '字段名（用于字段级算法配置覆盖）',
+            },
+        },
+        output: {
+            schema: { type: 'string' },
+            render: (_args, value) => [{ type: 'text', text: value }],
+        },
+        async execute(args) {
+            const { config } = loader.load();
+            const workingDir = process.cwd();
+            const filePath = args.filePath;
+            const strategy = args.strategy ?? 'PARTIAL';
+            const algorithm = args.algorithm;
+            const fieldName = args.fieldName;
+            try {
+                pathValidator.validate(filePath, workingDir);
+            }
+            catch {
+                return `错误：路径不合法 - ${filePath}`;
+            }
+            const fullPath = path.resolve(filePath);
+            if (!fs.existsSync(fullPath)) {
+                return `错误：文件不存在 - ${fullPath}`;
+            }
+            if (algorithm) {
+                const advancedConfig = config.advancedMasking ?? defaultAdvancedMaskingConfig;
+                const readResult = await formatAdapter.read(fullPath);
+                const lines = readResult.lines;
+                const maskedLines = [];
+                for (const line of lines) {
+                    const result = advancedExecutor.execute(line, algorithm, fieldName, advancedConfig);
+                    maskedLines.push(typeof result.maskedData === 'string' ? result.maskedData : JSON.stringify(result.maskedData));
+                }
+                const ext = path.extname(fullPath);
+                const baseName = path.basename(fullPath, ext);
+                const outputPath = path.join(path.dirname(fullPath), `${baseName}_masked${ext}`);
+                await formatAdapter.write(outputPath, maskedLines, readResult.format);
+                auditLogger.log({
+                    pluginName: 'data-masking',
+                    operation: 'mask_sensitive_data',
+                    inputPath: fullPath,
+                    outputPath,
+                    result: 'SUCCESS',
+                });
+                return `高级脱敏完成（算法: ${algorithm}）\n输入: ${fullPath}\n输出: ${outputPath}\n处理行数: ${maskedLines.length}`;
+            }
+            const readResult = await formatAdapter.read(fullPath);
+            const lines = readResult.lines;
+            const scanResult = scanner.scan(lines, config.sensitivePatterns);
+            let maskedLines = [...lines];
+            for (const field of scanResult.fields) {
+                const fieldType = field.type;
+                const patternConfig = config.sensitivePatterns[fieldType];
+                const fieldStrategy = patternConfig.level ?? strategy;
+                const maskedValue = strategyExecutor.execute(field.value, fieldType, fieldStrategy);
+                maskedLines = maskedLines.map(line => line.replace(field.value, maskedValue));
+            }
+            const ext = path.extname(fullPath);
+            const baseName = path.basename(fullPath, ext);
+            const outputPath = path.join(path.dirname(fullPath), `${baseName}_masked${ext}`);
+            await formatAdapter.write(outputPath, maskedLines, readResult.format);
+            const report = reportGenerator.generateMaskingReport({
+                inputPath: fullPath,
+                outputPath,
+                strategy: strategyExecutor.resolveStrategy(strategy),
+                findings: scanResult.fields,
+                fieldTypeCounts: scanResult.typeCounts,
+            });
+            auditLogger.log({
+                pluginName: 'data-masking',
+                operation: 'mask_sensitive_data',
+                inputPath: fullPath,
+                outputPath,
+                result: 'SUCCESS',
+            });
+            return report;
+        },
+    }));
+    console.log('[data-masking] 脱敏插件已加载');
 }
-//#endregion
-export { apply, inject, name };
+//# sourceMappingURL=index.js.map

package/lib/invariant.js

@@ -1,5 +1,4 @@
-//#region lib/types/invariant.js
-const invariant = "data-masking";
-function install() {}
-//#endregion
-export { install, invariant };
+export const invariant = 'data-masking';
+export function install() {
+}
+//# sourceMappingURL=invariant.js.map

package/lib/types/maskingStrategyExecutor.d.ts

@@ -1,4 +1,4 @@
-import type { MaskingStrategy, SensitiveFieldType } from '@deepseek-ai/dsh-data-asset-shared';
+import type { MaskingStrategy, SensitiveFieldType } from '@liuhange/dsh-data-asset-shared';
 export declare class MaskingStrategyExecutor {
     execute(value: string, _type: SensitiveFieldType, strategy: MaskingStrategy): string;
     resolveStrategy(strategy: MaskingStrategy): MaskingStrategy;

package/lib/types/sensitiveFieldScanner.d.ts

@@ -1,4 +1,4 @@
-import type { SensitiveField, SensitivePatterns } from '@deepseek-ai/dsh-data-asset-shared';
+import type { SensitiveField, SensitivePatterns } from '@liuhange/dsh-data-asset-shared';
 export interface ScanResult {
     fields: SensitiveField[];
     typeCounts: Record<string, number>;

package/package.json

@@ -1,13 +1,13 @@
-{
+{
     "name":  "@liuhange/dsh-data-masking",
     "description":  "Data masking plugin: sensitive field identification and graded masking (FULL/PARTIAL/GENERALIZE)",
-    "version":  "2.0.0",
+    "version": "2.0.2",
     "publishConfig":  {
                           "access":  "public"
                       },
     "repository":  {
                        "type":  "git",
-                       "url":  "git+https://github.com/deepseek-ai/deepseek-harness.git",
+                       "url":  "git+https://github.com/liuhange789/data-asset-inspector.git",
                        "directory":  "packages/data-asset/data-masking"
                    },
     "type":  "module",
@@ -31,9 +31,14 @@
                   "lib/types/**/*.d.ts"
               ],
     "license":  "MIT",
+    "scripts": {
+      "typecheck": "tsc --noEmit",
+      "test": "vitest run",
+      "build": "tsc"
+    },
     "peerDependencies":  {
-                             "@deepseek-ai/cordis":  "^0.1.0",
-                             "@deepseek-ai/dsh-tools":  "^0.1.0",
+                             "@deepseek-ai/cordis": "^4.0.0",
+                             "@deepseek-ai/dsh-tools": "0.0.1-rc.1",
                              "@liuhange/dsh-data-asset-shared":  "^2.0.0"
                          }
 }