@dommaker/harness 1.4.0 → 1.6.0

package/src/cli/session-mining/text.ts

@@ -1,117 +0,0 @@
-/**
- * 分词 / 停用词 / 相似度 / 文本清洗（工单 19-C）
- *
- * 合并 analyze-sessions 与 update-user-model 的重复实现。
- */
-
-/** N-gram 停用词表（中英） */
-export const STOP_WORDS = new Set([
-  '的', '了', '在', '是', '我', '有', '和', '就', '不', '人', '都', '一',
-  '一个', '上', '也', '很', '到', '说', '要', '去', '你', '会', '着',
-  '没有', '看', '好', '自己', '这', '他', '她', '它', '们', '那',
-  'the', 'a', 'an', 'is', 'are', 'was', 'were', 'be', 'been', 'being',
-  'have', 'has', 'had', 'do', 'does', 'did', 'will', 'would', 'could',
-  'should', 'may', 'might', 'can', 'shall', 'to', 'of', 'in', 'for',
-  'on', 'with', 'at', 'by', 'from', 'as', 'into', 'through', 'during',
-  'before', 'after', 'above', 'below', 'between', 'under', 'again',
-  'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why',
-  'how', 'all', 'both', 'each', 'few', 'more', 'most', 'other', 'some',
-  'such', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 'just',
-  '因为', '所以', '但是', '而且', '如果', '虽然', '可以', '这个', '那个',
-  '什么', '怎么', '哪', '吗', '吧', '呢', '啊', '嗯', '哦',
-]);
-
-const CN_CHAR = /[\u4e00-\u9fff\u3400-\u4dbf]/;
-
-/**
- * 中英文混合分词：中文产出 bigram/trigram（短序列保留整体），
- * 连续的拉丁/数字/路径字符作为整体 token。
- */
-export function tokenize(text: string): string[] {
-  const tokens: string[] = [];
-  let current = '';
-
-  for (const char of text) {
-    if (CN_CHAR.test(char)) {
-      current += char;
-    } else if (/[a-zA-Z0-9_\-/.]/.test(char)) {
-      current += char;
-    } else {
-      pushToken(current, tokens);
-      current = '';
-    }
-  }
-  pushToken(current, tokens);
-
-  return tokens;
-}
-
-function pushToken(current: string, tokens: string[]): void {
-  if (current.length < 2) return;
-  if (CN_CHAR.test(current)) {
-    for (let i = 0; i <= current.length - 2; i++) {
-      tokens.push(current.slice(i, i + 2));
-    }
-    for (let i = 0; i <= current.length - 3; i++) {
-      tokens.push(current.slice(i, i + 3));
-    }
-    if (current.length <= 6) tokens.push(current);
-  } else {
-    tokens.push(current);
-  }
-}
-
-/**
- * 基于 tokenize 的 Jaccard 相似度
- */
-export function jaccardSimilarity(a: string, b: string): number {
-  const setA = new Set(tokenize(a));
-  const setB = new Set(tokenize(b));
-  const intersection = new Set([...setA].filter(x => setB.has(x)));
-  const union = new Set([...setA, ...setB]);
-  return union.size === 0 ? 0 : intersection.size / union.size;
-}
-
-/**
- * 仅取汉字字符集的 Jaccard 相似度（概念聚类用）
- */
-export function jaccardChinese(a: string, b: string): number {
-  const charsA = new Set(a.split('').filter(c => /[\u4e00-\u9fff]/.test(c)));
-  const charsB = new Set(b.split('').filter(c => /[\u4e00-\u9fff]/.test(c)));
-  if (charsA.size === 0 || charsB.size === 0) return 0;
-  const intersection = new Set([...charsA].filter(x => charsB.has(x)));
-  const union = new Set([...charsA, ...charsB]);
-  return intersection.size / union.size;
-}
-
-/**
- * 剥离代码块 / 行内代码 / JSON 片段，保留自然语言
- */
-export function stripCodeBlocks(text: string): string {
-  return text
-    .replace(/```[\s\S]*?```/g, '')
-    .replace(/`[^`]+`/g, '')
-    .replace(/\{[^{}]*"[^"]+"\s*:\s*"[^"]*"[^}]*\}/g, '')
-    .replace(/\[[\s\S]*?\]/g, '')
-    .replace(/\\n/g, ' ')
-    .replace(/[{}[\]"':,\\]+/g, ' ');
-}
-
-/** 是否纯标点/空白 */
-export function isPunctuation(str: string): boolean {
-  return /^[，。！？、；：""''（）[\]【】「」『』《》〈〉\s]+$/.test(str);
-}
-
-/** 是否含语义内容（汉字，或 ≥7 字符的合法英文标识词） */
-export function hasSemanticContent(phrase: string): boolean {
-  if (/[\u4e00-\u9fff]/.test(phrase)) return true;
-  if (phrase.length >= 7 && /^[a-zA-Z][a-zA-Z0-9_]*[a-zA-Z]$/.test(phrase)) return true;
-  return false;
-}
-
-/** 是否代码噪声（短拉丁片段 / 常见 JSON 字段名碎片） */
-export function isCodeNoise(phrase: string): boolean {
-  if (!/[\u4e00-\u9fff]/.test(phrase) && phrase.length <= 6) return true;
-  if (/^(?:clla|cood|odde|oppe|laaw|peen|enne|ennc|ddec|deco|ecod|code|type|role|name|text|file|path|tool|user|session)$/i.test(phrase)) return true;
-  return false;
-}

package/src/cli/session-mining/transcript.ts

@@ -1,125 +0,0 @@
-/**
- * Claude Code transcript（.jsonl）读取（工单 19-C）
- *
- * 原 analyze-sessions.findSessions 与 update-user-model.findNewSessions
- * 各持一份解析逻辑，收敛于此。
- */
-
-import * as fs from 'fs';
-import * as path from 'path';
-import { readJsonl } from '../../utils/jsonl';
-
-export interface MinedTurn {
-  role: string;
-  content: string;
-}
-
-export interface MinedSession {
-  /** 会话 ID（文件名去掉 .jsonl，截断至 40 字符） */
-  id: string;
-  /** 文件修改日期（YYYY-MM-DD） */
-  date: string;
-  /** 修改时间戳（毫秒） */
-  mtimeMs: number;
-  turns: MinedTurn[];
-  /** 出现过的工具调用名（去重，保持出现顺序） */
-  toolCalls: string[];
-}
-
-function extractText(content: unknown): string {
-  if (typeof content === 'string') return content;
-  if (Array.isArray(content)) {
-    return content
-      .filter((b: any) => b.type === 'text')
-      .map((b: any) => b.text)
-      .join(' ');
-  }
-  return '';
-}
-
-/**
- * 扫描目录下的 .jsonl transcript 文件
- *
- * 不可读/损坏的行静默跳过（与原实现一致）。
- */
-export interface TranscriptFilter {
-  /** 只读 mtimeMs >= since 的文件（stat 级过滤，命中即不 parse）；缺省不过滤 */
-  since?: number;
-  /** 按会话 ID（文件名去 .jsonl、截断 40 字符）排除（文件名级过滤，不 stat 不 parse）；缺省不排除 */
-  excludeIds?: string[];
-}
-
-export function readTranscriptSessions(dir: string, filter: TranscriptFilter = {}): MinedSession[] {
-  const sessions: MinedSession[] = [];
-  const exclude = filter.excludeIds && filter.excludeIds.length > 0
-    ? new Set(filter.excludeIds)
-    : undefined;
-
-  let files: string[];
-  try {
-    files = fs.readdirSync(dir);
-  } catch {
-    return sessions;
-  }
-
-  for (const file of files) {
-    if (!file.endsWith('.jsonl')) continue;
-    const id = file.replace('.jsonl', '').slice(0, 40);
-    if (exclude?.has(id)) continue;
-    const filePath = path.join(dir, file);
-
-    let stat: fs.Stats;
-    try {
-      stat = fs.statSync(filePath);
-    } catch {
-      continue;
-    }
-    if (filter.since !== undefined && stat.mtimeMs < filter.since) continue;
-
-    const turns: MinedTurn[] = [];
-    const toolCalls: string[] = [];
-
-    let records: Record<string, any>[];
-    try {
-      // 损坏行静默跳过（坏行策略 skip，与原实现一致，harness#82 走 jsonl 正本）；
-      // 外层 catch 兜文件不可读，沿用原「跳过该文件」语义
-      // 计数去向：豁免（harness#100）——读的是外部 Claude Code transcript（非 harness 写入面）。
-      // 影响方向如实：坏行使该会话少几条 turn，而 analyze-sessions 的候选按次数阈值筛
-      // （frequency>=4 / 跨会话>=2 / confidence=len/10）→ 只会少出候选，不会造出候选，且候选须人审；
-      // 告知需改 MinedSession 形状并波及 analyze-sessions/update-user-model 两个调用方，另票收口
-      records = readJsonl<Record<string, any>>(filePath, 'skip').records;
-    } catch {
-      continue;
-    }
-
-    for (const entry of records) {
-      try {
-        const msg = entry.message;
-        if (msg?.role && (msg.content || entry.type === 'user')) {
-          turns.push({ role: msg.role, content: extractText(msg.content) });
-        }
-        if (msg?.role === 'assistant' && Array.isArray(msg.content)) {
-          for (const block of msg.content) {
-            if (block.type === 'tool_use' && block.name) {
-              toolCalls.push(block.name);
-            }
-          }
-        }
-      } catch {
-        // 单条记录提取异常跳过（原逐行 catch 语义：parse 已由 module skip，此处兜提取）
-      }
-    }
-
-    if (turns.length > 0) {
-      sessions.push({
-        id,
-        date: stat.mtime.toISOString().slice(0, 10),
-        mtimeMs: stat.mtimeMs,
-        turns,
-        toolCalls: [...new Set(toolCalls)],
-      });
-    }
-  }
-
-  return sessions;
-}

package/src/context/__tests__/compaction.test.ts

@@ -1,169 +0,0 @@
-/**
- * SessionCompaction 测试
- */
-
-import { SessionCompaction } from '../compaction';
-import type { SessionMessage } from '../types';
-
-function makeMessages(count: number, contentLength: number = 100): SessionMessage[] {
-  return Array.from({ length: count }, (_, i) => ({
-    role: (i % 2 === 0 ? 'user' : 'assistant') as 'user' | 'assistant',
-    content: `message ${i} `.repeat(contentLength / 10),
-    timestamp: new Date(Date.now() + i * 1000).toISOString(),
-  }));
-}
-
-describe('SessionCompaction', () => {
-  describe('shouldCompact', () => {
-    it('应该返回 true 当超过触发比例', () => {
-      const compaction = new SessionCompaction({ triggerRatio: 0.5 });
-      const messages = makeMessages(10, 500);
-      // 10 * ~50 tokens each ≈ 500 tokens, budget=800 → ratio > 0.5
-      expect(compaction.shouldCompact(messages, 800)).toBe(true);
-    });
-
-    it('应该返回 false 当未超过触发比例', () => {
-      const compaction = new SessionCompaction({ triggerRatio: 0.9 });
-      const messages = makeMessages(2, 10);
-      expect(compaction.shouldCompact(messages, 10000)).toBe(false);
-    });
-  });
-
-  describe('compact (eviction)', () => {
-    it('应该按比例丢弃最早消息', () => {
-      const compaction = new SessionCompaction({ level: 'eviction', preserveToolCallPairs: false });
-      const messages = makeMessages(20, 500);
-      const result = compaction.compact(messages, 1000);
-      expect(result.level).toBe('eviction');
-      expect(result.compacted.length).toBeLessThan(messages.length);
-      expect(result.compacted.length).toBeGreaterThanOrEqual(2);
-    });
-
-    it('应该保留最近的消息', () => {
-      const compaction = new SessionCompaction({ level: 'eviction', preserveToolCallPairs: false });
-      const messages = makeMessages(20, 100);
-      const result = compaction.compact(messages, 1000);
-      const lastOriginal = messages[messages.length - 1];
-      expect(result.compacted[result.compacted.length - 1]).toBe(lastOriginal);
-    });
-
-    it('应该保留 tool-call 边界对', () => {
-      const compaction = new SessionCompaction({ level: 'eviction', preserveToolCallPairs: true });
-      const messages: SessionMessage[] = [
-        { role: 'user', content: 'a'.repeat(1000), timestamp: 't1' },
-        { role: 'assistant', content: 'b'.repeat(1000), toolCallId: 'call-1', timestamp: 't2' },
-        { role: 'tool', content: 'c'.repeat(1000), toolCallId: 'call-1', timestamp: 't3' },
-        { role: 'user', content: 'd'.repeat(1000), timestamp: 't4' },
-        { role: 'assistant', content: 'e'.repeat(1000), timestamp: 't5' },
-      ];
-      const result = compaction.compact(messages, 500);
-      // 如果 tool result is kept, its paired assistant call should also be kept
-      const toolResult = result.compacted.find(m => m.role === 'tool');
-      if (toolResult) {
-        const paired = result.compacted.find(m => m.toolCallId === toolResult.toolCallId && m.role === 'assistant');
-        expect(paired).toBeDefined();
-      }
-    });
-  });
-
-  describe('compact (summary)', () => {
-    it('应该生成结构化摘要', () => {
-      const compaction = new SessionCompaction({ level: 'summary' });
-      const messages = makeMessages(20, 100);
-      const result = compaction.compact(messages, 1000);
-      expect(result.level).toBe('summary');
-      expect(result.summary).toBeDefined();
-      expect(result.summary).toContain('用户目标');
-      expect(result.summary).toContain('消息总数');
-    });
-
-    it('应该保留最近 6 条消息', () => {
-      const compaction = new SessionCompaction({ level: 'summary' });
-      const messages = makeMessages(20, 100);
-      const result = compaction.compact(messages, 1000);
-      // 1 summary message + 6 recent
-      expect(result.compacted.length).toBe(7);
-    });
-  });
-
-  describe('compact (checkpoint)', () => {
-    it('应该压缩为单条 checkpoint 消息', () => {
-      const compaction = new SessionCompaction({ level: 'checkpoint' });
-      const messages = makeMessages(20, 100);
-      const result = compaction.compact(messages, 1000);
-      expect(result.level).toBe('checkpoint');
-      expect(result.compacted.length).toBe(1);
-      expect(result.compacted[0].content).toContain('Checkpoint');
-    });
-  });
-
-  describe('generateStructuredSummary', () => {
-    it('应该从用户消息提取目标', () => {
-      const compaction = new SessionCompaction();
-      const messages: SessionMessage[] = [
-        { role: 'user', content: '帮我写一个函数', timestamp: 't1' },
-        { role: 'assistant', content: '好的', timestamp: 't2' },
-      ];
-      const summary = compaction.generateStructuredSummary(messages);
-      expect(summary).toContain('帮我写一个函数');
-    });
-
-    it('应该统计工具调用', () => {
-      const compaction = new SessionCompaction();
-      const messages: SessionMessage[] = [
-        { role: 'tool', content: 'result1', timestamp: 't1' },
-        { role: 'tool', content: 'result2', timestamp: 't2' },
-      ];
-      const summary = compaction.generateStructuredSummary(messages);
-      expect(summary).toContain('2 次');
-    });
-  });
-
-  describe('getConfig', () => {
-    it('应该返回配置', () => {
-      const compaction = new SessionCompaction({ level: 'summary' });
-      expect(compaction.getConfig().level).toBe('summary');
-    });
-  });
-
-  describe('compact (default level)', () => {
-    it('应该在未知 level 时回退到 eviction', () => {
-      const compaction = new SessionCompaction({ level: 'eviction' as any, preserveToolCallPairs: false });
-      // 直接设置一个不会匹配的 level 来触发 default 分支
-      (compaction as any).config.level = 'unknown_level';
-      const messages = makeMessages(20, 500);
-      const result = compaction.compact(messages, 1000);
-      expect(result.level).toBe('eviction');
-    });
-  });
-
-  describe('preserveToolCallPairs 边界情况', () => {
-    it('应该在 paired assistant 已在 compacted 中时跳过', () => {
-      const compaction = new SessionCompaction({ level: 'eviction', preserveToolCallPairs: true });
-      const messages: SessionMessage[] = [
-        { role: 'user', content: 'a'.repeat(100), timestamp: 't1' },
-        { role: 'assistant', content: 'b'.repeat(100), toolCallId: 'call-1', timestamp: 't2' },
-        { role: 'tool', content: 'c'.repeat(100), toolCallId: 'call-1', timestamp: 't3' },
-        { role: 'user', content: 'd'.repeat(100), timestamp: 't4' },
-        { role: 'assistant', content: 'e'.repeat(100), timestamp: 't5' },
-      ];
-      // budget 足够大，所有消息都会保留
-      const result = compaction.compact(messages, 100000);
-      // paired assistant 已在 compacted 中，不应重复添加
-      const assistantCalls = result.compacted.filter(m => m.toolCallId === 'call-1' && m.role === 'assistant');
-      expect(assistantCalls.length).toBe(1);
-    });
-
-    it('应该处理无 toolCallId 的消息', () => {
-      const compaction = new SessionCompaction({ level: 'eviction', preserveToolCallPairs: true });
-      const messages: SessionMessage[] = [
-        { role: 'user', content: 'a'.repeat(1000), timestamp: 't1' },
-        { role: 'assistant', content: 'b'.repeat(1000), timestamp: 't2' },
-        { role: 'user', content: 'c'.repeat(1000), timestamp: 't3' },
-        { role: 'assistant', content: 'd'.repeat(1000), timestamp: 't4' },
-      ];
-      const result = compaction.compact(messages, 500);
-      expect(result.compacted.length).toBeGreaterThan(0);
-    });
-  });
-});

package/src/context/compaction.ts

@@ -1,201 +0,0 @@
-/**
- * 会话压缩
- *
- * 三档策略：eviction / summary / checkpoint
- * 不调用 LLM，零 token 成本
- */
-
-import { TokenEstimator } from './token-budget';
-import type { CompactionConfig, CompactionLevel, SessionMessage } from './types';
-import { DEFAULT_COMPACTION_CONFIG } from './types';
-
-export interface CompactionResult {
-  compacted: SessionMessage[];
-  summary?: string;
-  level: CompactionLevel;
-  originalTokens: number;
-  compactedTokens: number;
-}
-
-export class SessionCompaction {
-  private config: CompactionConfig;
-
-  constructor(config?: Partial<CompactionConfig>) {
-    this.config = { ...DEFAULT_COMPACTION_CONFIG, ...config };
-  }
-
-  /**
-   * 检查是否需要压缩
-   */
-  shouldCompact(messages: SessionMessage[], budget: number): boolean {
-    const totalTokens = this.estimateMessages(messages);
-    return totalTokens / budget > this.config.triggerRatio;
-  }
-
-  /**
-   * 执行压缩
-   */
-  compact(messages: SessionMessage[], budget: number): CompactionResult {
-    const originalTokens = this.estimateMessages(messages);
-
-    switch (this.config.level) {
-      case 'eviction':
-        return this.evictionCompact(messages, budget, originalTokens);
-      case 'summary':
-        return this.summaryCompact(messages, budget, originalTokens);
-      case 'checkpoint':
-        return this.checkpointCompact(messages, budget, originalTokens);
-      default:
-        return this.evictionCompact(messages, budget, originalTokens);
-    }
-  }
-
-  /**
-   * 轻量压缩：按比例丢最早消息，保留 tool-call 边界对
-   */
-  private evictionCompact(messages: SessionMessage[], budget: number, originalTokens: number): CompactionResult {
-    const targetTokens = Math.floor(budget * 0.6); // 目标压缩到 60%
-    const keepRatio = targetTokens / originalTokens;
-
-    // 至少保留最近 2 条
-    const minKeep = Math.min(2, messages.length);
-    const keepCount = Math.max(minKeep, Math.floor(messages.length * keepRatio));
-
-    let compacted = messages.slice(-keepCount);
-
-    // 保留 tool-call 边界对
-    if (this.config.preserveToolCallPairs) {
-      compacted = this.preserveToolCallPairs(messages, compacted);
-    }
-
-    const compactedTokens = this.estimateMessages(compacted);
-
-    return {
-      compacted,
-      level: 'eviction',
-      originalTokens,
-      compactedTokens,
-    };
-  }
-
-  /**
-   * 中等压缩：结构化总结 + 保留最近 N 轮
-   */
-  private summaryCompact(messages: SessionMessage[], _budget: number, originalTokens: number): CompactionResult {
-    const recentCount = Math.min(6, messages.length); // 保留最近 6 条
-    const recent = messages.slice(-recentCount);
-    const older = messages.slice(0, -recentCount);
-
-    // 生成结构化摘要
-    const summary = this.generateStructuredSummary(older);
-
-    // 摘要作为系统消息插入
-    const summarySessionMessage: SessionMessage = {
-      role: 'assistant',
-      content: `[会话摘要]\n${summary}`,
-      timestamp: new Date().toISOString(),
-    };
-
-    const compacted = [summarySessionMessage, ...recent];
-    const compactedTokens = this.estimateMessages(compacted);
-
-    return {
-      compacted,
-      summary,
-      level: 'summary',
-      originalTokens,
-      compactedTokens,
-    };
-  }
-
-  /**
-   * 重型压缩：checkpoint + 清空历史
-   */
-  private checkpointCompact(messages: SessionMessage[], _budget: number, originalTokens: number): CompactionResult {
-    const summary = this.generateStructuredSummary(messages);
-
-    const checkpointSessionMessage: SessionMessage = {
-      role: 'assistant',
-      content: `[Checkpoint]\n${summary}\n\n[历史已压缩，共 ${messages.length} 条消息]`,
-      timestamp: new Date().toISOString(),
-    };
-
-    const compacted = [checkpointSessionMessage];
-    const compactedTokens = this.estimateMessages(compacted);
-
-    return {
-      compacted,
-      summary,
-      level: 'checkpoint',
-      originalTokens,
-      compactedTokens,
-    };
-  }
-
-  /**
-   * 生成结构化摘要（零 LLM 成本）
-   */
-  generateStructuredSummary(messages: SessionMessage[]): string {
-    const sections: string[] = [];
-
-    // 提取用户目标
-    const userSessionMessages = messages.filter(m => m.role === 'user');
-    if (userSessionMessages.length > 0) {
-      sections.push(`用户目标: ${userSessionMessages[0].content.slice(0, 200)}`);
-    }
-
-    // 提取工具调用
-    const toolCalls = messages.filter(m => m.role === 'tool');
-    if (toolCalls.length > 0) {
-      sections.push(`工具调用: ${toolCalls.length} 次`);
-    }
-
-    // 提取关键信息
-    const assistantSessionMessages = messages.filter(m => m.role === 'assistant');
-    if (assistantSessionMessages.length > 0) {
-      const lastAssistant = assistantSessionMessages[assistantSessionMessages.length - 1];
-      sections.push(`最近响应: ${lastAssistant.content.slice(0, 200)}`);
-    }
-
-    sections.push(`消息总数: ${messages.length}`);
-    sections.push(`时间范围: ${messages[0]?.timestamp || 'N/A'} → ${messages[messages.length - 1]?.timestamp || 'N/A'}`);
-
-    return sections.join('\n');
-  }
-
-  /**
-   * 保留 tool-call 边界对
-   *
-   * 如果 tool 结果被保留，对应的 tool 调用也要保留
-   */
-  private preserveToolCallPairs(all: SessionMessage[], compacted: SessionMessage[]): SessionMessage[] {
-    const result: SessionMessage[] = [];
-
-    for (const msg of compacted) {
-      // 如果是 tool 结果，确保对应的 tool 调用也在
-      if (msg.role === 'tool' && msg.toolCallId) {
-        const paired = all.find(m => m.toolCallId === msg.toolCallId && m.role === 'assistant');
-        if (paired && !compacted.some(m => m === paired)) {
-          result.push(paired);
-        }
-      }
-      result.push(msg);
-    }
-
-    return result;
-  }
-
-  /**
-   * 估算消息的 token 数
-   */
-  private estimateMessages(messages: SessionMessage[]): number {
-    return messages.reduce((sum, m) => sum + TokenEstimator.estimateText(m.content), 0);
-  }
-
-  /**
-   * 获取配置
-   */
-  getConfig(): CompactionConfig {
-    return { ...this.config };
-  }
-}