@reconcrap/boss-recommend-mcp 1.3.38 → 2.0.0

package/vendor/boss-chat-cli/src/services/llm.test.js

@@ -1,326 +0,0 @@
-import assert from 'node:assert/strict';
-import test from 'node:test';
-
-import { LlmClient, __testables } from './llm.js';
-
-function createJsonResponse(payload) {
-  return {
-    ok: true,
-    status: 200,
-    statusText: 'OK',
-    async json() {
-      return payload;
-    },
-    async text() {
-      return JSON.stringify(payload);
-    },
-  };
-}
-
-function createErrorResponse(status, message) {
-  return {
-    ok: false,
-    status,
-    statusText: '',
-    async json() {
-      return {};
-    },
-    async text() {
-      return message;
-    },
-  };
-}
-
-function createCompletionsPayload(content) {
-  return {
-    choices: [
-      {
-        message: {
-          content,
-        },
-      },
-    ],
-  };
-}
-
-function createClient(fetchImpl) {
-  return new LlmClient(
-    {
-      baseUrl: 'https://example.invalid/v1',
-      apiKey: 'test-key',
-      model: 'test-model',
-    },
-    {
-      fetchImpl,
-      preferCompletions: true,
-      maxRetries: 1,
-      timeoutMs: 5000,
-    },
-  );
-}
-
-function getPromptFromOptions(options = {}) {
-  const payload = JSON.parse(String(options.body || '{}'));
-  return String(
-    payload?.messages?.[0]?.content?.[0]?.text ||
-      payload?.input?.[0]?.content?.[0]?.text ||
-      '',
-  );
-}
-
-async function withChunkEnv(overrides, fn) {
-  const keys = [
-    'BOSS_CHAT_TEXT_CHUNK_SIZE_CHARS',
-    'BOSS_CHAT_TEXT_CHUNK_OVERLAP_CHARS',
-    'BOSS_CHAT_TEXT_MAX_CHUNKS',
-  ];
-  const previous = Object.fromEntries(keys.map((key) => [key, process.env[key]]));
-  for (const [key, value] of Object.entries(overrides || {})) {
-    process.env[key] = String(value);
-  }
-  try {
-    return await fn();
-  } finally {
-    for (const key of keys) {
-      if (previous[key] === undefined) {
-        delete process.env[key];
-      } else {
-        process.env[key] = previous[key];
-      }
-    }
-  }
-}
-
-function buildCandidate(resumeText) {
-  return {
-    name: '候选人A',
-    sourceJob: '算法工程师',
-    resumeProfile: {
-      primarySchool: '南京大学',
-      schools: ['南京大学'],
-      major: '人工智能',
-      majors: ['人工智能'],
-      company: '测试公司',
-      position: '算法工程师',
-    },
-    resumeText,
-    evidenceCorpus: resumeText,
-  };
-}
-
-test('evaluateTextResume keeps the full-text fast path when context allows', async () => {
-  const marker = '__END_OF_FULL_RESUME__';
-  const resumeText = `${'A'.repeat(32000)}${marker}`;
-  let callCount = 0;
-  let capturedPrompt = '';
-  const client = createClient(async (_url, options) => {
-    callCount += 1;
-    capturedPrompt = getPromptFromOptions(options);
-    return createJsonResponse(createCompletionsPayload('{"passed":false}'));
-  });
-
-  const result = await client.evaluateTextResume({
-    screeningCriteria: '有 AI 项目经验',
-    candidate: buildCandidate(resumeText),
-  });
-
-  assert.equal(result.passed, false);
-  assert.equal(result.evaluationMode, 'text');
-  assert.equal(result.aggregateRetryUsed, false);
-  assert.equal(callCount, 1);
-  assert.equal(capturedPrompt.includes(marker), true);
-});
-
-test('evaluateTextResume aggregates chunk evidence for long resumes instead of using any-pass shortcut', async () => {
-  const chunkOneMarker = '本科 2025 届，南京大学';
-  const chunkTwoMarker = '有 2 段 AI 项目与工作经历';
-  const resumeText = `${chunkOneMarker}\n${'A'.repeat(1300)}\n${chunkTwoMarker}\n${'B'.repeat(300)}`;
-  const expectedChunkCount = __testables.splitTextByChunks(resumeText, 1000, 1, 6).length;
-  let callCount = 0;
-  let aggregatePrompt = '';
-  const client = createClient(async (_url, options) => {
-    callCount += 1;
-    const prompt = getPromptFromOptions(options);
-    if (callCount === 1) {
-      return createErrorResponse(400, 'maximum context length exceeded');
-    }
-    const chunkMatch = prompt.match(/当前分段：\s*(\d+)\/(\d+)/);
-    if (chunkMatch) {
-      const chunkIndex = Number(chunkMatch[1]);
-      const chunkTotal = Number(chunkMatch[2]);
-      const isFirst = chunkIndex === 1;
-      const isLast = chunkIndex === chunkTotal;
-      return createJsonResponse(
-        createCompletionsPayload(
-          JSON.stringify({
-            chunk_passed: false,
-            chunk_summary: isFirst ? '教育信息命中' : isLast ? '项目与工作信息命中' : '中间分段主要是补充描述',
-            hard_evidence: isFirst ? ['本科 2025 届', '南京大学'] : isLast ? ['AI 项目', '2 段工作经历'] : [],
-            soft_evidence: [],
-            hard_blockers: [],
-            missing_or_uncertain: isLast ? [] : ['还需结合后续分段'],
-            quoted_spans: isFirst ? ['本科 2025 届'] : isLast ? ['2 段工作经历'] : [],
-            chunk_index: chunkIndex,
-            chunk_total: chunkTotal,
-          }),
-        ),
-      );
-    }
-    aggregatePrompt = prompt;
-    return createJsonResponse(
-      createCompletionsPayload(
-        JSON.stringify({
-          passed: true,
-          reason: '综合全部分段后，教育背景与 AI 项目/工作经历共同满足筛选条件。',
-          summary: '跨 chunk 证据成立',
-          evidence: ['本科 2025 届', 'AI 项目', '2 段工作经历'],
-        }),
-      ),
-    );
-  });
-
-  const result = await withChunkEnv(
-    {
-      BOSS_CHAT_TEXT_CHUNK_SIZE_CHARS: '1000',
-      BOSS_CHAT_TEXT_CHUNK_OVERLAP_CHARS: '1',
-      BOSS_CHAT_TEXT_MAX_CHUNKS: '6',
-    },
-    () =>
-      client.evaluateTextResume({
-        screeningCriteria: '有 AI 项目经验',
-        candidate: buildCandidate(resumeText),
-      }),
-  );
-
-  assert.equal(result.passed, true);
-  assert.equal(result.evaluationMode, 'text-chunk-aggregate');
-  assert.equal(result.chunkIndex, null);
-  assert.equal(result.chunkTotal, expectedChunkCount);
-  assert.equal(result.aggregateRetryUsed, false);
-  assert.equal(callCount, expectedChunkCount + 2);
-  assert.equal(aggregatePrompt.includes(`"chunk_count": ${expectedChunkCount}`), true);
-  assert.equal(aggregatePrompt.includes('本科 2025 届'), true);
-  assert.equal(aggregatePrompt.includes('AI 项目'), true);
-});
-
-test('evaluateTextResume does not pass when one chunk passes but aggregate decision rejects', async () => {
-  const resumeText = `${'A'.repeat(900)}\n局部命中关键词\n${'B'.repeat(900)}`;
-  const client = createClient(async (_url, options) => {
-    const prompt = getPromptFromOptions(options);
-    if (prompt.includes('简历文本:') && !prompt.includes('当前分段：')) {
-      return createErrorResponse(400, 'maximum context length exceeded');
-    }
-    const chunkMatch = prompt.match(/当前分段：\s*(\d+)\/(\d+)/);
-    if (chunkMatch) {
-      const chunkIndex = Number(chunkMatch[1]);
-      return createJsonResponse(
-        createCompletionsPayload(
-          JSON.stringify({
-            chunk_passed: chunkIndex === 1,
-            chunk_summary: chunkIndex === 1 ? '局部关键词命中' : '缺少完整证据',
-            hard_evidence: chunkIndex === 1 ? ['局部关键词'] : [],
-            soft_evidence: [],
-            hard_blockers: chunkIndex === 1 ? [] : ['缺少连续工作/项目证据'],
-            missing_or_uncertain: ['完整时间线不足'],
-            quoted_spans: chunkIndex === 1 ? ['局部关键词'] : ['缺少连续工作/项目证据'],
-            chunk_index: chunkIndex,
-            chunk_total: Number(chunkMatch[2]),
-          }),
-        ),
-      );
-    }
-    return createJsonResponse(
-      createCompletionsPayload(
-        JSON.stringify({
-          passed: false,
-          reason: '综合全部分段后仍缺少完整项目与工作链路，不能通过。',
-          summary: '聚合后不通过',
-          evidence: ['缺少连续工作/项目证据'],
-        }),
-      ),
-    );
-  });
-
-  const result = await withChunkEnv(
-    {
-      BOSS_CHAT_TEXT_CHUNK_SIZE_CHARS: '1000',
-      BOSS_CHAT_TEXT_CHUNK_OVERLAP_CHARS: '1',
-      BOSS_CHAT_TEXT_MAX_CHUNKS: '6',
-    },
-    () =>
-      client.evaluateTextResume({
-        screeningCriteria: '有 AI 项目经验',
-        candidate: buildCandidate(resumeText),
-      }),
-  );
-
-  assert.equal(result.passed, false);
-  assert.equal(result.evaluationMode, 'text-chunk-aggregate');
-});
-
-test('evaluateTextResume retries aggregate once with compacted evidence and fails on invalid aggregate JSON', async () => {
-  const resumeText = `${'第一段证据 '.repeat(160)}\n${'第二段证据 '.repeat(160)}`;
-  const aggregatePrompts = [];
-  const client = createClient(async (url, options) => {
-    const prompt = getPromptFromOptions(options);
-    if (prompt.includes('简历文本:') && !prompt.includes('当前分段：')) {
-      return createErrorResponse(400, 'maximum context length exceeded');
-    }
-    if (prompt.includes('当前分段：')) {
-      const chunkMatch = prompt.match(/当前分段：\s*(\d+)\/(\d+)/);
-      const chunkIndex = Number(chunkMatch[1]);
-      return createJsonResponse(
-        createCompletionsPayload(
-          JSON.stringify({
-            chunk_passed: false,
-            chunk_summary: `分段 ${chunkIndex} 提取到多条证据`,
-            hard_evidence: [`关键证据 ${chunkIndex}-1`, `关键证据 ${chunkIndex}-2`, `关键证据 ${chunkIndex}-3`],
-            soft_evidence: [`补充证据 ${chunkIndex}-1`, `补充证据 ${chunkIndex}-2`],
-            hard_blockers: [],
-            missing_or_uncertain: [`待确认信息 ${chunkIndex}-1`, `待确认信息 ${chunkIndex}-2`],
-            quoted_spans: [`原文片段 ${chunkIndex}-1`, `原文片段 ${chunkIndex}-2`],
-            chunk_index: chunkIndex,
-            chunk_total: Number(chunkMatch[2]),
-          }),
-        ),
-      );
-    }
-    aggregatePrompts.push(prompt);
-    if (aggregatePrompts.length === 1) {
-      return createErrorResponse(400, 'maximum context length exceeded');
-    }
-    if (String(url || '').includes('/responses')) {
-      return createJsonResponse({
-        output_text: '{"reason":"missing passed field","summary":"invalid"}',
-      });
-    }
-    return createJsonResponse(
-      createCompletionsPayload(
-        JSON.stringify({
-          reason: 'missing passed field',
-          summary: 'invalid',
-        }),
-      ),
-    );
-  });
-
-  await assert.rejects(
-    () =>
-      withChunkEnv(
-        {
-          BOSS_CHAT_TEXT_CHUNK_SIZE_CHARS: '1000',
-          BOSS_CHAT_TEXT_CHUNK_OVERLAP_CHARS: '1',
-          BOSS_CHAT_TEXT_MAX_CHUNKS: '6',
-        },
-        () =>
-          client.evaluateTextResume({
-            screeningCriteria: '有 AI 项目经验',
-            candidate: buildCandidate(resumeText),
-          }),
-      ),
-    /missing boolean "passed"|unparsable/i,
-  );
-
-  assert.equal(aggregatePrompts.length >= 2, true);
-  assert.equal(aggregatePrompts[1].length < aggregatePrompts[0].length, true);
-});

package/vendor/boss-chat-cli/src/services/profile-store.js

@@ -1,173 +0,0 @@
-import { mkdir, readFile, writeFile } from 'node:fs/promises';
-import path from 'node:path';
-
-export const DEFAULT_GREETING_TEXT = 'Hi同学，能麻烦发下简历吗？';
-
-const DEFAULT_PROFILE = {
-  screeningCriteria: '',
-  targetCount: null,
-  startFrom: 'unread',
-  greetingText: DEFAULT_GREETING_TEXT,
-  jobSelection: null,
-  llm: {
-    baseUrl: '',
-    apiKey: '',
-    model: '',
-    thinkingLevel: '',
-    timeoutMs: 60000,
-    maxRetries: 3,
-  },
-  chrome: {
-    port: 9222,
-  },
-  runtime: {
-    batchRestEnabled: true,
-    safePacing: true,
-  },
-};
-
-function cloneDefaultProfile() {
-  return JSON.parse(JSON.stringify(DEFAULT_PROFILE));
-}
-
-function mergeProfile(base, override) {
-  return {
-    ...base,
-    ...override,
-    llm: {
-      ...base.llm,
-      ...(override?.llm || {}),
-    },
-    chrome: {
-      ...base.chrome,
-      ...(override?.chrome || {}),
-    },
-    runtime: {
-      ...base.runtime,
-      ...(override?.runtime || {}),
-    },
-  };
-}
-
-function normalizeNumber(value, fallback) {
-  const parsed = Number.parseInt(String(value), 10);
-  return Number.isFinite(parsed) && parsed > 0 ? parsed : fallback;
-}
-
-function normalizeOptionalPositiveNumber(value, fallback = null) {
-  if (value === null || value === undefined || String(value).trim() === '') {
-    return fallback;
-  }
-  const parsed = Number.parseInt(String(value), 10);
-  return Number.isFinite(parsed) && parsed > 0 ? parsed : fallback;
-}
-
-function normalizeJobSelection(jobSelection) {
-  if (!jobSelection || typeof jobSelection !== 'object') {
-    return null;
-  }
-  const value = String(jobSelection.value || '').trim();
-  const label = String(jobSelection.label || '').trim();
-  if (!value && !label) {
-    return null;
-  }
-  return {
-    value: value || null,
-    label: label || null,
-  };
-}
-
-export function toPersistentProfile(profile = {}) {
-  const normalized = normalizeProfile(profile);
-  return {
-    greetingText: normalized.greetingText,
-    llm: {
-      baseUrl: normalized.llm.baseUrl,
-      apiKey: normalized.llm.apiKey,
-      model: normalized.llm.model,
-      thinkingLevel: normalized.llm.thinkingLevel,
-      timeoutMs: normalized.llm.timeoutMs,
-      maxRetries: normalized.llm.maxRetries,
-    },
-    chrome: {
-      port: normalized.chrome.port,
-    },
-    runtime: {
-      batchRestEnabled: normalized.runtime.batchRestEnabled,
-      safePacing: normalized.runtime.safePacing,
-    },
-  };
-}
-
-export function normalizeProfile(profile = {}) {
-  const merged = mergeProfile(cloneDefaultProfile(), profile);
-  merged.screeningCriteria = String(merged.screeningCriteria || '').trim();
-  merged.startFrom = String(merged.startFrom || '').trim().toLowerCase() === 'all' ? 'all' : 'unread';
-  merged.targetCount = normalizeOptionalPositiveNumber(merged.targetCount, null);
-  merged.greetingText = String(merged.greetingText || '').trim() || DEFAULT_GREETING_TEXT;
-  merged.jobSelection = normalizeJobSelection(merged.jobSelection);
-  merged.chrome.port = normalizeNumber(merged.chrome.port, DEFAULT_PROFILE.chrome.port);
-  merged.llm.baseUrl = String(merged.llm.baseUrl || '').trim().replace(/\/+$/, '');
-  merged.llm.apiKey = String(merged.llm.apiKey || '').trim();
-  merged.llm.model = String(merged.llm.model || '').trim();
-  merged.llm.thinkingLevel = String(
-    merged.llm.thinkingLevel || merged.llm.llmThinkingLevel || merged.llm.reasoningEffort || merged.llm.reasoning_effort || '',
-  ).trim();
-  merged.llm.timeoutMs = normalizeNumber(merged.llm.timeoutMs, DEFAULT_PROFILE.llm.timeoutMs);
-  merged.llm.maxRetries = normalizeNumber(merged.llm.maxRetries, DEFAULT_PROFILE.llm.maxRetries);
-  merged.runtime.batchRestEnabled = merged.runtime.batchRestEnabled !== false;
-  merged.runtime.safePacing = merged.runtime.safePacing !== false;
-  return merged;
-}
-
-export function validateProfile(profile) {
-  const missing = [];
-  if (!profile.llm.baseUrl) missing.push('llm.baseUrl');
-  if (!profile.llm.apiKey) missing.push('llm.apiKey');
-  if (!profile.llm.model) missing.push('llm.model');
-  if (!profile.chrome.port) missing.push('chrome.port');
-  return missing;
-}
-
-export class ProfileStore {
-  constructor(baseDir) {
-    this.baseDir = baseDir;
-    this.profilesDir = path.join(baseDir, 'profiles');
-  }
-
-  profilePath(profileName) {
-    return path.join(this.profilesDir, `${profileName}.json`);
-  }
-
-  async ensureDir() {
-    await mkdir(this.profilesDir, { recursive: true });
-  }
-
-  async load(profileName) {
-    await this.ensureDir();
-    try {
-      const raw = await readFile(this.profilePath(profileName), 'utf8');
-      return normalizeProfile(JSON.parse(raw));
-    } catch (error) {
-      if (error && error.code === 'ENOENT') {
-        return null;
-      }
-      throw error;
-    }
-  }
-
-  async save(profileName, profile) {
-    await this.ensureDir();
-    const normalized = toPersistentProfile(profile);
-    await writeFile(
-      this.profilePath(profileName),
-      `${JSON.stringify(normalized, null, 2)}\n`,
-      'utf8',
-    );
-    return normalizeProfile(normalized);
-  }
-
-  mergeWithOverrides(profile, overrides) {
-    return normalizeProfile(mergeProfile(profile || cloneDefaultProfile(), overrides));
-  }
-}