import fs from 'node:fs';
|
import path from 'node:path';
|
import crypto from 'node:crypto';
|
|
const ROOT = process.cwd();
|
const CASE_ID = 'ANA-ROBOT-INDUSTRY-001';
|
const ACTION_ID = 'NEXT-ROBOT-071';
|
const RUN_ID = 'RUN-ANA-ROBOT-CONTENT-FIRST-REMAINING231-001';
|
const BATCH_ID = 'BATCH-ANA-ROBOT-CONTENT-FIRST-REMAINING231-001';
|
const CASE_REL = 'ana-data/cases/机器人案例/ANA-ROBOT-INDUSTRY-001';
|
const CASE_DIR = path.join(ROOT, ...CASE_REL.split('/'));
|
const UPSTREAM_ROOT = 'G:/industry/doc/research';
|
|
const QUEUE_REL = `${CASE_REL}/evidence/next_robot_059_company_content_completion_queue_20260801.csv`;
|
const SNAPSHOT_REL = 'ana-data/cases/机器人案例/manifest/robot_source_snapshot_CMP-ANA-ROBOT-INDUSTRY-20260723-001.csv';
|
const PRIOR_PROFILE_REL = `${CASE_REL}/evidence/next_robot_062_company_content_completion_profile_backed_evidence_20260801.csv`;
|
const MANIFEST_REL = `${CASE_REL}/manifest/artifact_manifest.csv`;
|
|
const OUTPUTS = {
|
source: `${CASE_REL}/evidence/next_robot_071_remaining231_source_authority_20260804.csv`,
|
evidence: `${CASE_REL}/evidence/next_robot_071_remaining231_content_deepening_evidence_20260804.csv`,
|
status: `${CASE_REL}/evidence/next_robot_071_remaining231_content_deepening_status_20260804.csv`,
|
progress: `${CASE_REL}/evidence/next_robot_071_company_content_deepening_progress_20260804.csv`,
|
human: `${CASE_REL}/outputs/核心文档/产业链公司内容深化_剩余231家_20260804.md`,
|
validation: `${CASE_REL}/manifest/next_robot_071_remaining231_content_deepening_validation_20260804.csv`,
|
};
|
|
const CORRECTION_RUN_ID = 'RUN-ANA-ROBOT-CONTENT-FIRST-REMAINING231-QUALITY-CORRECTION-001';
|
const CORRECTION_BATCH_ID = 'BATCH-ANA-ROBOT-CONTENT-FIRST-REMAINING231-QUALITY-CORRECTION-001';
|
const CORRECTION_OUTPUTS = {
|
evidence: `${CASE_REL}/evidence/next_robot_072_remaining231_content_deepening_quality_corrections_20260804.csv`,
|
status: `${CASE_REL}/evidence/next_robot_072_remaining231_content_deepening_quality_correction_status_20260804.csv`,
|
human: `${CASE_REL}/outputs/核心文档/产业链公司内容深化_剩余231家_完成版_20260804.md`,
|
validation: `${CASE_REL}/manifest/next_robot_072_remaining231_content_deepening_quality_correction_validation_20260804.csv`,
|
};
|
|
const CATEGORIES = [
|
{
|
id: 'CHAIN_POSITION_DETAIL',
|
section: /公司结论|一句话结论|基本信息|所属产业链|产业链环节|公司定位/i,
|
cue: /所属|环节|定位|角色|上游|中游|下游|平台|供应商|客户|本体|零部件/i,
|
interpretation: '补充公司在机器人产业链中的具体角色或相邻位置。',
|
},
|
{
|
id: 'PRODUCT_OR_CAPABILITY_DETAIL',
|
section: /机器人相关|产品|技术|能力|解决方案|核心业务/i,
|
cue: /产品|技术|能力|传感|电机|减速|编码|控制|驱动|软件|平台|机器人|模组|材料|服务/i,
|
interpretation: '补充产品、技术能力或机器人应用载体。',
|
},
|
{
|
id: 'CUSTOMER_OR_COMMERCIAL_DETAIL',
|
section: /客户|供应链|供货|合作|应用|市场|商业化|订单/i,
|
cue: /客户|合作|供货|部署|订单|应用|市场|验证|量产|试制|样品|采购|协议/i,
|
interpretation: '补充客户、合作、供货或商业化阶段信息。',
|
},
|
{
|
id: 'QUANTITATIVE_OR_OPERATION_DETAIL',
|
section: /收入|利润|营收|产量|销量|产能|ASP|寿命|交付|成本|经营|财务/i,
|
cue: /\d|收入|利润|产能|销量|产量|毛利|研发|寿命|交付|成本|单价|ASP|万元|亿元|%/i,
|
interpretation: '补充财务、产能、规格、寿命或运营量化信息。',
|
},
|
{
|
id: 'EXPLICIT_HIGH_VALUE_GAP',
|
section: /待验证|缺口|风险|硬字段|后续补充/i,
|
cue: /未披露|待|缺|需|不能|尚无|不足|不明确|后续|验证|确认/i,
|
interpretation: '明确保留客户、订单、量产、产能、ASP、寿命、收入或利润等尚未闭合的缺口。',
|
},
|
];
|
|
function abs(rel) {
|
return path.isAbsolute(rel) ? rel : path.join(ROOT, ...rel.split('/'));
|
}
|
|
function sha256Buffer(buf) {
|
return crypto.createHash('sha256').update(buf).digest('hex');
|
}
|
|
function sha256Text(text) {
|
return sha256Buffer(Buffer.from(text, 'utf8'));
|
}
|
|
function readUtf8(relOrAbs) {
|
return fs.readFileSync(abs(relOrAbs), 'utf8').replace(/^\uFEFF/, '');
|
}
|
|
function parseCsv(text) {
|
const rows = [];
|
let row = [];
|
let field = '';
|
let quoted = false;
|
for (let i = 0; i < text.length; i += 1) {
|
const c = text[i];
|
if (quoted) {
|
if (c === '"' && text[i + 1] === '"') {
|
field += '"';
|
i += 1;
|
} else if (c === '"') {
|
quoted = false;
|
} else {
|
field += c;
|
}
|
} else if (c === '"') {
|
quoted = true;
|
} else if (c === ',') {
|
row.push(field);
|
field = '';
|
} else if (c === '\n') {
|
row.push(field.replace(/\r$/, ''));
|
rows.push(row);
|
row = [];
|
field = '';
|
} else {
|
field += c;
|
}
|
}
|
if (field.length || row.length) {
|
row.push(field.replace(/\r$/, ''));
|
rows.push(row);
|
}
|
while (rows.length && rows.at(-1).every((v) => v === '')) rows.pop();
|
const headers = rows.shift();
|
return rows
|
.filter((r) => r.length === headers.length && r.join('') !== headers.join(''))
|
.map((r) => Object.fromEntries(headers.map((h, i) => [h, r[i] ?? ''])));
|
}
|
|
function csvCell(value) {
|
return `"${String(value ?? '').replaceAll('"', '""')}"`;
|
}
|
|
function toCsv(rows, columns) {
|
return `${columns.map(csvCell).join(',')}\r\n${rows.map((r) => columns.map((c) => csvCell(r[c])).join(',')).join('\r\n')}\r\n`;
|
}
|
|
function cleanMarkdownLine(raw) {
|
const text = raw.trim();
|
if (text.startsWith('|')) {
|
const cells = text.slice(1, text.endsWith('|') ? -1 : undefined).split('|').map((v) => v.trim()).filter(Boolean);
|
if (cells.length === 1) return cells[0];
|
return `${cells[0]}：${cells.slice(1).join('；')}`;
|
}
|
return text.replace(/^[-*+]\s+/, '').replace(/^>\s*/, '').replace(/\s{2,}/g, ' ').trim();
|
}
|
|
function isSeparator(line) {
|
const t = line.trim();
|
return /^\|?(?:\s*:?-{3,}:?\s*\|)+\s*$/.test(t);
|
}
|
|
function isMetadata(line) {
|
return /^>\s*(日期|版本|角色口径|资料边界|证据等级|档案状态)[:：]/.test(line.trim())
|
|| /^(日期|版本|角色口径|资料边界|证据等级|档案状态)[:：]/.test(line.trim());
|
}
|
|
function isPureGapSection(section) {
|
const normalized = String(section || '').replace(/^\d+(?:\.\d+)?\.\s*/, '');
|
return /^(关键缺口|硬字段缺口|待验证问题|待验证$|风险$|后续补充)/.test(normalized);
|
}
|
|
function candidateRows(lines, priorUsed) {
|
let section = '';
|
const out = [];
|
for (let i = 0; i < lines.length; i += 1) {
|
const raw = lines[i];
|
const lineNo = i + 1;
|
const trimmed = raw.trim();
|
const heading = trimmed.match(/^#{2,3}\s+(.+)$/);
|
if (heading) {
|
section = heading[1].trim();
|
continue;
|
}
|
if (!trimmed || trimmed.startsWith('#') || isSeparator(trimmed) || isMetadata(trimmed)) continue;
|
if (/资料来源|参考资料|证据等级/.test(section)) continue;
|
if (trimmed.startsWith('|')) {
|
let j = i + 1;
|
while (j < lines.length && !lines[j].trim()) j += 1;
|
if (j < lines.length && isSeparator(lines[j])) continue;
|
}
|
if (/^https?:\/\//i.test(trimmed) || /^[-*+]\s*`?[A-Z]:[\\/]/.test(trimmed)) continue;
|
const text = cleanMarkdownLine(raw);
|
if (text.length < 10 || priorUsed.has(lineNo)) continue;
|
if (/^(字段|项目|产品|方向|层级|资料|优先级|线索|口径|成本项|增量来源)[:：；]?/.test(text) && text.length < 35) continue;
|
out.push({ lineNo, section, raw, text, isTable: trimmed.startsWith('|') });
|
}
|
return out;
|
}
|
|
function scoreCandidate(c, category) {
|
let score = c.isTable ? 8 : 5;
|
if (category.section.test(c.section)) score += 18;
|
if (category.cue.test(c.text)) score += 9;
|
if (/\d/.test(c.text)) score += category.id === 'QUANTITATIVE_OR_OPERATION_DETAIL' ? 12 : 3;
|
if (/A\s*[/；|]|证据.*A|年报|公告|官网|公开合作/.test(c.text)) score += 3;
|
if (/未披露|尚无|不足|不明确|待验证|需补|不能/.test(c.text)) {
|
score += category.id === 'EXPLICIT_HIGH_VALUE_GAP' ? 14 : -22;
|
}
|
if (category.id === 'CHAIN_POSITION_DETAIL' && /一句话结论|公司结论/.test(c.section)) score += 5;
|
if (category.id === 'CHAIN_POSITION_DETAIL' && /风险|不确定|不等于|未披露|不能|需验证|尚未|未被|但.*未/.test(c.text)) score -= 30;
|
if (category.id === 'PRODUCT_OR_CAPABILITY_DETAIL' && /机器人相关/.test(c.section)) score += 6;
|
if (category.id === 'CUSTOMER_OR_COMMERCIAL_DETAIL' && /客户|合作|供货/.test(c.section)) score += 6;
|
if (category.id === 'CUSTOMER_OR_COMMERCIAL_DETAIL' && /^(机器人客户|客户).*未披露/.test(c.text)) score -= 30;
|
if (category.id === 'EXPLICIT_HIGH_VALUE_GAP' && /P0|优先级/.test(c.text)) score += 6;
|
if (c.text.length > 420) score -= 8;
|
return score;
|
}
|
|
function selectFive(lines, priorUsed) {
|
const candidates = candidateRows(lines, priorUsed);
|
const used = new Set();
|
const selected = [];
|
for (const category of CATEGORIES) {
|
const ranked = candidates
|
.filter((c) => !used.has(c.lineNo))
|
.filter((c) => category.id === 'EXPLICIT_HIGH_VALUE_GAP' || !isPureGapSection(c.section))
|
.map((c) => ({ ...c, score: scoreCandidate(c, category) }))
|
.sort((a, b) => b.score - a.score || a.lineNo - b.lineNo);
|
const choice = ranked[0];
|
if (!choice) throw new Error(`NO_CANDIDATE:${category.id}`);
|
used.add(choice.lineNo);
|
selected.push({ ...choice, factType: category.id, interpretation: category.interpretation });
|
}
|
if (selected.length !== 5 || new Set(selected.map((r) => r.lineNo)).size !== 5) {
|
throw new Error('EXACT_FIVE_SELECTION_FAILED');
|
}
|
return selected;
|
}
|
|
function fileIdentity(rel) {
|
const full = abs(rel);
|
const stat = fs.statSync(full);
|
const buf = fs.readFileSync(full);
|
return {
|
sha256: sha256Buffer(buf),
|
bytes: buf.length,
|
mtime: stat.mtime.toISOString(),
|
};
|
}
|
|
function appendManifestRows(paths, runId = RUN_ID, batchId = BATCH_ID) {
|
const manifestPath = abs(MANIFEST_REL);
|
const before = fs.readFileSync(manifestPath);
|
const beforeHash = sha256Buffer(before);
|
const raw = before.toString('utf8');
|
if (raw.includes(`"${runId}"`)) throw new Error('RUN_ALREADY_PRESENT_IN_MANIFEST');
|
const rows = paths.map((artifactPath) => {
|
const id = fileIdentity(artifactPath);
|
return {
|
case_id: CASE_ID,
|
run_id: runId,
|
batch_id: batchId,
|
artifact_path: artifactPath,
|
...id,
|
};
|
});
|
const appended = rows.map((r) => [r.case_id, r.run_id, r.batch_id, r.artifact_path, r.sha256, r.bytes, r.mtime].map(csvCell).join(',')).join('\r\n') + '\r\n';
|
fs.appendFileSync(manifestPath, appended, 'utf8');
|
const after = fs.readFileSync(manifestPath);
|
const actual = parseCsv(after.toString('utf8')).filter((r) => r.run_id === runId);
|
if (actual.length !== rows.length) throw new Error(`MANIFEST_RUN_COUNT:${actual.length}:${rows.length}`);
|
for (const expected of rows) {
|
const got = actual.find((r) => r.artifact_path === expected.artifact_path);
|
if (!got) throw new Error(`MANIFEST_PATH_MISSING:${expected.artifact_path}`);
|
for (const key of ['case_id', 'run_id', 'batch_id', 'sha256', 'bytes', 'mtime']) {
|
if (String(got[key]) !== String(expected[key])) throw new Error(`MANIFEST_FIELD_MISMATCH:${key}:${expected.artifact_path}`);
|
}
|
}
|
return { beforeBytes: before.length, beforeHash, afterBytes: after.length, afterHash: sha256Buffer(after), rows };
|
}
|
|
function sectionForEvidence(row) {
|
const physicalPath = row.source_root_alias === 'project_robot_existing_research'
|
? path.join(ROOT, ...`${CASE_REL}/../raw/existing_research/${row.source_relative_path}`.split('/'))
|
: path.join(UPSTREAM_ROOT, ...row.source_relative_path.split('/'));
|
const lines = fs.readFileSync(physicalPath, 'utf8').replace(/^\uFEFF/, '').replaceAll('\r\n', '\n').split('\n');
|
const lineNo = Number(String(row.content_locator).replace(/^L/, ''));
|
let section = '';
|
for (let i = 0; i < Math.min(lineNo, lines.length); i += 1) {
|
const heading = lines[i].trim().match(/^#{2,3}\s+(.+)$/);
|
if (heading) section = heading[1].trim();
|
}
|
return section;
|
}
|
|
function executeQualityCorrection() {
|
for (const rel of Object.values(CORRECTION_OUTPUTS)) {
|
if (fs.existsSync(abs(rel))) throw new Error(`OUTPUT_ALREADY_EXISTS:${rel}`);
|
}
|
const prospective = build();
|
const priorEvidence = parseCsv(readUtf8(OUTPUTS.evidence));
|
const priorById = new Map(priorEvidence.map((r) => [r.evidence_id, r]));
|
const changed = prospective.evidenceRows.filter((row) => {
|
const old = priorById.get(row.evidence_id);
|
if (!old) throw new Error(`PRIOR_EVIDENCE_ID_MISSING:${row.evidence_id}`);
|
return ['content_locator', 'fact_text', 'source_line_text_sha256'].some((key) => String(old[key]) !== String(row[key]));
|
});
|
const changedCompanies = new Set(changed.map((r) => r.company_id));
|
const correctionRows = changed.map((row, index) => {
|
const old = priorById.get(row.evidence_id);
|
return {
|
correction_id: `N072-COR-${String(index + 1).padStart(4, '0')}`,
|
supersedes_evidence_id: row.evidence_id,
|
company_id: row.company_id,
|
company_name: row.company_name,
|
fact_type: row.fact_type,
|
old_content_locator: old.content_locator,
|
old_fact_text: old.fact_text,
|
corrected_content_locator: row.content_locator,
|
corrected_fact_text: row.fact_text,
|
source_root_alias: row.source_root_alias,
|
source_relative_path: row.source_relative_path,
|
source_document_sha256: row.source_document_sha256,
|
corrected_source_line_text_sha256: row.source_line_text_sha256,
|
correction_reason: isPureGapSection(sectionForEvidence(old))
|
? 'NON_GAP_DIMENSION_PREVIOUSLY_SELECTED_FROM_PURE_GAP_SECTION'
|
: 'CASCADE_RESELECTION_AFTER_PURE_GAP_DIMENSION_CORRECTION',
|
evidence_boundary: row.evidence_boundary,
|
status: 'QUALITY_CORRECTED_EFFECTIVE_ROW_WITHOUT_STRENGTH_UPGRADE',
|
};
|
});
|
const statusRows = prospective.statusRows
|
.filter((r) => changedCompanies.has(r.company_id))
|
.map((r) => ({
|
company_id: r.company_id,
|
company_name: r.company_name,
|
corrected_row_count: correctionRows.filter((c) => c.company_id === r.company_id).length,
|
effective_content_state: 'FIVE_DIMENSION_CONTENT_DEEPENING_COMPLETED_QUALITY_CORRECTED',
|
remaining_high_value_gap: r.remaining_high_value_gap,
|
formal_pool_effect: 'NO_AUTOMATIC_FORMAL_POOL_CHANGE',
|
evidence_strength_effect: 'NO_UPGRADE',
|
}));
|
const effectivePureGapMisuse = prospective.evidenceRows.filter((r) => r.fact_type !== 'EXPLICIT_HIGH_VALUE_GAP' && isPureGapSection(sectionForEvidence(r)));
|
const priorPureGapMisuse = priorEvidence.filter((r) => r.fact_type !== 'EXPLICIT_HIGH_VALUE_GAP' && isPureGapSection(sectionForEvidence(r)));
|
if (!correctionRows.length) throw new Error('NO_CORRECTIONS_DERIVED');
|
if (effectivePureGapMisuse.length) throw new Error(`EFFECTIVE_PURE_GAP_MISUSE:${effectivePureGapMisuse.length}`);
|
if (new Set(correctionRows.map((r) => r.supersedes_evidence_id)).size !== correctionRows.length) throw new Error('CORRECTION_SUPERSESSION_NOT_UNIQUE');
|
|
const correctionColumns = ['correction_id', 'supersedes_evidence_id', 'company_id', 'company_name', 'fact_type', 'old_content_locator', 'old_fact_text', 'corrected_content_locator', 'corrected_fact_text', 'source_root_alias', 'source_relative_path', 'source_document_sha256', 'corrected_source_line_text_sha256', 'correction_reason', 'evidence_boundary', 'status'];
|
const statusColumns = ['company_id', 'company_name', 'corrected_row_count', 'effective_content_state', 'remaining_high_value_gap', 'formal_pool_effect', 'evidence_strength_effect'];
|
const validationRows = [
|
['N072VAL-001', 'prior_evidence_rows_preserved', '1155', String(priorEvidence.length), 'PASS', 'N071 evidence remains append-only history'],
|
['N072VAL-002', 'prior_non_gap_rows_from_pure_gap_sections', String(priorPureGapMisuse.length), String(priorPureGapMisuse.length), 'INFO', 'identified quality issue, not deleted'],
|
['N072VAL-003', 'correction_row_count', String(changed.length), String(correctionRows.length), 'PASS', 'all changed effective rows recorded'],
|
['N072VAL-004', 'correction_company_count', String(changedCompanies.size), String(statusRows.length), 'PASS', 'one status per affected company'],
|
['N072VAL-005', 'unique_supersedes_evidence_id', String(correctionRows.length), String(new Set(correctionRows.map((r) => r.supersedes_evidence_id)).size), 'PASS', 'one correction per old evidence row'],
|
['N072VAL-006', 'effective_non_gap_rows_from_pure_gap_sections', '0', String(effectivePureGapMisuse.length), 'PASS', 'dimension-label integrity restored'],
|
['N072VAL-007', 'effective_evidence_rows', '1155', String(prospective.evidenceRows.length), 'PASS', 'five effective rows per company'],
|
['N072VAL-008', 'effective_company_count', '231', String(new Set(prospective.evidenceRows.map((r) => r.company_id)).size), 'PASS', 'all remaining companies'],
|
['N072VAL-009', 'effective_explicit_gap_rows', '231', String(prospective.evidenceRows.filter((r) => r.fact_type === 'EXPLICIT_HIGH_VALUE_GAP').length), 'PASS', 'gap retained separately'],
|
['N072VAL-010', 'formal_pool_effect', 'NO_AUTOMATIC_FORMAL_POOL_CHANGE', 'NO_AUTOMATIC_FORMAL_POOL_CHANGE', 'PASS', 'no pool promotion'],
|
['N072VAL-011', 'evidence_strength_effect', 'NO_UPGRADE', 'NO_UPGRADE', 'PASS', 'no strength upgrade'],
|
['N072VAL-012', 'remaining_company_count', '0', '0', 'PASS', '307/307 deepening complete after correction overlay'],
|
].map(([check_id, check_name, expected, actual, status, note]) => ({ check_id, check_name, expected, actual, status, note }));
|
|
const human = [...prospective.human];
|
human[0] = '# 机器人产业链剩余231家公司内容深化完成版';
|
human.splice(6, 0,
|
'> 质量修正：原N071中误从纯“待验证/硬字段缺口”章节抽取到非缺口维度的记录已通过N072逐行替换；原行仅作为历史保留。',
|
'');
|
|
fs.writeFileSync(abs(CORRECTION_OUTPUTS.evidence), toCsv(correctionRows, correctionColumns), 'utf8');
|
fs.writeFileSync(abs(CORRECTION_OUTPUTS.status), toCsv(statusRows, statusColumns), 'utf8');
|
fs.writeFileSync(abs(CORRECTION_OUTPUTS.human), `${human.join('\n')}\n`, 'utf8');
|
fs.writeFileSync(abs(CORRECTION_OUTPUTS.validation), toCsv(validationRows, ['check_id', 'check_name', 'expected', 'actual', 'status', 'note']), 'utf8');
|
const manifest = appendManifestRows(Object.values(CORRECTION_OUTPUTS), CORRECTION_RUN_ID, CORRECTION_BATCH_ID);
|
process.stdout.write(JSON.stringify({
|
status: 'PASS', action_id: 'NEXT-ROBOT-072', prior_pure_gap_misuse: priorPureGapMisuse.length,
|
correction_rows: correctionRows.length, affected_companies: statusRows.length,
|
effective_companies: 231, effective_evidence_rows: 1155, remaining: 0,
|
manifest_prefix_bytes: manifest.beforeBytes, manifest_prefix_sha256: manifest.beforeHash,
|
manifest_current_bytes: manifest.afterBytes, manifest_current_sha256: manifest.afterHash,
|
manifest_run_rows: manifest.rows.length,
|
}, null, 2));
|
}
|
|
function build() {
|
const queue = parseCsv(readUtf8(QUEUE_REL)).filter((r) => Number(r.queue_rank) >= 33).sort((a, b) => Number(a.queue_rank) - Number(b.queue_rank));
|
const snapshots = parseCsv(readUtf8(SNAPSHOT_REL));
|
const prior = parseCsv(readUtf8(PRIOR_PROFILE_REL));
|
if (queue.length !== 231) throw new Error(`QUEUE_COUNT:${queue.length}`);
|
|
const priorLinesByCompany = new Map();
|
for (const row of prior) {
|
const lines = [...String(row.content_locator || '').matchAll(/L(\d+)/g)].map((m) => Number(m[1]));
|
if (!priorLinesByCompany.has(row.company_id)) priorLinesByCompany.set(row.company_id, new Set());
|
for (const lineNo of lines) priorLinesByCompany.get(row.company_id).add(lineNo);
|
}
|
|
const sourceRows = [];
|
const evidenceRows = [];
|
const statusRows = [];
|
const humanCompanies = [];
|
let evidenceSeq = 1;
|
|
for (const q of queue) {
|
const sourceRelativePath = `公司档案/${path.basename(q.source_profile_path)}`.replaceAll('\\', '/');
|
const sourceAlias = q.source_profile_exists === 'YES' ? 'project_robot_existing_research' : 'industry_research_current';
|
const physicalPath = q.source_profile_exists === 'YES'
|
? abs(q.source_profile_path)
|
: path.join(UPSTREAM_ROOT, ...sourceRelativePath.split('/'));
|
if (!fs.existsSync(physicalPath)) throw new Error(`SOURCE_MISSING:${q.queue_rank}:${q.company_name}`);
|
const snapshot = snapshots.find((s) => s.source_root_alias === sourceAlias && s.relative_path === sourceRelativePath);
|
if (!snapshot) throw new Error(`SNAPSHOT_MISSING:${q.queue_rank}:${sourceAlias}:${sourceRelativePath}`);
|
const sourceBuffer = fs.readFileSync(physicalPath);
|
const sourceSha = sha256Buffer(sourceBuffer);
|
if (![snapshot.raw_sha256, snapshot.normalized_sha256].includes(sourceSha)) {
|
throw new Error(`SOURCE_SNAPSHOT_HASH_MISMATCH:${q.queue_rank}:${q.company_name}`);
|
}
|
const text = sourceBuffer.toString('utf8').replace(/^\uFEFF/, '').replaceAll('\r\n', '\n');
|
const lines = text.split('\n');
|
if (lines.at(-1) === '') lines.pop();
|
const priorUsed = q.source_profile_exists === 'YES' ? (priorLinesByCompany.get(q.company_id) || new Set()) : new Set();
|
let selected;
|
try {
|
selected = selectFive(lines, priorUsed);
|
} catch (error) {
|
throw new Error(`${error.message}:${q.queue_rank}:${q.company_name}`);
|
}
|
const sourceDateMatch = text.match(/日期[:：]\s*(\d{4}-\d{2}-\d{2})/);
|
const sourceDate = sourceDateMatch?.[1] || String(snapshot.captured_at || '').slice(0, 10);
|
const sourceUrl = `source-root://${sourceAlias}/${sourceRelativePath}`;
|
|
sourceRows.push({
|
source_authority_id: `N071-SRC-${String(sourceRows.length + 1).padStart(4, '0')}`,
|
queue_rank: q.queue_rank,
|
company_id: q.company_id,
|
company_name: q.company_name,
|
source_root_alias: sourceAlias,
|
source_relative_path: sourceRelativePath,
|
snapshot_row_id: snapshot.source_snapshot_row_id,
|
snapshot_id: snapshot.snapshot_id,
|
source_sha256: sourceSha,
|
snapshot_raw_sha256: snapshot.raw_sha256,
|
source_bytes: sourceBuffer.length,
|
source_lines: lines.length,
|
captured_at: snapshot.captured_at,
|
physical_identity_status: 'MATCHED_FROZEN_SOURCE_SNAPSHOT',
|
});
|
|
for (const fact of selected) {
|
evidenceRows.push({
|
evidence_id: `N071-EV-${String(evidenceSeq).padStart(4, '0')}`,
|
company_id: q.company_id,
|
company_name: q.company_name,
|
source_title: `${q.company_name}公司信息梳理`,
|
source_url: sourceUrl,
|
source_date: sourceDate,
|
source_type: q.source_profile_exists === 'YES' ? 'FROZEN_PROJECT_PROFILE_DETAIL_EXTRACTION' : 'FROZEN_UPSTREAM_PROFILE_RECOVERY_AND_DETAIL_EXTRACTION',
|
content_locator: `L${fact.lineNo}`,
|
fact_type: fact.factType,
|
fact_text: fact.text,
|
value: '',
|
unit: '',
|
period: 'frozen_profile_snapshot',
|
robot_chain_interpretation: fact.interpretation,
|
evidence_boundary: 'PENDING_VERIFICATION_MAX_NO_UPGRADE|NO_AUTOMATIC_FORMAL_POOL_CHANGE',
|
status: 'CONTENT_DEEPENING_COMPLETED_WITH_EXPLICIT_GAP_RETAINED',
|
source_root_alias: sourceAlias,
|
source_relative_path: sourceRelativePath,
|
source_document_sha256: sourceSha,
|
source_line_text_sha256: sha256Text(fact.raw),
|
});
|
evidenceSeq += 1;
|
}
|
|
const gap = selected.find((r) => r.factType === 'EXPLICIT_HIGH_VALUE_GAP');
|
statusRows.push({
|
queue_rank: q.queue_rank,
|
company_id: q.company_id,
|
company_name: q.company_name,
|
source_mode: q.source_profile_exists === 'YES' ? 'PROJECT_PROFILE' : 'UPSTREAM_PROFILE_RECOVERED',
|
deepening_fact_count: 5,
|
content_state: 'FIVE_DIMENSION_CONTENT_DEEPENING_COMPLETED_WITH_OPEN_GAPS',
|
confirmed_increment: '产业链定位、产品/能力、客户/商业化、量化/运营和高价值缺口均形成逐行可复核记录',
|
remaining_high_value_gap: gap.text,
|
next_valid_trigger: '官方公告、年报、产品手册、客户侧披露、采购/订单或量产运营资料',
|
formal_pool_effect: 'NO_AUTOMATIC_FORMAL_POOL_CHANGE',
|
evidence_strength_effect: 'NO_UPGRADE',
|
});
|
humanCompanies.push({ q, sourceAlias, sourceRelativePath, selected });
|
}
|
|
const sourceColumns = ['source_authority_id', 'queue_rank', 'company_id', 'company_name', 'source_root_alias', 'source_relative_path', 'snapshot_row_id', 'snapshot_id', 'source_sha256', 'snapshot_raw_sha256', 'source_bytes', 'source_lines', 'captured_at', 'physical_identity_status'];
|
const evidenceColumns = ['evidence_id', 'company_id', 'company_name', 'source_title', 'source_url', 'source_date', 'source_type', 'content_locator', 'fact_type', 'fact_text', 'value', 'unit', 'period', 'robot_chain_interpretation', 'evidence_boundary', 'status', 'source_root_alias', 'source_relative_path', 'source_document_sha256', 'source_line_text_sha256'];
|
const statusColumns = ['queue_rank', 'company_id', 'company_name', 'source_mode', 'deepening_fact_count', 'content_state', 'confirmed_increment', 'remaining_high_value_gap', 'next_valid_trigger', 'formal_pool_effect', 'evidence_strength_effect'];
|
const progressRows = [
|
{ metric: 'company_universe', before_bulk: '307', after_bulk: '307', remaining: '0', note: '机器人公司全集保持307家不变' },
|
{ metric: 'curated_or_deepened_company_count', before_bulk: '76', after_bulk: '307', remaining: '0', note: '剩余231家全部完成五维内容深化' },
|
{ metric: 'project_profile_pending_deepening', before_bulk: '114', after_bulk: '0', remaining: '0', note: '114家项目内档案逐行提取未被旧基线使用的新增细节' },
|
{ metric: 'upstream_profile_pending_recovery', before_bulk: '117', after_bulk: '0', remaining: '0', note: '117家项目内缺档企业从冻结上游档案恢复并深化' },
|
{ metric: 'new_deepening_evidence_rows', before_bulk: '0', after_bulk: String(evidenceRows.length), remaining: '0', note: '每家公司5条：定位、产品、商业化、量化运营、明确缺口' },
|
{ metric: 'content_completion_definition', before_bulk: 'BASELINE_ONLY', after_bulk: 'DEEPENED_WITH_EXPLICIT_GAPS', remaining: '0', note: '完成指内容深化覆盖完成，不等于缺口关闭、正式入池或证据升级' },
|
];
|
|
const human = [
|
'# 机器人产业链剩余231家公司内容深化',
|
'',
|
'> 日期：2026-08-04 ',
|
'> 批次：NEXT-ROBOT-071 / REMAINING231 ',
|
'> 方法：内容优先，一次性复用114家项目档案和117家冻结上游档案；每家公司抽取5条逐行可复核内容，避免重复旧基线。',
|
'',
|
'## 1. 完成结果',
|
'',
|
'- 本批覆盖：231/231 家；累计深化：307/307 家；剩余：0 家。',
|
'- 新增逐行证据：1,155 条；每家公司固定覆盖产业链定位、产品/能力、客户/商业化、量化/运营、高价值缺口。',
|
'- 114家来自项目内公司档案；117家从冻结上游 `industry_research_current` 档案恢复，物理哈希均与既有 source snapshot 一致。',
|
'- “完成”仅表示公司内容深化和缺口显式化完成；不表示客户、订单、量产、产能、ASP、寿命、收入或利润已全部核实。',
|
'',
|
'## 2. 公司明细',
|
'',
|
];
|
for (const company of humanCompanies) {
|
human.push(`### ${company.q.queue_rank}. ${company.q.company_name}`);
|
human.push('');
|
human.push(`- 来源：\`${company.sourceAlias}/${company.sourceRelativePath}\``);
|
for (const category of CATEGORIES) {
|
const fact = company.selected.find((r) => r.factType === category.id);
|
human.push(`- ${category.id}（L${fact.lineNo}）：${fact.text}`);
|
}
|
human.push('- 边界：PENDING_VERIFICATION_MAX_NO_UPGRADE；NO_AUTOMATIC_FORMAL_POOL_CHANGE。');
|
human.push('');
|
}
|
|
const errors = [];
|
if (sourceRows.length !== 231) errors.push(`SOURCE_COUNT:${sourceRows.length}`);
|
if (new Set(sourceRows.map((r) => r.company_id)).size !== 231) errors.push('SOURCE_COMPANY_UNIQUE');
|
if (evidenceRows.length !== 1155) errors.push(`EVIDENCE_COUNT:${evidenceRows.length}`);
|
if (new Set(evidenceRows.map((r) => r.evidence_id)).size !== 1155) errors.push('EVIDENCE_ID_UNIQUE');
|
for (const q of queue) {
|
const rows = evidenceRows.filter((r) => r.company_id === q.company_id);
|
if (rows.length !== 5) errors.push(`COMPANY_FACT_COUNT:${q.queue_rank}:${rows.length}`);
|
const types = new Set(rows.map((r) => r.fact_type));
|
for (const category of CATEGORIES) if (!types.has(category.id)) errors.push(`COMPANY_CATEGORY_MISSING:${q.queue_rank}:${category.id}`);
|
}
|
if (statusRows.length !== 231) errors.push(`STATUS_COUNT:${statusRows.length}`);
|
if (sourceRows.filter((r) => r.source_root_alias === 'project_robot_existing_research').length !== 114) errors.push('PROJECT_SOURCE_COUNT');
|
if (sourceRows.filter((r) => r.source_root_alias === 'industry_research_current').length !== 117) errors.push('UPSTREAM_SOURCE_COUNT');
|
const projectPriorCollision = evidenceRows.filter((r) => {
|
if (r.source_root_alias !== 'project_robot_existing_research') return false;
|
const lineNo = Number(r.content_locator.slice(1));
|
return (priorLinesByCompany.get(r.company_id) || new Set()).has(lineNo);
|
});
|
if (projectPriorCollision.length) errors.push(`PRIOR_LOCATOR_COLLISION:${projectPriorCollision.length}`);
|
if (evidenceRows.some((r) => !r.fact_text || /^[-|\s]+$/.test(r.fact_text))) errors.push('STRUCTURAL_OR_EMPTY_FACT');
|
if (errors.length) throw new Error(errors.join('|'));
|
|
const validationRows = [
|
['N071VAL-001', 'remaining_company_universe', '231', String(queue.length), 'PASS', 'completion queue ranks 33..263'],
|
['N071VAL-002', 'project_profile_company_count', '114', String(sourceRows.filter((r) => r.source_root_alias === 'project_robot_existing_research').length), 'PASS', 'project-local profiles'],
|
['N071VAL-003', 'upstream_profile_recovered_count', '117', String(sourceRows.filter((r) => r.source_root_alias === 'industry_research_current').length), 'PASS', 'frozen upstream profiles'],
|
['N071VAL-004', 'source_snapshot_identity_match', '231', String(sourceRows.filter((r) => r.physical_identity_status === 'MATCHED_FROZEN_SOURCE_SNAPSHOT').length), 'PASS', 'physical SHA matched raw or normalized snapshot authority'],
|
['N071VAL-005', 'evidence_row_count', '1155', String(evidenceRows.length), 'PASS', 'exactly five rows per company'],
|
['N071VAL-006', 'unique_evidence_id', '1155', String(new Set(evidenceRows.map((r) => r.evidence_id)).size), 'PASS', 'unique evidence IDs'],
|
['N071VAL-007', 'five_rows_per_company', '231', String(queue.filter((q) => evidenceRows.filter((r) => r.company_id === q.company_id).length === 5).length), 'PASS', 'all companies'],
|
['N071VAL-008', 'five_categories_per_company', '231', String(queue.filter((q) => new Set(evidenceRows.filter((r) => r.company_id === q.company_id).map((r) => r.fact_type)).size === 5).length), 'PASS', 'position/product/commercial/quant/gap'],
|
['N071VAL-009', 'prior_profile_locator_collision', '0', String(projectPriorCollision.length), 'PASS', 'does not reuse N062 profile locator rows'],
|
['N071VAL-010', 'explicit_gap_rows', '231', String(evidenceRows.filter((r) => r.fact_type === 'EXPLICIT_HIGH_VALUE_GAP').length), 'PASS', 'one explicit gap per company'],
|
['N071VAL-011', 'status_row_count', '231', String(statusRows.length), 'PASS', 'one status row per company'],
|
['N071VAL-012', 'formal_pool_effect', 'NO_AUTOMATIC_FORMAL_POOL_CHANGE:231', `NO_AUTOMATIC_FORMAL_POOL_CHANGE:${statusRows.filter((r) => r.formal_pool_effect === 'NO_AUTOMATIC_FORMAL_POOL_CHANGE').length}`, 'PASS', 'no pool promotion'],
|
['N071VAL-013', 'evidence_strength_effect', 'NO_UPGRADE:231', `NO_UPGRADE:${statusRows.filter((r) => r.evidence_strength_effect === 'NO_UPGRADE').length}`, 'PASS', 'no strength upgrade'],
|
['N071VAL-014', 'remaining_company_count', '0', '0', 'PASS', '307/307 content deepening pass complete'],
|
['N071VAL-015', 'completion_boundary', 'DEEPENED_WITH_EXPLICIT_GAPS', 'DEEPENED_WITH_EXPLICIT_GAPS', 'PASS', 'coverage completion is not factual gap closure'],
|
].map(([check_id, check_name, expected, actual, status, note]) => ({ check_id, check_name, expected, actual, status, note }));
|
|
return { sourceRows, sourceColumns, evidenceRows, evidenceColumns, statusRows, statusColumns, progressRows, human, validationRows };
|
}
|
|
function execute() {
|
for (const rel of Object.values(OUTPUTS)) {
|
if (fs.existsSync(abs(rel))) throw new Error(`OUTPUT_ALREADY_EXISTS:${rel}`);
|
}
|
const built = build();
|
fs.writeFileSync(abs(OUTPUTS.source), toCsv(built.sourceRows, built.sourceColumns), 'utf8');
|
fs.writeFileSync(abs(OUTPUTS.evidence), toCsv(built.evidenceRows, built.evidenceColumns), 'utf8');
|
fs.writeFileSync(abs(OUTPUTS.status), toCsv(built.statusRows, built.statusColumns), 'utf8');
|
fs.writeFileSync(abs(OUTPUTS.progress), toCsv(built.progressRows, ['metric', 'before_bulk', 'after_bulk', 'remaining', 'note']), 'utf8');
|
fs.writeFileSync(abs(OUTPUTS.human), `${built.human.join('\n')}\n`, 'utf8');
|
fs.writeFileSync(abs(OUTPUTS.validation), toCsv(built.validationRows, ['check_id', 'check_name', 'expected', 'actual', 'status', 'note']), 'utf8');
|
const manifest = appendManifestRows(Object.values(OUTPUTS));
|
process.stdout.write(JSON.stringify({
|
status: 'PASS', action_id: ACTION_ID, companies: 231, evidence_rows: 1155,
|
project_profiles: 114, upstream_profiles: 117, remaining: 0,
|
manifest_prefix_bytes: manifest.beforeBytes, manifest_prefix_sha256: manifest.beforeHash,
|
manifest_current_bytes: manifest.afterBytes, manifest_current_sha256: manifest.afterHash,
|
manifest_run_rows: manifest.rows.length,
|
}, null, 2));
|
}
|
|
function preview() {
|
const built = build();
|
const categoryCounts = Object.fromEntries(CATEGORIES.map((c) => [c.id, built.evidenceRows.filter((r) => r.fact_type === c.id).length]));
|
const qualityFlags = {
|
chain_position_gap_like: built.evidenceRows.filter((r) => r.fact_type === 'CHAIN_POSITION_DETAIL' && /风险|不确定|不等于|未披露|不能|需验证|尚未|未被|但.*未/.test(r.fact_text)).length,
|
commercial_missing_like: built.evidenceRows.filter((r) => r.fact_type === 'CUSTOMER_OR_COMMERCIAL_DETAIL' && /^(机器人客户|客户).*未披露/.test(r.fact_text)).length,
|
quantitative_without_number: built.evidenceRows.filter((r) => r.fact_type === 'QUANTITATIVE_OR_OPERATION_DETAIL' && !/\d/.test(r.fact_text)).length,
|
overly_long_fact: built.evidenceRows.filter((r) => r.fact_text.length > 500).length,
|
};
|
process.stdout.write(JSON.stringify({
|
status: 'PREVIEW_PASS', companies: built.sourceRows.length, evidence_rows: built.evidenceRows.length,
|
project_profiles: built.sourceRows.filter((r) => r.source_root_alias === 'project_robot_existing_research').length,
|
upstream_profiles: built.sourceRows.filter((r) => r.source_root_alias === 'industry_research_current').length,
|
category_counts: categoryCounts,
|
quality_flags: qualityFlags,
|
samples: built.evidenceRows.slice(0, 10).map((r) => ({ company: r.company_name, type: r.fact_type, locator: r.content_locator, fact: r.fact_text })),
|
}, null, 2));
|
}
|
|
const mode = process.argv[2] || '--preview';
|
if (mode === '--preview') preview();
|
else if (mode === '--execute') execute();
|
else if (mode === '--execute-quality-correction') executeQualityCorrection();
|
else throw new Error(`UNKNOWN_MODE:${mode}`);
|
