memind/page-data-workspace-ensure.mjs

import fs from 'node:fs';
import path from 'node:path';
import { assessPageDataHtmlBinding } from './page-data-delivery-assess.mjs';
import {
  detectPageDataDatasetUsageFromHtml,
  htmlUsesPageDataApi,
  inferPageDataBindAccessMode,
} from './page-data-html-detect.mjs';
import { bindWorkspaceHtmlForPageData } from './page-data-workspace-bind.mjs';
import { createUserDataSpaceService } from './user-data-space-service.mjs';
import { assertSafeSqlIdentifier } from './user-data-space-service.mjs';

const PAGE_DATA_ADMIN_PASSWORD = '88888888';

function readPublicHtmlFiles(workspaceRoot) {
  const publicDir = path.join(path.resolve(String(workspaceRoot ?? '')), 'public');
  if (!fs.existsSync(publicDir) || !fs.statSync(publicDir).isDirectory()) return [];
  return fs
    .readdirSync(publicDir)
    .filter((name) => name.toLowerCase().endsWith('.html'))
    .map((name) => {
      const relativePath = `public/${name}`;
      const absolutePath = path.join(publicDir, name);
      const content = fs.readFileSync(absolutePath, 'utf8');
      return { relativePath, absolutePath, content };
    });
}

export function inferInsertColumnsFromHtml(html, datasetName) {
  const text = String(html ?? '');
  const safeName = String(datasetName ?? '').replace(/[.*+?^${}()|[\]\\]/g, '\\$&');
  const patterns = [
    new RegExp(`\\.insertRow\\(\\s*['"]${safeName}['"]\\s*,\\s*\\{([\\s\\S]*?)\\}\\s*[,)]`, 'm'),
    new RegExp(
      `\\.insertRow\\(\\s*([A-Za-z_$][\\w$]*)\\s*,\\s*\\{([\\s\\S]*?)\\}\\s*[,)]`,
      'm',
    ),
  ];

  const constants = new Map();
  for (const match of text.matchAll(/(?:const|let|var)\s+([A-Za-z_$][\w$]*)\s*=\s*['"]([^'"]+)['"]/g)) {
    constants.set(match[1], match[2]);
  }

  for (const pattern of patterns) {
    const match = pattern.exec(text);
    if (!match) continue;
    const resolvedDataset = match.length === 3 ? constants.get(match[1]) ?? match[1] : datasetName;
    if (String(resolvedDataset).trim() !== String(datasetName).trim()) continue;
    const body = match[match.length - 1] ?? '';
    const columns = [];
    for (const fieldMatch of body.matchAll(/([A-Za-z_][\w$]*)\s*:/g)) {
      const column = assertSafeSqlIdentifier(fieldMatch[1], 'insert 字段');
      if (!columns.includes(column)) columns.push(column);
    }
    if (columns.length) return columns;
  }
  return [];
}

export async function ensureRegisteredDatasetFromHtml({
  workspaceRoot,
  userId = null,
  query = null,
  html,
  datasetName,
}) {
  const dataSpace = createUserDataSpaceService({ workspaceRoot, userId, query });
  const existing = await dataSpace.getDataset(datasetName);
  if (existing) return existing;

  const insertColumns = inferInsertColumnsFromHtml(html, datasetName);
  if (!insertColumns.length) {
    throw Object.assign(
      new Error(`无法从 HTML 推断 dataset「${datasetName}」的 insert 字段，请先 register_dataset`),
      { code: 'insert_columns_unknown', datasetName },
    );
  }

  const tableName = assertSafeSqlIdentifier(datasetName, 'dataset 表名');
  const columnSql = insertColumns
    .map((column) => `${column} TEXT NOT NULL DEFAULT ''`)
    .join(',\n  ');
  await dataSpace.executeSql(
    `CREATE TABLE IF NOT EXISTS ${tableName} (
  id INTEGER PRIMARY KEY AUTOINCREMENT,
  ${columnSql},
  created_at TEXT NOT NULL DEFAULT (datetime('now', '+8 hours'))
);`,
  );

  const readColumns = ['id', ...insertColumns, 'created_at'];
  return dataSpace.upsertDataset({
    name: datasetName,
    table: tableName,
    description: `Auto-registered from workspace HTML (${datasetName})`,
    actions: ['read', 'insert'],
    columns: {
      read: readColumns,
      insert: insertColumns,
    },
  });
}

export function listPageDataHtmlFiles(workspaceRoot) {
  return readPublicHtmlFiles(workspaceRoot).filter((file) => htmlUsesPageDataApi(file.content));
}

export async function ensurePageDataHtmlPagesBound({
  pool,
  h5Root,
  storageRoot,
  userId,
  workspaceRoot,
  findPageByRelativePath = null,
  onlyRelativePaths = null,
  logger = console,
} = {}) {
  if (!pool || !userId || !workspaceRoot) {
    return {
      bound: [],
      skipped: [],
      errors: [{ code: 'missing_context', message: '缺少 pool/userId/workspaceRoot' }],
    };
  }

  const allowList = onlyRelativePaths ? new Set(onlyRelativePaths) : null;
  const bound = [];
  const skipped = [];
  const errors = [];

  for (const file of listPageDataHtmlFiles(workspaceRoot)) {
    if (allowList && !allowList.has(file.relativePath)) continue;
    const usage = detectPageDataDatasetUsageFromHtml(file.content);
    if (!usage.size) {
      skipped.push({ relativePath: file.relativePath, reason: 'no_dataset_usage' });
      continue;
    }

    try {
      // Keep every page's assessment and repair isolated. A malformed legacy
      // page must be reported in `errors`, not reject the entire MindSpace page
      // listing that invokes this best-effort maintenance pass.
      const assessment = await assessPageDataHtmlBinding({
        pool,
        userId,
        publishDir: workspaceRoot,
        relativePath: file.relativePath,
        html: file.content,
        findPageByRelativePath,
      });
      if (assessment.bound) {
        skipped.push({ relativePath: file.relativePath, reason: 'already_bound' });
        continue;
      }

      for (const datasetName of usage.keys()) {
        await ensureRegisteredDatasetFromHtml({
          workspaceRoot,
          userId,
          query: pool.query.bind(pool),
          html: file.content,
          datasetName,
        });
      }
      const accessMode = inferPageDataBindAccessMode(file.relativePath, file.content);
      const result = await bindWorkspaceHtmlForPageData({
        pool,
        h5Root,
        storageRoot,
        userId,
        workspaceRoot,
        relativePath: file.relativePath,
        accessMode,
        password: accessMode === 'password' ? PAGE_DATA_ADMIN_PASSWORD : null,
      });
      bound.push({
        relativePath: file.relativePath,
        pageId: result.pageId,
        workspaceUrl: result.workspaceUrl,
      });
    } catch (err) {
      const message = err instanceof Error ? err.message : String(err);
      errors.push({
        relativePath: file.relativePath,
        message,
        code: err?.code ?? 'bind_failed',
      });
      logger.warn?.(
        `[PageData] ensure bind failed for ${file.relativePath}: ${message}`,
      );
    }
  }

  return { bound, skipped, errors };
}