const fs = require('fs')
const path = require('path')

const root = path.resolve(__dirname, '..')
const reportDir = path.join(root, 'reports')
const excluded = new Set(['.git', 'node_modules', 'unpackage', '.zcode'])

function walk(dir) {
  if (!fs.existsSync(dir)) return []
  const result = []
  for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
    if (entry.isDirectory() && excluded.has(entry.name)) continue
    const file = path.join(dir, entry.name)
    if (entry.isDirectory()) result.push(...walk(file))
    else result.push(file)
  }
  return result
}

function rel(file) {
  return path.relative(root, file).replace(/\\/g, '/')
}

function parseJson(file) {
  try {
    return JSON.parse(stripJsonComments(fs.readFileSync(file, 'utf8')))
  } catch (error) {
    return { __error: error.message }
  }
}

// B1: schema 允许 // 行注释与 /* 块注释头（uniCloud JSONC），比较前剥离
function stripJsonComments(text) {
  let out = ''
  let i = 0
  const n = text.length
  let inStr = false
  while (i < n) {
    const c = text[i]
    if (inStr) {
      out += c
      if (c === '\\') { out += text[i + 1] || ''; i += 2; continue }
      if (c === '"') inStr = false
      i++
      continue
    }
    if (c === '"') { inStr = true; out += c; i++; continue }
    if (c === '/' && text[i + 1] === '/') {
      const end = text.indexOf('\n', i)
      const stop = end === -1 ? n : end
      out += ' '.repeat(stop - i)
      i = stop
      continue
    }
    if (c === '/' && text[i + 1] === '*') {
      const end = text.indexOf('*/', i + 2)
      const stop = end === -1 ? n : end + 2
      const skipped = text.slice(i, stop)
      out += skipped.replace(/[^\n]/g, ' ')
      i = stop
      continue
    }
    out += c
    i++
  }
  return out
}

function schemaFiles() {
  const localSchemas = walk(path.join(root, 'uniCloud-alipay', 'database'))
    .filter(file => file.endsWith('.schema.json'))
  const moduleSchemas = walk(path.join(root, 'uni_modules'))
    .filter(file => file.endsWith('.schema.json'))
  return [...localSchemas, ...moduleSchemas]
}

function sourceFiles() {
  return [
    ...walk(path.join(root, 'pages')),
    ...walk(path.join(root, 'components')),
    ...walk(path.join(root, 'cloudfunctions'))
  ].filter(file => /\.(vue|js)$/.test(file))
}

function getSchemas() {
  const result = new Map()
  for (const file of schemaFiles()) {
    result.set(path.basename(file, '.schema.json'), { file: rel(file), schema: parseJson(file) })
  }
  return result
}

function addMap(map, key, value) {
  if (!map.has(key)) map.set(key, [])
  map.get(key).push(value)
}

function collectCollections(sources) {
  const collections = new Map()
  const patterns = [
    /collectionList\s*:\s*["']([^"']+)["']/g,
    /collection\s*:\s*["']([^"']+)["']/g,
    /collection\(["']([^"']+)["']\)/g
  ]
  for (const file of sources) {
    const text = fs.readFileSync(file, 'utf8')
    for (const pattern of patterns) {
      let match
      while ((match = pattern.exec(text))) addMap(collections, match[1], rel(file))
    }
  }
  return collections
}

function collectFields(sources, collections) {
  const fields = new Map()
  // B1-1: 前置边界 [\s"] 排除 group-field/groupField；只认属性 field=（不认 .field( JS 调用）
  const fieldPattern = /[\s"]field\s*=\s*["']([^"']+)["']/g
  // 主集合判定：<unicloud-db ...> 标签内的 collection 绑定
  const udbRe = /<unicloud-db\b[^>]*>/g
  const udbCollectionRe = /collectionList\s*:\s*["']([^"']+)["']|collection\s*:\s*["']([^"']+)["']/
  for (const file of sources.filter(file => file.endsWith('.vue'))) {
    const text = fs.readFileSync(file, 'utf8')
    // script 区 collectionList: "x" 字面量（:collection="collectionList" 绑定的解析目标）
    const scriptCollections = []
    const scriptRe = /collectionList\s*:\s*["']([^"']+)["']/g
    let sm
    while ((sm = scriptRe.exec(text))) scriptCollections.push(sm[1])
    const arrayRe = /collectionList\s*:\s*\[([^\]]*)\]/g
    let am
    while ((am = arrayRe.exec(text))) {
      const inner = am[1]
      const strRe = /db\.collection\(\s*["']([^"']+)["']\s*\)/g
      let im
      while ((im = strRe.exec(inner))) scriptCollections.push(im[1])
    }
    let um
    let foundMain = false
    const pendingTags = []
    while ((um = udbRe.exec(text))) {
      const tag = um[0]
      const cm = udbCollectionRe.exec(tag)
      if (!cm) {
        if (/[:\s]collection\s*=/.test(tag)) pendingTags.push(tag)
        continue
      }
      const main = cm[1] || cm[2]
      foundMain = true
      collectTagFields(tag, main, file, fields)
    }
    if (pendingTags.length && scriptCollections.length) {
      // :collection="collectionList" 绑定的 unicloud-db 标签：只收该标签自身的 field，
      // 不收 uni-data-select / uni-data-checkbox 等辅助组件的 field
      foundMain = true
      const main = scriptCollections[0]
      for (const tag of pendingTags) collectTagFields(tag, main, file, fields)
    }
  }
  return fields
}

function collectTagFields(tag, main, file, fields) {
  const fieldRe = /[\s"]field\s*=\s*["']([^"']+)["']/g
  let fm
  while ((fm = fieldRe.exec(tag))) addFieldEntries(main, fm[1], file, fields)
}

function addFieldEntries(main, raw, file, fields) {
  for (const field of raw.split(',')) {
    // B1-1: 去 as 别名，去表前缀；聚合表达式（含括号）直接丢弃
    const name = field.trim().split(/\s+as\s+/i)[0].split('.')[0].trim()
    if (!name || name.includes('(') || name.includes(')')) continue
    addMap(fields, `${main}:${name}`, rel(file))
  }
}

function buildReport() {
  const schemas = getSchemas()
  const sources = sourceFiles()
  const collections = collectCollections(sources)
  const fields = collectFields(sources, collections)
  const missingSchemas = []
  for (const [collection, files] of collections) {
    if (!schemas.has(collection)) missingSchemas.push({ collection, files: [...new Set(files)] })
  }
  const fieldDrift = []
  for (const [key, files] of fields) {
    const [collection, field] = key.split(':')
    const schema = schemas.get(collection)
    if (!schema || schema.schema.__error) {
      fieldDrift.push({ collection, field, files: [...new Set(files)], schema: schema ? schema.file : null, driftKind: 'missing-in-schema' })
    } else if (!schema.schema.properties || !Object.prototype.hasOwnProperty.call(schema.schema.properties, field)) {
      fieldDrift.push({ collection, field, files: [...new Set(files)], schema: schema.file, driftKind: 'page-field-typo' })
    }
  }
  const knownMigrations = [
    {
      source: 'opendb-news-articles',
      target: 'uni-cms-articles',
      status: 'requires-review',
      mapping: { avatar: 'thumbnail[]', content: 'content Delta/object', category_id: 'uni-cms-categories._id' }
    },
    {
      source: 'opendb-news-comments',
      target: 'comment',
      status: 'missing-target-schema',
      mapping: { comment_content: 'content', comment_date: 'create_time', comment_ip: 'ip_location' }
    },
    {
      source: 'opendb-news-favorite',
      target: 'comment_like',
      status: 'requires-review',
      mapping: { create_date: 'create_time', article_id: 'comment_id (semantic mismatch)' }
    }
  ]
  const report = {
    generatedAt: new Date().toISOString(),
    schemas: [...schemas.entries()].map(([name, value]) => ({ name, file: value.file })),
    collections: [...collections.entries()].map(([name, files]) => ({ name, files: [...new Set(files)] })),
    missingSchemas,
    fieldDrift,
    knownMigrations,
    summary: {
      schemaCount: schemas.size,
      collectionCount: collections.size,
      missingSchemaCount: missingSchemas.length,
      fieldDriftCount: fieldDrift.length
    }
  }
  fs.mkdirSync(reportDir, { recursive: true })
  fs.writeFileSync(path.join(reportDir, 'data-contract-report.json'), JSON.stringify(report, null, 2) + '\n')
  const lines = [
    '# Data contract report',
    '',
    // G1收尾：去掉 Generated 时间戳行 —— 每次跑门禁必变导致 reports/*.md 常脏；
    // 时间戳仍保留在 gitignored 的 data-contract-report.json（generatedAt）中备查
    `- Schemas: ${report.summary.schemaCount}`,
    `- Referenced collections: ${report.summary.collectionCount}`,
    `- Missing schemas: ${report.summary.missingSchemaCount}`,
    `- Field drift entries: ${report.summary.fieldDriftCount}`,
    '',
    '## Missing schemas',
    ''
  ]
  if (!missingSchemas.length) lines.push('None', '')
  for (const item of missingSchemas) lines.push(`- **${item.collection}** — ${item.files.join(', ')}`)
  lines.push('', '## Field drift', '')
  if (!fieldDrift.length) lines.push('None', '')
  for (const item of fieldDrift.slice(0, 200)) lines.push(`- **${item.collection}.${item.field}** [\`${item.driftKind}\`] — ${item.files.join(', ')}`)
  lines.push('', '## Migration review queue', '')
  for (const item of knownMigrations) lines.push(`- **${item.source} → ${item.target}** (${item.status}) — ${JSON.stringify(item.mapping)}`)
  fs.writeFileSync(path.join(reportDir, 'data-contract-report.md'), lines.join('\n') + '\n')
  return report
}

const report = buildReport()
console.log(JSON.stringify(report.summary))