import { BaseCommand, flags } from '@adonisjs/core/ace'
import { CommandOptions } from '@adonisjs/core/types/ace'
import db from '@adonisjs/lucid/services/db'
import GarbageFilter from '#services/GarbageFilter'

/**
 * Learns the garbage upload signatures now (gc:scheduler does it every
 * GARBAGE_LEARN_INTERVAL_MS) and lists them (CONCEPT.md D42). --seed first
 * fills signature_probes from a sample of the files probed before it
 * existed, so there is something to learn from right away.
 */
export default class GarbageLearn extends BaseCommand {
  public static commandName = 'garbage:learn'
  public static description = 'Learns which upload signatures are garbage and lists them.'

  public static options: CommandOptions = {
    startApp: true,
  }

  @flags.number({ description: 'First add probed releases from TABLESAMPLE SYSTEM (percent) of files' })
  declare seed?: number

  public async run() {
    if (this.seed) {
      const started = Date.now()
      const added = await GarbageFilter.seed(this.seed)
      this.logger.info(`Seeded ${added} probed releases from ${this.seed}% of files in ${Math.round((Date.now() - started) / 1000)} s.`)
    }
    const { signatures, dropped, changed } = await GarbageFilter.learn()
    for (const line of changed) this.logger.info(`Changed: ${line}`)
    const rows = await db.from('upload_signatures').orderBy('probed', 'desc').limit(40)
    for (const row of rows) {
      const rate = row.probed ? ((100 * row.named) / row.probed).toFixed(1) : '-'
      this.logger.info(`${row.dropped ? 'DROP' : 'keep'}  ${row.tool.padEnd(10)} ${(row.group_name || '(no group)').padEnd(36)} ${String(row.named).padStart(6)}/${String(row.probed).padEnd(6)} named (${rate}%)`)
    }
    this.logger.info(`${dropped} of ${signatures} signatures dropped.`)
  }
}