feat: replace Kaggle CSV with Wikipedia scraper for historical match data
Add scripts/scrape-wikipedia.ts that fetches all 22 World Cups (1930–2022) from English Wikipedia via MediaWiki API, handles group sub-pages, AET/penalty detection, and goal parsing, writing openfootball-format JSON to app/data/openfootball/. Rewrite scripts/seed.ts to read these local JSON files instead of the Kaggle CSV, producing 965 matches and 2716 goals with per-group assignments for all historical tournaments (enabling group standings on tournament pages). Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
This commit is contained in:
@@ -0,0 +1,549 @@
|
||||
import { load } from 'cheerio'
|
||||
import type { CheerioAPI } from 'cheerio'
|
||||
import type { Cheerio } from 'cheerio'
|
||||
import type { Element } from 'domhandler'
|
||||
import { mkdirSync, writeFileSync } from 'fs'
|
||||
import path from 'path'
|
||||
import { fileURLToPath } from 'url'
|
||||
|
||||
const __dirname = path.dirname(fileURLToPath(import.meta.url))
|
||||
const DATA_DIR = path.join(__dirname, '../app/data/openfootball')
|
||||
|
||||
const YEARS = [
|
||||
1930,1934,1938,1950,1954,1958,1962,1966,1970,1974,
|
||||
1978,1982,1986,1990,1994,1998,2002,2006,2010,2014,2018,2022,
|
||||
]
|
||||
|
||||
const delay = (ms: number) => new Promise(r => setTimeout(r, ms))
|
||||
|
||||
// ── Types ──────────────────────────────────────────────────────────────────
|
||||
|
||||
type Goal = {
|
||||
name: string
|
||||
minute?: number
|
||||
offset?: number
|
||||
penalty?: boolean
|
||||
owngoal?: boolean
|
||||
}
|
||||
|
||||
type ScoreObj = {
|
||||
ft?: [number, number]
|
||||
et?: [number, number]
|
||||
p?: [number, number]
|
||||
}
|
||||
|
||||
type Match = {
|
||||
round: string
|
||||
group?: string
|
||||
date?: string
|
||||
time?: string
|
||||
team1: string
|
||||
team2: string
|
||||
score?: ScoreObj
|
||||
goals1?: Goal[]
|
||||
goals2?: Goal[]
|
||||
ground?: string
|
||||
}
|
||||
|
||||
type Stadium = { name: string; city: string }
|
||||
type Player = { name: string; number?: number; pos?: string; date_of_birth?: string }
|
||||
type Squad = { name: string; players: Player[] }
|
||||
type Group = { name: string; teams: string[] }
|
||||
|
||||
// ── Fetch ──────────────────────────────────────────────────────────────────
|
||||
|
||||
async function fetchWikiHtml(page: string, retries = 5): Promise<string | null> {
|
||||
const url = `https://en.wikipedia.org/w/api.php?action=parse&page=${encodeURIComponent(page)}&format=json&prop=text&disabletoc=1`
|
||||
for (let attempt = 0; attempt < retries; attempt++) {
|
||||
try {
|
||||
if (attempt > 0) await delay(3000 * attempt)
|
||||
const res = await fetch(url, { headers: { 'User-Agent': 'WorldCupScraper/1.0 (github.com/worldcup)' } })
|
||||
if (!res.ok) continue
|
||||
const data = await res.json() as { parse?: { text?: { '*': string } } }
|
||||
const html = data?.parse?.text?.['*']
|
||||
if (html) return html
|
||||
} catch {
|
||||
// retry
|
||||
}
|
||||
}
|
||||
return null
|
||||
}
|
||||
|
||||
// ── Score parsing ──────────────────────────────────────────────────────────
|
||||
|
||||
function parseScoreText(text: string): [number, number] | null {
|
||||
const m = text.match(/(\d+)\s*[–\-]\s*(\d+)/)
|
||||
if (!m) return null
|
||||
return [parseInt(m[1]), parseInt(m[2])]
|
||||
}
|
||||
|
||||
// ── Team name extraction ───────────────────────────────────────────────────
|
||||
|
||||
function extractTeam($: CheerioAPI, $cell: Cheerio<Element>): string {
|
||||
let name = ''
|
||||
$cell.find('a').each((_, a) => {
|
||||
const $a = $(a)
|
||||
if (!$a.find('img').length && $a.text().trim()) {
|
||||
name = $a.text().trim()
|
||||
return false
|
||||
}
|
||||
})
|
||||
return name
|
||||
}
|
||||
|
||||
// ── Goal parsing ───────────────────────────────────────────────────────────
|
||||
|
||||
function parseGoals($: CheerioAPI, $td: Cheerio<Element>): Goal[] {
|
||||
const goals: Goal[] = []
|
||||
|
||||
$td.find('li').each((_, li) => {
|
||||
const $li = $(li)
|
||||
|
||||
// Player name: first <a> NOT inside .fb-goal
|
||||
let playerName = ''
|
||||
$li.find('a').each((_, a) => {
|
||||
if (!$(a).closest('.fb-goal').length) {
|
||||
const t = $(a).text().trim()
|
||||
if (t) { playerName = t; return false }
|
||||
}
|
||||
})
|
||||
if (!playerName) return
|
||||
|
||||
const $fbGoal = $li.find('.fb-goal')
|
||||
if (!$fbGoal.length) return
|
||||
|
||||
// Each direct child <span> inside .fb-goal (excluding image wrapper)
|
||||
$fbGoal.children('span').each((_, span) => {
|
||||
const $span = $(span)
|
||||
if ($span.attr('typeof')) return // image wrapper
|
||||
|
||||
const text = $span.text()
|
||||
const minMatch = text.match(/(\d+)(?:\+(\d+))?['′]/)
|
||||
if (!minMatch) return
|
||||
|
||||
const minute = parseInt(minMatch[1])
|
||||
const offset = minMatch[2] ? parseInt(minMatch[2]) : 0
|
||||
const isPen = text.includes('pen.')
|
||||
const isOG = text.includes('o.g.')
|
||||
|
||||
const goal: Goal = { name: playerName }
|
||||
if (!isNaN(minute)) goal.minute = minute
|
||||
if (offset) goal.offset = offset
|
||||
if (isPen) goal.penalty = true
|
||||
if (isOG) goal.owngoal = true
|
||||
goals.push(goal)
|
||||
})
|
||||
})
|
||||
|
||||
return goals
|
||||
}
|
||||
|
||||
// ── Ground extraction ──────────────────────────────────────────────────────
|
||||
|
||||
function extractGround($: CheerioAPI, $box: Cheerio<Element>): string {
|
||||
const $loc = $box.find('[itemprop="name address"]').first()
|
||||
if ($loc.length) return $loc.text().trim()
|
||||
return $box.find('.fright').first().text().split('\n')[0].trim()
|
||||
}
|
||||
|
||||
function parseGroundParts(ground: string): { name: string; city: string } {
|
||||
const commaIdx = ground.indexOf(',')
|
||||
if (commaIdx !== -1) {
|
||||
return {
|
||||
name: ground.slice(0, commaIdx).trim(),
|
||||
city: ground.slice(commaIdx + 1).trim(),
|
||||
}
|
||||
}
|
||||
return { name: ground, city: '' }
|
||||
}
|
||||
|
||||
// ── Footballbox parsing ────────────────────────────────────────────────────
|
||||
|
||||
function parseBox(
|
||||
$: CheerioAPI,
|
||||
$box: Cheerio<Element>,
|
||||
round: string,
|
||||
group: string | null,
|
||||
): Match | null {
|
||||
const team1 = extractTeam($, $box.find('.fhome'))
|
||||
const team2 = extractTeam($, $box.find('.faway'))
|
||||
if (!team1 || !team2) return null
|
||||
|
||||
const dateStr = $box.find('.bday, .dtstart').first().text().trim() || undefined
|
||||
|
||||
const timeText = $box.find('.ftime').first().text().trim()
|
||||
const timeMatch = timeText.match(/(\d{2}:\d{2})/)
|
||||
const timeStr = timeMatch?.[1]
|
||||
|
||||
const scoreText = $box.find('.fscore').first().text().trim()
|
||||
const hasAET = scoreText.toLowerCase().includes('a.e.t.')
|
||||
const scoreArr = parseScoreText(scoreText)
|
||||
|
||||
// Use first fgoals row only (exclude penalty shootout row)
|
||||
const $regularRow = $box.find('tr.fgoals').first()
|
||||
const goals1 = parseGoals($, $regularRow.find('.fhgoal'))
|
||||
const goals2 = parseGoals($, $regularRow.find('.fagoal'))
|
||||
|
||||
// Penalty shootout score: row after "Penalties" header tr
|
||||
let penScore: [number, number] | undefined
|
||||
$box.find('tr').each((_, tr) => {
|
||||
const $tr = $(tr)
|
||||
if ($tr.find('th[colspan]').text().toLowerCase().includes('penalt')) {
|
||||
const penText = $tr.next('tr').find('th').not('.fhome,.faway').first().text().trim()
|
||||
const ps = parseScoreText(penText)
|
||||
if (ps) penScore = ps
|
||||
return false
|
||||
}
|
||||
})
|
||||
|
||||
let score: ScoreObj | undefined
|
||||
if (scoreArr) {
|
||||
if (hasAET) {
|
||||
// scoreArr is ET total; compute FT from goals in ≤90 min
|
||||
const ftGoals = (gs: Goal[], includeOG = false) =>
|
||||
gs.filter(g => {
|
||||
const w90 = g.minute === undefined || g.minute <= 90
|
||||
return includeOG ? g.owngoal === true && w90 : !g.owngoal && w90
|
||||
}).length
|
||||
const ftHome = ftGoals(goals1) + ftGoals(goals2, true)
|
||||
const ftAway = ftGoals(goals2) + ftGoals(goals1, true)
|
||||
score = { ft: [ftHome, ftAway], et: scoreArr }
|
||||
} else {
|
||||
score = { ft: scoreArr }
|
||||
}
|
||||
if (penScore) score.p = penScore
|
||||
}
|
||||
|
||||
const ground = extractGround($, $box) || undefined
|
||||
|
||||
return {
|
||||
round,
|
||||
...(group ? { group } : {}),
|
||||
...(dateStr ? { date: dateStr } : {}),
|
||||
...(timeStr ? { time: timeStr } : {}),
|
||||
team1,
|
||||
team2,
|
||||
...(score ? { score } : {}),
|
||||
...(goals1.length ? { goals1 } : {}),
|
||||
...(goals2.length ? { goals2 } : {}),
|
||||
...(ground ? { ground } : {}),
|
||||
}
|
||||
}
|
||||
|
||||
// ── Collect matches from a pre-loaded page ─────────────────────────────────
|
||||
|
||||
function collectBoxes(
|
||||
$: CheerioAPI,
|
||||
round: string,
|
||||
group: string | null,
|
||||
): Match[] {
|
||||
const matches: Match[] = []
|
||||
$('.footballbox').each((_, el) => {
|
||||
const m = parseBox($, $(el), round, group)
|
||||
if (m) matches.push(m)
|
||||
})
|
||||
return matches
|
||||
}
|
||||
|
||||
// ── Section heading state machine ──────────────────────────────────────────
|
||||
|
||||
type State = {
|
||||
active: boolean
|
||||
round: string
|
||||
group: string | null
|
||||
}
|
||||
|
||||
function processHeading(text: string, level: number, state: State): void {
|
||||
const t = text.toLowerCase().trim()
|
||||
|
||||
if (level === 2) {
|
||||
if (/group stage/i.test(t) && !/second/i.test(t)) {
|
||||
state.active = true; state.round = 'Group stage'; state.group = null
|
||||
} else if (/first group stage/i.test(t)) {
|
||||
state.active = true; state.round = 'Group stage'; state.group = null
|
||||
} else if (/second group stage/i.test(t)) {
|
||||
state.active = true; state.round = 'Second group stage'; state.group = null
|
||||
} else if (t === 'final round') {
|
||||
state.active = true; state.round = 'Final round'; state.group = null
|
||||
} else if (/final tournament/i.test(t)) {
|
||||
state.active = true; state.round = ''; state.group = null
|
||||
} else if (/knock.?out stage/i.test(t)) {
|
||||
state.active = true; state.round = ''; state.group = null
|
||||
} else if (/round of 16/i.test(t)) {
|
||||
state.active = true; state.round = 'Round of 16'; state.group = null
|
||||
} else if (/quarter.final/i.test(t)) {
|
||||
state.active = true; state.round = 'Quarter-finals'; state.group = null
|
||||
} else if (/semi.final/i.test(t)) {
|
||||
state.active = true; state.round = 'Semi-finals'; state.group = null
|
||||
} else if (/third.place|match for third|play.off for third/i.test(t)) {
|
||||
state.active = true; state.round = 'Third-place match'; state.group = null
|
||||
} else if (t === 'final') {
|
||||
state.active = true; state.round = 'Final'; state.group = null
|
||||
} else {
|
||||
state.active = false
|
||||
}
|
||||
return
|
||||
}
|
||||
|
||||
if (!state.active) return
|
||||
|
||||
if (level === 3 || level === 4) {
|
||||
if (/^group [a-h1-9]+$/i.test(t)) {
|
||||
state.group = text.trim()
|
||||
} else if (/round of 32/i.test(t)) {
|
||||
state.round = 'Round of 32'; state.group = null
|
||||
} else if (/round of 16/i.test(t)) {
|
||||
state.round = 'Round of 16'; state.group = null
|
||||
} else if (/quarter.final/i.test(t)) {
|
||||
state.round = 'Quarter-finals'; state.group = null
|
||||
} else if (/semi.final/i.test(t)) {
|
||||
state.round = 'Semi-finals'; state.group = null
|
||||
} else if (/third.place|match for third|play.off for third/i.test(t)) {
|
||||
state.round = 'Third-place match'; state.group = null
|
||||
} else if (t === 'final') {
|
||||
state.round = 'Final'; state.group = null
|
||||
}
|
||||
// bracket, draw, seeding, replay → keep current state
|
||||
}
|
||||
}
|
||||
|
||||
// ── Main year scraper ──────────────────────────────────────────────────────
|
||||
|
||||
type YearResult = {
|
||||
matches: Match[]
|
||||
stadiums: Map<string, Stadium>
|
||||
groups: Map<string, Set<string>>
|
||||
}
|
||||
|
||||
async function scrapeYear(year: number, mainHtml: string): Promise<YearResult> {
|
||||
const $ = load(mainHtml)
|
||||
const matches: Match[] = []
|
||||
const stadiums = new Map<string, Stadium>()
|
||||
const groups = new Map<string, Set<string>>()
|
||||
|
||||
const state: State = { active: false, round: '', group: null }
|
||||
|
||||
// Maps group name → sub-page to fetch (if main page has no matches for that group)
|
||||
const groupSubpages = new Map<string, string>()
|
||||
// Groups that got at least one match from the main page
|
||||
const groupsOnMainPage = new Set<string>()
|
||||
|
||||
function recordMatch(m: Match) {
|
||||
matches.push(m)
|
||||
if (m.group) groupsOnMainPage.add(m.group)
|
||||
if (m.ground) {
|
||||
const { name, city } = parseGroundParts(m.ground)
|
||||
if (name && !stadiums.has(name)) stadiums.set(name, { name, city })
|
||||
}
|
||||
if (m.group) {
|
||||
if (!groups.has(m.group)) groups.set(m.group, new Set())
|
||||
groups.get(m.group)!.add(m.team1)
|
||||
groups.get(m.group)!.add(m.team2)
|
||||
}
|
||||
}
|
||||
|
||||
// Walk elements in document order: headings, hatnotes, footballboxes
|
||||
$('.mw-parser-output').find('div.mw-heading, .footballbox, .hatnote').each((_, el) => {
|
||||
const $el = $(el)
|
||||
|
||||
if ($el.hasClass('mw-heading')) {
|
||||
const $h = $el.find('h2, h3, h4').first()
|
||||
const level = parseInt($h.prop('tagName')?.slice(1) ?? '9')
|
||||
const text = $h.text().replace(/\[edit\]/g, '').trim()
|
||||
processHeading(text, level, state)
|
||||
|
||||
} else if ($el.hasClass('hatnote') && $el.text().includes('Main article')) {
|
||||
// Record sub-page link for current group context (for fallback if no main-page matches)
|
||||
if (state.active && state.group) {
|
||||
const link = $el.find('a[href^="/wiki/"]').first().attr('href')
|
||||
if (link) {
|
||||
const page = link.replace('/wiki/', '').split('#')[0]
|
||||
if (/World_Cup_Group/i.test(page) && !groupSubpages.has(state.group)) {
|
||||
groupSubpages.set(state.group, page)
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
} else if ($el.hasClass('footballbox')) {
|
||||
if (!state.active) return
|
||||
const round = state.round || state.group || 'Unknown'
|
||||
const m = parseBox($, $el, round, state.group)
|
||||
if (m) recordMatch(m)
|
||||
}
|
||||
})
|
||||
|
||||
// Fetch group sub-pages for any group that got 0 matches from main page
|
||||
for (const [group, page] of groupSubpages) {
|
||||
if (groupsOnMainPage.has(group)) continue
|
||||
|
||||
await delay(1200)
|
||||
const subHtml = await fetchWikiHtml(page)
|
||||
if (!subHtml) { process.stdout.write(`(failed: ${page}) `); continue }
|
||||
|
||||
// Determine the round for this group from the state machine result
|
||||
// (we'll reconstruct from the main-page walk state — use the round that was active when this group was seen)
|
||||
// Since we can't easily recover state here, we re-walk to find the round for this group
|
||||
let round = 'Group stage'
|
||||
let foundGroup = false
|
||||
const stateTemp: State = { active: false, round: '', group: null }
|
||||
$('.mw-parser-output').find('div.mw-heading').each((_, el) => {
|
||||
const $h = $(el).find('h2, h3, h4').first()
|
||||
const level = parseInt($h.prop('tagName')?.slice(1) ?? '9')
|
||||
const text = $h.text().replace(/\[edit\]/g, '').trim()
|
||||
processHeading(text, level, stateTemp)
|
||||
if (stateTemp.group === group) {
|
||||
round = stateTemp.round || 'Group stage'
|
||||
foundGroup = true
|
||||
return false
|
||||
}
|
||||
})
|
||||
|
||||
const $sub = load(subHtml)
|
||||
const subMatches = collectBoxes($sub, round || 'Group stage', group)
|
||||
for (const m of subMatches) {
|
||||
recordMatch(m)
|
||||
}
|
||||
process.stdout.write(`[+${page.slice(-8)}] `)
|
||||
}
|
||||
|
||||
return { matches, stadiums, groups }
|
||||
}
|
||||
|
||||
// ── Squad page scraper ─────────────────────────────────────────────────────
|
||||
|
||||
function scrapeSquads(html: string): Squad[] {
|
||||
const $ = load(html)
|
||||
const squads: Squad[] = []
|
||||
let currentTeam: Squad | null = null
|
||||
|
||||
$('.mw-parser-output').find('div.mw-heading, tr.nat-fs-player').each((_, el) => {
|
||||
const $el = $(el)
|
||||
|
||||
if ($el.hasClass('mw-heading')) {
|
||||
const $h = $el.find('h3, h4').first()
|
||||
if (!$h.length) return
|
||||
const level = parseInt($h.prop('tagName')?.slice(1) ?? '9')
|
||||
if (level !== 3) return
|
||||
const name = $h.text().replace(/\[edit\]/g, '').trim()
|
||||
if (/^group /i.test(name)) return // skip group headers
|
||||
currentTeam = { name, players: [] }
|
||||
squads.push(currentTeam)
|
||||
return
|
||||
}
|
||||
|
||||
if (!currentTeam) return
|
||||
|
||||
let number: number | undefined
|
||||
let pos: string | undefined
|
||||
let playerName = ''
|
||||
let dob: string | undefined
|
||||
|
||||
$el.find('td, th[scope="row"]').each((i, td) => {
|
||||
const $td = $(td)
|
||||
const text = $td.text().trim()
|
||||
|
||||
if ($td.is('th[scope="row"]')) {
|
||||
playerName = $td.find('a').first().text().trim() || text
|
||||
} else if (i === 0 && !playerName) {
|
||||
const n = parseInt(text)
|
||||
if (!isNaN(n)) number = n
|
||||
} else if (i === 1 && !playerName && !pos) {
|
||||
const posLink = $td.find('a').first().text().trim()
|
||||
if (['GK', 'DF', 'MF', 'FW'].includes(posLink)) pos = posLink
|
||||
}
|
||||
|
||||
const $bday = $td.find('.bday')
|
||||
if ($bday.length) dob = $bday.text().trim()
|
||||
})
|
||||
|
||||
if (!playerName) return
|
||||
|
||||
const player: Player = { name: playerName }
|
||||
if (number !== undefined) player.number = number
|
||||
if (pos) player.pos = pos
|
||||
if (dob) player.date_of_birth = dob
|
||||
currentTeam.players.push(player)
|
||||
})
|
||||
|
||||
return squads
|
||||
}
|
||||
|
||||
// ── Output ─────────────────────────────────────────────────────────────────
|
||||
|
||||
function writeOutput(
|
||||
year: number,
|
||||
matches: Match[],
|
||||
stadiums: Map<string, Stadium>,
|
||||
groups: Map<string, Set<string>>,
|
||||
squads: Squad[],
|
||||
): void {
|
||||
const dir = path.join(DATA_DIR, String(year))
|
||||
mkdirSync(dir, { recursive: true })
|
||||
|
||||
writeFileSync(
|
||||
path.join(dir, 'worldcup.json'),
|
||||
JSON.stringify({ matches }, null, 2),
|
||||
'utf-8',
|
||||
)
|
||||
|
||||
if (stadiums.size > 0) {
|
||||
writeFileSync(
|
||||
path.join(dir, 'worldcup.stadiums.json'),
|
||||
JSON.stringify({ stadiums: Array.from(stadiums.values()) }, null, 2),
|
||||
'utf-8',
|
||||
)
|
||||
}
|
||||
|
||||
const groupList: Group[] = []
|
||||
groups.forEach((teams, name) => {
|
||||
groupList.push({ name, teams: Array.from(teams) })
|
||||
})
|
||||
if (groupList.length > 0) {
|
||||
writeFileSync(
|
||||
path.join(dir, 'worldcup.groups.json'),
|
||||
JSON.stringify({ groups: groupList }, null, 2),
|
||||
'utf-8',
|
||||
)
|
||||
}
|
||||
|
||||
if (squads.length > 0) {
|
||||
writeFileSync(
|
||||
path.join(dir, 'worldcup.squads.json'),
|
||||
JSON.stringify(squads, null, 2),
|
||||
'utf-8',
|
||||
)
|
||||
}
|
||||
}
|
||||
|
||||
// ── Entry point ────────────────────────────────────────────────────────────
|
||||
|
||||
async function main() {
|
||||
const onlyYear = process.argv[2] ? parseInt(process.argv[2]) : null
|
||||
const yearsToScrape = onlyYear ? [onlyYear] : YEARS
|
||||
|
||||
console.log(`Scraping ${yearsToScrape.length} World Cup(s) from Wikipedia...`)
|
||||
|
||||
for (const year of yearsToScrape) {
|
||||
process.stdout.write(` ${year}... `)
|
||||
|
||||
const mainHtml = await fetchWikiHtml(`${year}_FIFA_World_Cup`)
|
||||
if (!mainHtml) { console.log('FAILED'); continue }
|
||||
|
||||
const { matches, stadiums, groups } = await scrapeYear(year, mainHtml)
|
||||
|
||||
await delay(600)
|
||||
|
||||
const squadHtml = await fetchWikiHtml(`${year}_FIFA_World_Cup_squads`)
|
||||
const squads = squadHtml ? scrapeSquads(squadHtml) : []
|
||||
|
||||
writeOutput(year, matches, stadiums, groups, squads)
|
||||
|
||||
console.log(`${matches.length} matches, ${stadiums.size} stadiums, ${groups.size} groups, ${squads.length} teams`)
|
||||
|
||||
await delay(600)
|
||||
}
|
||||
|
||||
console.log('\nDone! Files written to app/data/openfootball/{year}/')
|
||||
}
|
||||
|
||||
main().catch(e => { console.error(e); process.exit(1) })
|
||||
+193
-127
@@ -1,16 +1,23 @@
|
||||
import postgres from 'postgres'
|
||||
import { drizzle } from 'drizzle-orm/postgres-js'
|
||||
import { sql } from 'drizzle-orm'
|
||||
import { readFileSync } from 'fs'
|
||||
import { readFileSync, existsSync } from 'fs'
|
||||
import path from 'path'
|
||||
import { fileURLToPath } from 'url'
|
||||
import { getIso } from '../lib/iso-codes'
|
||||
|
||||
const DATABASE_URL = process.env.DATABASE_URL ?? 'postgres://wc:wc@localhost:5432/worldcup'
|
||||
const __dirname = path.dirname(fileURLToPath(import.meta.url))
|
||||
const DATA_DIR = path.join(__dirname, '../app/data/kaggle')
|
||||
const DATA_DIR = path.join(__dirname, '../app/data')
|
||||
const KAGGLE_DIR = path.join(DATA_DIR, 'kaggle')
|
||||
const WC_DIR = path.join(DATA_DIR, 'openfootball')
|
||||
|
||||
// Third/fourth place not present in Kaggle world_cup.csv
|
||||
const YEARS = [
|
||||
1930,1934,1938,1950,1954,1958,1962,1966,1970,1974,
|
||||
1978,1982,1986,1990,1994,1998,2002,2006,2010,2014,2018,2022,
|
||||
]
|
||||
|
||||
// Third/fourth place not reliably in source data for older years
|
||||
const PLACEMENTS: Record<number, { third?: string; fourth?: string }> = {
|
||||
1930: { third: 'USA', fourth: 'Yugoslavia' },
|
||||
1934: { third: 'Germany', fourth: 'Austria' },
|
||||
@@ -35,7 +42,7 @@ const PLACEMENTS: Record<number, { third?: string; fourth?: string }> = {
|
||||
2022: { third: 'Croatia', fourth: 'Morocco' },
|
||||
}
|
||||
|
||||
// Normalize Kaggle team names to match openfootball / our canonical names
|
||||
// Normalize team names from Wikipedia to canonical DB names
|
||||
const TEAM_ALIASES: Record<string, string> = {
|
||||
'West Germany': 'Germany',
|
||||
'Korea Republic': 'South Korea',
|
||||
@@ -46,7 +53,7 @@ function normTeam(name: string): string {
|
||||
return TEAM_ALIASES[name] ?? name
|
||||
}
|
||||
|
||||
// Minimal RFC-4180 CSV parser — no external dependency needed
|
||||
// Minimal RFC-4180 CSV parser
|
||||
function parseCsv(content: string): Record<string, string>[] {
|
||||
const rows: string[][] = []
|
||||
let row: string[] = []
|
||||
@@ -78,38 +85,36 @@ function parseCsv(content: string): Record<string, string>[] {
|
||||
.map(r => Object.fromEntries(headers.map((h, i) => [h.trim(), (r[i] ?? '').trim()])))
|
||||
}
|
||||
|
||||
type GoalEntry = { name: string; minute: number | null; offset: number; isPenalty: boolean; isOwnGoal: boolean }
|
||||
|
||||
// Parse "Player Name · 57" or "Player (OG) · 90+3" → GoalEntry
|
||||
function parseGoalStr(entry: string, isPenalty = false, isOwnGoal = false): GoalEntry | null {
|
||||
const dot = entry.lastIndexOf('·')
|
||||
if (dot === -1) return null
|
||||
const name = entry.slice(0, dot).trim()
|
||||
.replace(/\s*\(P\)\s*$/, '').replace(/\s*\(OG\)\s*$/, '').trim()
|
||||
if (!name) return null
|
||||
const minRaw = entry.slice(dot + 1).trim()
|
||||
const plusIdx = minRaw.indexOf('+')
|
||||
let minute: number | null, offset = 0
|
||||
if (plusIdx !== -1) {
|
||||
minute = parseInt(minRaw.slice(0, plusIdx))
|
||||
offset = parseInt(minRaw.slice(plusIdx + 1)) || 0
|
||||
} else {
|
||||
const m = parseInt(minRaw)
|
||||
minute = isNaN(m) ? null : m
|
||||
}
|
||||
return { name, minute, offset, isPenalty, isOwnGoal }
|
||||
function readJson<T>(filePath: string): T | null {
|
||||
if (!existsSync(filePath)) return null
|
||||
try { return JSON.parse(readFileSync(filePath, 'utf-8')) as T } catch { return null }
|
||||
}
|
||||
|
||||
function parseGoalCol(col: string, isPenalty = false, isOwnGoal = false): GoalEntry[] {
|
||||
if (!col?.trim()) return []
|
||||
return col.split('|').map(e => parseGoalStr(e.trim(), isPenalty, isOwnGoal)).filter(Boolean) as GoalEntry[]
|
||||
// ── Types matching scrape-wikipedia.ts output ──────────────────────────────
|
||||
|
||||
type RawGoal = { name: string; minute?: string | number; offset?: number; penalty?: boolean; owngoal?: boolean }
|
||||
type RawScore = { ft?: number[]; ht?: number[]; et?: number[]; p?: number[] }
|
||||
type RawMatch = {
|
||||
round?: string; date?: string; time?: string;
|
||||
team1: string; team2: string; score?: RawScore;
|
||||
goals1?: RawGoal[]; goals2?: RawGoal[];
|
||||
group?: string; ground?: string;
|
||||
}
|
||||
type RawData = { matches: RawMatch[] }
|
||||
type RawStadiums = { stadiums: { name: string; city: string; cc?: string; capacity?: number; timezone?: string; coords?: string }[] }
|
||||
type RawSquad = { name: string; players: { name: string; number?: number; pos?: string; date_of_birth?: string }[] }
|
||||
|
||||
function parseScore(score: RawScore | undefined) {
|
||||
if (!score) return {}
|
||||
if (Array.isArray(score)) return { ft: score as number[] }
|
||||
return { ft: score.ft, ht: score.ht, et: score.et, p: score.p }
|
||||
}
|
||||
|
||||
async function run() {
|
||||
const client = postgres(DATABASE_URL, { max: 5 })
|
||||
const db = drizzle(client)
|
||||
|
||||
// Create tables (mirrors sync.ts DDL — runs first on a fresh DB)
|
||||
// Create tables
|
||||
await db.execute(sql`
|
||||
CREATE TABLE IF NOT EXISTS tournaments (
|
||||
year INTEGER PRIMARY KEY,
|
||||
@@ -203,7 +208,6 @@ async function run() {
|
||||
|
||||
const force = process.argv.includes('--force') || process.argv.includes('-f')
|
||||
|
||||
// Skip if already seeded (idempotency check)
|
||||
if (!force) {
|
||||
const existing = await db.execute(sql`SELECT COUNT(*)::int AS cnt FROM tournaments WHERE year < 2026`)
|
||||
if ((existing[0] as { cnt: number }).cnt > 0) {
|
||||
@@ -216,20 +220,24 @@ async function run() {
|
||||
if (force) {
|
||||
console.log('--force: clearing historical data...')
|
||||
await db.execute(sql`DELETE FROM goals WHERE match_id IN (SELECT id FROM matches WHERE tournament_year < 2026)`)
|
||||
await db.execute(sql`DELETE FROM squads WHERE tournament_year < 2026`)
|
||||
await db.execute(sql`DELETE FROM group_standings WHERE tournament_year < 2026`)
|
||||
await db.execute(sql`DELETE FROM stadiums WHERE tournament_year < 2026`)
|
||||
await db.execute(sql`DELETE FROM matches WHERE tournament_year < 2026`)
|
||||
await db.execute(sql`DELETE FROM tournaments WHERE year < 2026`)
|
||||
}
|
||||
|
||||
console.log('Seeding from Kaggle data (1930–2022)...')
|
||||
console.log('Seeding historical data (1930–2022)...')
|
||||
|
||||
const teamCache = new Map<string, number>()
|
||||
|
||||
async function upsertTeam(rawName: string): Promise<number> {
|
||||
const name = normTeam(rawName)
|
||||
if (teamCache.has(name)) return teamCache.get(name)!
|
||||
const iso2 = getIso(name)
|
||||
const [row] = await db.execute(sql`
|
||||
INSERT INTO teams (name, iso2)
|
||||
VALUES (${name}, ${getIso(name) ?? null})
|
||||
VALUES (${name}, ${iso2 ?? null})
|
||||
ON CONFLICT (name) DO UPDATE SET name = EXCLUDED.name
|
||||
RETURNING id
|
||||
`)
|
||||
@@ -238,8 +246,8 @@ async function run() {
|
||||
return id
|
||||
}
|
||||
|
||||
// 1. Tournaments from world_cup.csv
|
||||
const wcRows = parseCsv(readFileSync(path.join(DATA_DIR, 'world_cup.csv'), 'utf-8'))
|
||||
// 1. Tournaments from world_cup.csv (host, winner, runner_up)
|
||||
const wcRows = parseCsv(readFileSync(path.join(KAGGLE_DIR, 'world_cup.csv'), 'utf-8'))
|
||||
for (const r of wcRows) {
|
||||
const year = parseInt(r['Year'])
|
||||
if (isNaN(year)) continue
|
||||
@@ -247,135 +255,193 @@ async function run() {
|
||||
const runnerUp = normTeam(r['Runner-Up'] || '')
|
||||
const p = PLACEMENTS[year] ?? {}
|
||||
await db.execute(sql`
|
||||
INSERT INTO tournaments (year, host, winner, runner_up, third_place, fourth_place, teams_count, matches_count)
|
||||
INSERT INTO tournaments (year, host, winner, runner_up, third_place, fourth_place, teams_count)
|
||||
VALUES (
|
||||
${year}, ${r['Host']},
|
||||
${winner || null}, ${runnerUp || null},
|
||||
${p.third ?? null}, ${p.fourth ?? null},
|
||||
${parseInt(r['Teams']) || null}, ${parseInt(r['Matches']) || null}
|
||||
${parseInt(r['Teams']) || null}
|
||||
)
|
||||
ON CONFLICT (year) DO UPDATE SET
|
||||
host = EXCLUDED.host,
|
||||
winner = EXCLUDED.winner,
|
||||
runner_up = EXCLUDED.runner_up,
|
||||
host = EXCLUDED.host,
|
||||
winner = EXCLUDED.winner,
|
||||
runner_up = EXCLUDED.runner_up,
|
||||
third_place = EXCLUDED.third_place,
|
||||
fourth_place = EXCLUDED.fourth_place,
|
||||
teams_count = EXCLUDED.teams_count,
|
||||
matches_count = EXCLUDED.matches_count
|
||||
teams_count = EXCLUDED.teams_count
|
||||
`)
|
||||
}
|
||||
|
||||
// 2. Matches + goals from matches_1930_2022.csv
|
||||
const matchRows = parseCsv(readFileSync(path.join(DATA_DIR, 'matches_1930_2022.csv'), 'utf-8'))
|
||||
// 2. Per-year match/stadium/squad data from openfootball JSON files
|
||||
let totalMatches = 0
|
||||
let totalGoals = 0
|
||||
|
||||
let totalMatches = 0, totalGoals = 0
|
||||
for (const r of matchRows) {
|
||||
const year = parseInt(r['Year'])
|
||||
if (isNaN(year)) continue
|
||||
|
||||
const t1Id = await upsertTeam(r['home_team'])
|
||||
const t2Id = await upsertTeam(r['away_team'])
|
||||
|
||||
const homeScore = r['home_score'] !== '' ? parseInt(r['home_score']) : null
|
||||
const awayScore = r['away_score'] !== '' ? parseInt(r['away_score']) : null
|
||||
const homePen = r['home_penalty'] !== '' ? parseInt(r['home_penalty']) : null
|
||||
const awayPen = r['away_penalty'] !== '' ? parseInt(r['away_penalty']) : null
|
||||
const dateStr = r['Date'] || null
|
||||
|
||||
// Parse all goal columns
|
||||
const homeGoals = parseGoalCol(r['home_goal'])
|
||||
const awayGoals = parseGoalCol(r['away_goal'])
|
||||
const homePenGoals = parseGoalCol(r['home_penalty_goal'], true)
|
||||
const awayPenGoals = parseGoalCol(r['away_penalty_goal'], true)
|
||||
// home_own_goal = home player scored OG → goal credited to AWAY team
|
||||
const homeOgGoals = parseGoalCol(r['home_own_goal'], false, true)
|
||||
// away_own_goal = away player scored OG → goal credited to HOME team
|
||||
const awayOgGoals = parseGoalCol(r['away_own_goal'], false, true)
|
||||
|
||||
// Determine FT vs ET score split from goal minutes
|
||||
const allGoals = [...homeGoals, ...awayGoals, ...homePenGoals, ...awayPenGoals]
|
||||
const hasEt = allGoals.some(g => g.minute !== null && g.minute > 90)
|
||||
|
||||
let scoreFtHome: number | null, scoreFtAway: number | null
|
||||
let scoreEtHome: number | null = null, scoreEtAway: number | null = null
|
||||
|
||||
if (hasEt) {
|
||||
// Compute FT from goals in minutes 1–90
|
||||
const ftGoalCount = (goals: GoalEntry[]) =>
|
||||
goals.filter(g => g.minute === null || g.minute <= 90).length
|
||||
scoreFtHome = ftGoalCount(homeGoals) + ftGoalCount(homePenGoals) + ftGoalCount(awayOgGoals)
|
||||
scoreFtAway = ftGoalCount(awayGoals) + ftGoalCount(awayPenGoals) + ftGoalCount(homeOgGoals)
|
||||
scoreEtHome = homeScore
|
||||
scoreEtAway = awayScore
|
||||
} else {
|
||||
scoreFtHome = homeScore
|
||||
scoreFtAway = awayScore
|
||||
for (const year of YEARS) {
|
||||
const yearDir = path.join(WC_DIR, String(year))
|
||||
const mainData = readJson<RawData>(path.join(yearDir, 'worldcup.json'))
|
||||
if (!mainData?.matches) {
|
||||
console.log(` ${year}: no data file, skipping`)
|
||||
continue
|
||||
}
|
||||
|
||||
const [matchRow] = await db.execute(sql`
|
||||
INSERT INTO matches (
|
||||
tournament_year, round, date, team1_id, team2_id,
|
||||
score_ft_home, score_ft_away, score_et_home, score_et_away,
|
||||
score_p_home, score_p_away, is_quali_playoff
|
||||
) VALUES (
|
||||
${year}, ${r['Round'] || 'Unknown'}, ${dateStr},
|
||||
${t1Id}, ${t2Id},
|
||||
${scoreFtHome}, ${scoreFtAway}, ${scoreEtHome}, ${scoreEtAway},
|
||||
${homePen}, ${awayPen}, false
|
||||
)
|
||||
ON CONFLICT (tournament_year, team1_id, team2_id, date, is_quali_playoff) DO UPDATE SET
|
||||
round = EXCLUDED.round,
|
||||
score_ft_home = EXCLUDED.score_ft_home,
|
||||
score_ft_away = EXCLUDED.score_ft_away,
|
||||
score_et_home = EXCLUDED.score_et_home,
|
||||
score_et_away = EXCLUDED.score_et_away,
|
||||
score_p_home = EXCLUDED.score_p_home,
|
||||
score_p_away = EXCLUDED.score_p_away
|
||||
RETURNING id
|
||||
`)
|
||||
const matchId = (matchRow as { id: number }).id
|
||||
let matchCount = 0, goalCount = 0
|
||||
|
||||
await db.execute(sql`DELETE FROM goals WHERE match_id = ${matchId}`)
|
||||
// Stadiums
|
||||
const stadiumsData = readJson<RawStadiums>(path.join(yearDir, 'worldcup.stadiums.json'))
|
||||
if (stadiumsData?.stadiums) {
|
||||
for (const s of stadiumsData.stadiums) {
|
||||
await db.execute(sql`
|
||||
INSERT INTO stadiums (tournament_year, name, city)
|
||||
VALUES (${year}, ${s.name}, ${s.city ?? null})
|
||||
ON CONFLICT DO NOTHING
|
||||
`)
|
||||
}
|
||||
}
|
||||
|
||||
// home team goals (+ away player own goals that benefit home)
|
||||
for (const g of [...homeGoals, ...homePenGoals, ...awayOgGoals]) {
|
||||
await db.execute(sql`
|
||||
INSERT INTO goals (match_id, team_id, player_name, minute, minute_offset, is_penalty, is_own_goal)
|
||||
VALUES (${matchId}, ${t1Id}, ${g.name}, ${g.minute}, ${g.offset}, ${g.isPenalty}, ${g.isOwnGoal})
|
||||
// Matches and goals
|
||||
for (const m of mainData.matches) {
|
||||
const t1Id = await upsertTeam(m.team1)
|
||||
const t2Id = await upsertTeam(m.team2)
|
||||
const score = parseScore(m.score)
|
||||
|
||||
const [matchRow] = await db.execute(sql`
|
||||
INSERT INTO matches (
|
||||
tournament_year, round, group_name, date, time_local,
|
||||
team1_id, team2_id,
|
||||
score_ft_home, score_ft_away,
|
||||
score_ht_home, score_ht_away,
|
||||
score_et_home, score_et_away,
|
||||
score_p_home, score_p_away,
|
||||
is_quali_playoff
|
||||
) VALUES (
|
||||
${year}, ${m.round ?? 'Unknown'}, ${m.group ?? null},
|
||||
${m.date ?? null}, ${m.time ?? null},
|
||||
${t1Id}, ${t2Id},
|
||||
${score.ft?.[0] ?? null}, ${score.ft?.[1] ?? null},
|
||||
${score.ht?.[0] ?? null}, ${score.ht?.[1] ?? null},
|
||||
${score.et?.[0] ?? null}, ${score.et?.[1] ?? null},
|
||||
${score.p?.[0] ?? null}, ${score.p?.[1] ?? null},
|
||||
false
|
||||
)
|
||||
ON CONFLICT (tournament_year, team1_id, team2_id, date, is_quali_playoff) DO UPDATE SET
|
||||
round = EXCLUDED.round,
|
||||
group_name = COALESCE(EXCLUDED.group_name, matches.group_name),
|
||||
time_local = COALESCE(EXCLUDED.time_local, matches.time_local),
|
||||
score_ft_home = COALESCE(EXCLUDED.score_ft_home, matches.score_ft_home),
|
||||
score_ft_away = COALESCE(EXCLUDED.score_ft_away, matches.score_ft_away),
|
||||
score_ht_home = COALESCE(EXCLUDED.score_ht_home, matches.score_ht_home),
|
||||
score_ht_away = COALESCE(EXCLUDED.score_ht_away, matches.score_ht_away),
|
||||
score_et_home = COALESCE(EXCLUDED.score_et_home, matches.score_et_home),
|
||||
score_et_away = COALESCE(EXCLUDED.score_et_away, matches.score_et_away),
|
||||
score_p_home = COALESCE(EXCLUDED.score_p_home, matches.score_p_home),
|
||||
score_p_away = COALESCE(EXCLUDED.score_p_away, matches.score_p_away)
|
||||
RETURNING id
|
||||
`)
|
||||
totalGoals++
|
||||
const matchId = (matchRow as { id: number }).id
|
||||
|
||||
// Goals (delete + re-insert)
|
||||
await db.execute(sql`DELETE FROM goals WHERE match_id = ${matchId}`)
|
||||
|
||||
for (const [rawGoals, teamId, ogTeamId] of [
|
||||
[m.goals1 ?? [], t1Id, t2Id],
|
||||
[m.goals2 ?? [], t2Id, t1Id],
|
||||
] as [RawGoal[], number, number][]) {
|
||||
for (const g of rawGoals) {
|
||||
if (!g.name) continue
|
||||
const minute = g.minute != null ? parseInt(String(g.minute)) : null
|
||||
const actualTeamId = g.owngoal ? ogTeamId : teamId
|
||||
await db.execute(sql`
|
||||
INSERT INTO goals (match_id, team_id, player_name, minute, minute_offset, is_penalty, is_own_goal)
|
||||
VALUES (${matchId}, ${actualTeamId}, ${g.name}, ${!minute || isNaN(minute) ? null : minute},
|
||||
${g.offset ?? 0}, ${g.penalty ?? false}, ${g.owngoal ?? false})
|
||||
`)
|
||||
goalCount++
|
||||
}
|
||||
}
|
||||
|
||||
matchCount++
|
||||
}
|
||||
// away team goals (+ home player own goals that benefit away)
|
||||
for (const g of [...awayGoals, ...awayPenGoals, ...homeOgGoals]) {
|
||||
await db.execute(sql`
|
||||
INSERT INTO goals (match_id, team_id, player_name, minute, minute_offset, is_penalty, is_own_goal)
|
||||
VALUES (${matchId}, ${t2Id}, ${g.name}, ${g.minute}, ${g.offset}, ${g.isPenalty}, ${g.isOwnGoal})
|
||||
`)
|
||||
totalGoals++
|
||||
|
||||
// Squads
|
||||
const squadsData = readJson<RawSquad[]>(path.join(yearDir, 'worldcup.squads.json'))
|
||||
if (squadsData && Array.isArray(squadsData)) {
|
||||
for (const sq of squadsData) {
|
||||
const teamId = await upsertTeam(sq.name)
|
||||
for (const p of sq.players) {
|
||||
if (!p.name) continue
|
||||
const dob = p.date_of_birth ? p.date_of_birth.replace(/\s/g, '') : null
|
||||
await db.execute(sql`
|
||||
INSERT INTO squads (tournament_year, team_id, player_name, shirt_number, position, date_of_birth)
|
||||
VALUES (${year}, ${teamId}, ${p.name}, ${p.number ?? null},
|
||||
${p.pos ?? null}, ${dob})
|
||||
ON CONFLICT (tournament_year, team_id, shirt_number) DO UPDATE SET
|
||||
player_name = EXCLUDED.player_name,
|
||||
position = EXCLUDED.position,
|
||||
date_of_birth = EXCLUDED.date_of_birth
|
||||
`)
|
||||
}
|
||||
}
|
||||
}
|
||||
totalMatches++
|
||||
|
||||
console.log(` ${year}: ${matchCount} matches, ${goalCount} goals`)
|
||||
totalMatches += matchCount
|
||||
totalGoals += goalCount
|
||||
}
|
||||
|
||||
// 3. Update tournament aggregates
|
||||
// 3. Group standings (computed from match results)
|
||||
console.log('Computing group standings...')
|
||||
await db.execute(sql`
|
||||
DELETE FROM group_standings WHERE tournament_year < 2026
|
||||
`)
|
||||
await db.execute(sql`
|
||||
INSERT INTO group_standings (tournament_year, group_name, team_id, played, won, drawn, lost,
|
||||
goals_for, goals_against, goal_diff, pts)
|
||||
WITH match_results AS (
|
||||
SELECT tournament_year, group_name, team1_id AS team_id, score_ft_home AS gf, score_ft_away AS ga
|
||||
FROM matches WHERE tournament_year < 2026 AND group_name IS NOT NULL
|
||||
AND is_quali_playoff = false AND score_ft_home IS NOT NULL
|
||||
UNION ALL
|
||||
SELECT tournament_year, group_name, team2_id, score_ft_away, score_ft_home
|
||||
FROM matches WHERE tournament_year < 2026 AND group_name IS NOT NULL
|
||||
AND is_quali_playoff = false AND score_ft_home IS NOT NULL
|
||||
)
|
||||
SELECT tournament_year, group_name, team_id,
|
||||
COUNT(*)::int,
|
||||
SUM(CASE WHEN gf > ga THEN 1 ELSE 0 END)::int,
|
||||
SUM(CASE WHEN gf = ga THEN 1 ELSE 0 END)::int,
|
||||
SUM(CASE WHEN gf < ga THEN 1 ELSE 0 END)::int,
|
||||
SUM(gf)::int, SUM(ga)::int, SUM(gf - ga)::int,
|
||||
SUM(CASE WHEN gf > ga THEN 3 WHEN gf = ga THEN 1 ELSE 0 END)::int
|
||||
FROM match_results
|
||||
GROUP BY tournament_year, group_name, team_id
|
||||
ON CONFLICT (tournament_year, group_name, team_id) DO UPDATE SET
|
||||
played = EXCLUDED.played, won = EXCLUDED.won, drawn = EXCLUDED.drawn,
|
||||
lost = EXCLUDED.lost, goals_for = EXCLUDED.goals_for,
|
||||
goals_against = EXCLUDED.goals_against, goal_diff = EXCLUDED.goal_diff,
|
||||
pts = EXCLUDED.pts
|
||||
`)
|
||||
|
||||
// 4. Tournament aggregates
|
||||
await db.execute(sql`
|
||||
UPDATE tournaments t SET
|
||||
matches_count = (
|
||||
SELECT COUNT(*)::int FROM matches WHERE tournament_year = t.year AND is_quali_playoff = false
|
||||
),
|
||||
total_goals = (
|
||||
SELECT COUNT(g.id)::int
|
||||
FROM goals g JOIN matches m ON g.match_id = m.id
|
||||
WHERE m.tournament_year = t.year AND m.is_quali_playoff = false
|
||||
),
|
||||
matches_count = (
|
||||
SELECT COUNT(*)::int FROM matches WHERE tournament_year = t.year AND is_quali_playoff = false
|
||||
),
|
||||
avg_goals_per_game = (
|
||||
SELECT ROUND(COUNT(g.id)::numeric / NULLIF(COUNT(DISTINCT m.id), 0), 2)
|
||||
FROM goals g JOIN matches m ON g.match_id = m.id
|
||||
WHERE m.tournament_year = t.year AND m.is_quali_playoff = false
|
||||
AND m.score_ft_home IS NOT NULL
|
||||
)
|
||||
WHERE t.year < 2026
|
||||
`)
|
||||
|
||||
console.log(`✅ Seed complete: ${totalMatches} matches, ${totalGoals} goals (1930–2022)`)
|
||||
console.log(`\n✅ Seed complete: ${totalMatches} matches, ${totalGoals} goals (1930–2022)`)
|
||||
await client.end()
|
||||
}
|
||||
|
||||
|
||||
Reference in New Issue
Block a user