""" MSport Nigeria scraper. API endpoints and required headers discovered via Playwright network interception. Key headers: operid=2, clientid=WEB, apilevel=2, platform=WEB """ import logging from datetime import datetime from typing import List, Optional from scrapers.base import BaseScraper, normalize_cs_score from core.models import Event, Outcome logger = logging.getLogger(__name__) BASE_URL = "https://www.msport.com/api/ng/facts-center/query/frontend" LIVE_ENDPOINT = f"{BASE_URL}/live-matches" SPORT_IDS = { 'football': 'sr:sport:1', 'basketball': 'sr:sport:2', 'tennis': 'sr:sport:5', } # Market IDs we care about for arb ARBI_MARKETS = { 'football': {1: '1X2', 10: 'Double Chance', 11: 'Asian Handicap', 18: 'Over/Under 2.5', 26: 'Draw No Bet', 29: 'BTTS', 47: 'Correct Score'}, 'basketball': {219: 'Home/Away'}, 'tennis': {186: 'Home/Away'}, } # MSport outcome ID → name for football markets OUTCOME_MAP = { '1': 'Home', '2': 'Draw', '3': 'Away', # 1X2 '9': '1X', '10': '12', '11': 'X2', # Double Chance '12': 'Over', '13': 'Under', # O/U '60': 'Home', '61': 'Away', # Draw No Bet '74': 'Yes', '76': 'No', # BTTS '714': 'Home', '715': 'Away', # Asian Handicap } # Per-market outcome ID → label (overrides OUTCOME_MAP) MARKET_OUTCOME_MAP = { 219: {'4': 'Home', '5': 'Away'}, # Basketball moneyline 186: {'4': 'Home', '5': 'Away'}, # Tennis winner } def _parse_ah_line(specifier: str): """Extract and normalise AH line from a hcp= specifier string. Returns '+0.5', '-0.5', '0' etc. or None for quarter-ball lines.""" for key in ('hcp=', 'handicap='): if key in specifier: try: raw = specifier.split(key)[1].split('&')[0].strip() val = float(raw) if abs(val * 4) % 2 != 0: # reject quarter-ball (0.25, 0.75, ...) return None return f'+{val:g}' if val > 0 else f'{val:g}' except (ValueError, IndexError): return None return None class MSportScraper(BaseScraper): def __init__(self): super().__init__('MSport') # Replace session headers with exact set that works self.session.headers = { 'User-Agent': 'Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/120.0.0.0 Safari/537.36', 'Accept': 'application/json, text/plain, */*', 'Referer': 'https://www.msport.com/ng/web', 'clientid': 'WEB', 'operid': '2', 'apilevel': '2', 'platform': 'WEB', } def get_events(self, sport: str) -> List[Event]: sport_id = SPORT_IDS.get(sport) if not sport_id: return [] events: List[Event] = [] page = 1 while True: try: raw_events, has_more = self._fetch_page(sport_id, page) for raw in raw_events: if raw.get('betStop') or raw.get('active') == 0: continue events.extend(self._parse(raw, sport)) if not has_more or page >= 10: # cap at 1000 events break page += 1 except Exception as ex: logger.error(f"[MSport] {sport} page {page} error: {ex}") break return events def get_live_events(self, sport: str) -> List[Event]: sport_id = SPORT_IDS.get(sport) if not sport_id: return [] events: List[Event] = [] try: resp = self.session.get( LIVE_ENDPOINT, params={'sportId': sport_id, 'pageSize': 200, 'pageNum': 1}, timeout=15, ) resp.raise_for_status() for raw in resp.json().get('data', {}).get('events', []): events.extend(self._parse(raw, sport, live=True)) except Exception as ex: logger.error(f'[MSport] live {sport} fetch error: {ex}') return events def _fetch_page(self, sport_id: str, page: int): url = f"{BASE_URL}/all-matches/next-7-days" params = { 'sportId': sport_id, 'pageSize': 100, 'pageNum': page, } resp = self.session.get(url, params=params, timeout=15) resp.raise_for_status() data = resp.json().get('data', {}) raw_events = data.get('events', []) # has_more: if we got a full page, there might be more has_more = len(raw_events) == 100 return raw_events, has_more def _parse(self, raw: dict, sport: str, live: bool = False) -> List[Event]: home = (raw.get('homeTeam') or '').strip() away = (raw.get('awayTeam') or '').strip() if not home or not away: return [] league = (raw.get('tournament') or '').strip() event_id = raw.get('eventId', '') sport_s = raw.get('sport') or 'Soccer' category = (raw.get('category') or '').replace(' ', '_') tourney = league.replace(' ', '_') home_s = home.replace(' ', '_') away_s = away.replace(' ', '_') if event_id: if live: event_url = ( f'https://www.msport.com/ng/live/{sport_s}' f'/{category}_{tourney}' f'/{home_s}_vs_{away_s}' f'/{event_id}' ) elif category and tourney: event_url = ( f'https://www.msport.com/ng/web/sports/{sport_s}' f'/{category}_{tourney}' f'/{home_s}_vs_{away_s}' f'/{event_id}' ) else: event_url = None else: event_url = None ts = raw.get('startTime') or raw.get('estimateStartTime') try: starts_at = datetime.utcfromtimestamp(ts / 1000) if ts else None except Exception: starts_at = None arb_market_ids = ARBI_MARKETS.get(sport, {}) result: List[Event] = [] for market in raw.get('markets', []): market_id = market.get('id') if market_id not in arb_market_ids: continue market_name = arb_market_ids[market_id] mkt_oc_map = MARKET_OUTCOME_MAP.get(market_id) ou_line = '' ah_line = '' if market_id == 47: pass # Correct Score — no line detection needed elif market_id == 18: # Detect which line this market covers from the first active outcome description first_desc = next( (o.get('description', '') for o in market.get('outcomes', []) if o.get('isActive', 1)), '' ) for _lv in ('0.5', '1.5', '2.5', '3.5', '4.5'): if _lv in first_desc: ou_line = _lv market_name = f'Over/Under {_lv}' break if not ou_line: continue elif market_id == 11: # Asian Handicap — line from hcp= specifier specifier = market.get('specifier', '') ah_line = _parse_ah_line(specifier) if ah_line is None: continue market_name = f'Asian Handicap {ah_line}' # Skip market if suspended/locked at the market level if (market.get('betStop') or market.get('active') == 0 or market.get('isActive') == 0 or str(market.get('status', '')).lower() in ('suspended', 'closed')): continue outcomes = [] for o in market.get('outcomes', []): if not o.get('isActive', 1): continue oid = str(o.get('id', '')) desc = o.get('description', '') if market_id == 47: # Correct Score — description IS the score string name = normalize_cs_score(desc) if not name: continue elif market_id == 18: if ou_line not in desc: continue name = 'Over' if desc.startswith('Over') else 'Under' elif market_id == 11: # AH: use outcome ID map (714=Home, 715=Away) name = OUTCOME_MAP.get(oid) if name is None: continue elif mkt_oc_map is not None: name = mkt_oc_map.get(oid) if name is None: continue else: name = OUTCOME_MAP.get(oid, desc) try: odds = float(o.get('odds', 0)) except (TypeError, ValueError): continue if odds > 1.0: outcomes.append(Outcome(name=name, odds=odds, bookmaker='MSport', event_url=event_url)) if len(outcomes) >= 2: _line_sfx = ou_line or ah_line or '' result.append(Event( event_id=f"ms_{raw.get('eventId', '')}_{market_id}{'_' + _line_sfx if _line_sfx else ''}", bookmaker='MSport', sport=sport, home_team=home, away_team=away, market=market_name, outcomes=outcomes, starts_at=starts_at, league=league, )) return result