wip

Scripts/sportstime_parser/scrapers/nwsl.py

@@ -1,6 +1,6 @@
 """NWSL scraper implementation with multi-source fallback."""
 
-from datetime import datetime, date
+from datetime import datetime, date, timedelta
 from typing import Optional
 
 from .base import BaseScraper, RawGameData, ScrapeResult
@@ -73,33 +73,30 @@ class NWSLScraper(BaseScraper):
             raise ValueError(f"Unknown source: {source}")
 
     def _scrape_espn(self) -> list[RawGameData]:
-        """Scrape games from ESPN API."""
-        all_games: list[RawGameData] = []
+        """Scrape games from ESPN API using date range query."""
+        # Build date range for entire season (March-November)
+        season_months = self._get_season_months()
+        start_year, start_month = season_months[0]
+        end_year, end_month = season_months[-1]
 
-        for year, month in self._get_season_months():
-            # Get number of days in month
-            if month == 12:
-                next_month = date(year + 1, 1, 1)
-            else:
-                next_month = date(year, month + 1, 1)
+        # Get last day of end month
+        if end_month == 12:
+            end_date = date(end_year + 1, 1, 1) - timedelta(days=1)
+        else:
+            end_date = date(end_year, end_month + 1, 1) - timedelta(days=1)
 
-            days_in_month = (next_month - date(year, month, 1)).days
+        start_date = date(start_year, start_month, 1)
+        date_range = f"{start_date.strftime('%Y%m%d')}-{end_date.strftime('%Y%m%d')}"
 
-            for day in range(1, days_in_month + 1):
-                try:
-                    game_date = date(year, month, day)
-                    date_str = game_date.strftime("%Y%m%d")
-                    url = self._get_source_url("espn", date=date_str)
+        url = f"https://site.api.espn.com/apis/site/v2/sports/soccer/usa.nwsl/scoreboard?limit=1000&dates={date_range}"
+        self._logger.info(f"Fetching NWSL schedule: {date_range}")
 
-                    data = self.session.get_json(url)
-                    games = self._parse_espn_response(data, url)
-                    all_games.extend(games)
-
-                except Exception as e:
-                    self._logger.debug(f"ESPN error for {year}-{month}-{day}: {e}")
-                    continue
-
-        return all_games
+        try:
+            data = self.session.get_json(url)
+            return self._parse_espn_response(data, url)
+        except Exception as e:
+            self._logger.error(f"ESPN error: {e}")
+            return []
 
     def _parse_espn_response(
         self,