52d445bca4
Complete Python package for scraping, normalizing, and uploading sports schedule data to CloudKit. Includes: - Multi-source scrapers for NBA, MLB, NFL, NHL, MLS, WNBA, NWSL - Canonical ID system for teams, stadiums, and games - Fuzzy matching with manual alias support - CloudKit uploader with batch operations and deduplication - Comprehensive test suite with fixtures - WNBA abbreviation aliases for improved team resolution - Alias validation script to detect orphan references All 5 phases of data remediation plan completed: - Phase 1: Alias fixes (team/stadium alias additions) - Phase 2: NHL stadium coordinate fixes - Phase 3: Re-scrape validation - Phase 4: iOS bundle update - Phase 5: Code quality improvements (WNBA aliases) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
33 lines
672 B
Python
33 lines
672 B
Python
"""Validators for scraped data."""
|
|
|
|
from .report import (
|
|
ValidationReport,
|
|
ValidationSummary,
|
|
generate_report,
|
|
detect_duplicate_games,
|
|
validate_games,
|
|
)
|
|
|
|
from .schema import (
|
|
SchemaValidationError,
|
|
validate_canonical_stadium,
|
|
validate_canonical_team,
|
|
validate_canonical_game,
|
|
validate_and_raise,
|
|
validate_batch,
|
|
)
|
|
|
|
__all__ = [
|
|
"ValidationReport",
|
|
"ValidationSummary",
|
|
"generate_report",
|
|
"detect_duplicate_games",
|
|
"validate_games",
|
|
"SchemaValidationError",
|
|
"validate_canonical_stadium",
|
|
"validate_canonical_team",
|
|
"validate_canonical_game",
|
|
"validate_and_raise",
|
|
"validate_batch",
|
|
]
|