63acf7accb
Adds the full Django application layer on top of sportstime_parser: - core: Sport, Team, Stadium, Game models with aliases and league structure - scraper: orchestration engine, adapter, job management, Celery tasks - cloudkit: CloudKit sync client, sync state tracking, sync jobs - dashboard: staff dashboard for monitoring scrapers, sync, review queue - notifications: email reports for scrape/sync results - Docker setup for deployment (Dockerfile, docker-compose, entrypoint) Game exports now use game_datetime_utc (ISO 8601 UTC) instead of venue-local date+time strings, matching the canonical format used by the iOS app. Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
395 lines
13 KiB
Python
395 lines
13 KiB
Python
from django.db import models
|
|
from django.conf import settings
|
|
from simple_history.models import HistoricalRecords
|
|
|
|
|
|
class CloudKitConfiguration(models.Model):
|
|
"""
|
|
CloudKit configuration for syncing.
|
|
"""
|
|
ENVIRONMENT_CHOICES = [
|
|
('development', 'Development'),
|
|
('production', 'Production'),
|
|
]
|
|
|
|
name = models.CharField(
|
|
max_length=100,
|
|
unique=True,
|
|
help_text='Configuration name (e.g., "Production", "Development")'
|
|
)
|
|
environment = models.CharField(
|
|
max_length=20,
|
|
choices=ENVIRONMENT_CHOICES,
|
|
default='development'
|
|
)
|
|
container_id = models.CharField(
|
|
max_length=200,
|
|
default=settings.CLOUDKIT_CONTAINER,
|
|
help_text='CloudKit container ID (e.g., iCloud.com.sportstime.app)'
|
|
)
|
|
key_id = models.CharField(
|
|
max_length=200,
|
|
blank=True,
|
|
help_text='CloudKit API key ID'
|
|
)
|
|
private_key = models.TextField(
|
|
blank=True,
|
|
help_text='EC P-256 private key content (PEM format). Paste key here OR use path below.'
|
|
)
|
|
private_key_path = models.CharField(
|
|
max_length=500,
|
|
blank=True,
|
|
help_text='Path to EC P-256 private key file (alternative to pasting key above)'
|
|
)
|
|
is_active = models.BooleanField(
|
|
default=False,
|
|
help_text='Whether this configuration is active for syncing'
|
|
)
|
|
|
|
# Sync settings
|
|
batch_size = models.PositiveIntegerField(
|
|
default=200,
|
|
help_text='Maximum records per batch upload'
|
|
)
|
|
auto_sync_after_scrape = models.BooleanField(
|
|
default=False,
|
|
help_text='Automatically sync after scraper jobs complete'
|
|
)
|
|
|
|
# Metadata
|
|
created_at = models.DateTimeField(auto_now_add=True)
|
|
updated_at = models.DateTimeField(auto_now=True)
|
|
|
|
# Audit trail
|
|
history = HistoricalRecords()
|
|
|
|
class Meta:
|
|
verbose_name = 'CloudKit Configuration'
|
|
verbose_name_plural = 'CloudKit Configurations'
|
|
|
|
def __str__(self):
|
|
return f"{self.name} ({self.environment})"
|
|
|
|
def save(self, *args, **kwargs):
|
|
# Ensure only one active configuration
|
|
if self.is_active:
|
|
CloudKitConfiguration.objects.filter(is_active=True).exclude(pk=self.pk).update(is_active=False)
|
|
super().save(*args, **kwargs)
|
|
|
|
def get_client(self):
|
|
"""Create a CloudKitClient from this configuration."""
|
|
from cloudkit.client import CloudKitClient
|
|
return CloudKitClient(
|
|
container_id=self.container_id,
|
|
environment=self.environment,
|
|
key_id=self.key_id,
|
|
private_key=self.private_key,
|
|
private_key_path=self.private_key_path,
|
|
)
|
|
|
|
@classmethod
|
|
def get_active(cls):
|
|
"""Get the active CloudKit configuration."""
|
|
return cls.objects.filter(is_active=True).first()
|
|
|
|
|
|
class CloudKitSyncState(models.Model):
|
|
"""
|
|
Tracks sync state for individual records.
|
|
"""
|
|
RECORD_TYPE_CHOICES = [
|
|
('Sport', 'Sport'),
|
|
('Conference', 'Conference'),
|
|
('Division', 'Division'),
|
|
('Team', 'Team'),
|
|
('Stadium', 'Stadium'),
|
|
('TeamAlias', 'Team Alias'),
|
|
('StadiumAlias', 'Stadium Alias'),
|
|
('Game', 'Game'),
|
|
]
|
|
|
|
SYNC_STATUS_CHOICES = [
|
|
('pending', 'Pending Sync'),
|
|
('synced', 'Synced'),
|
|
('failed', 'Failed'),
|
|
('deleted', 'Deleted'),
|
|
]
|
|
|
|
record_type = models.CharField(
|
|
max_length=20,
|
|
choices=RECORD_TYPE_CHOICES
|
|
)
|
|
record_id = models.CharField(
|
|
max_length=100,
|
|
help_text='Local record ID (canonical ID)'
|
|
)
|
|
cloudkit_record_name = models.CharField(
|
|
max_length=200,
|
|
blank=True,
|
|
help_text='CloudKit record name (may differ from local ID)'
|
|
)
|
|
local_hash = models.CharField(
|
|
max_length=64,
|
|
blank=True,
|
|
help_text='Hash of local record data for change detection'
|
|
)
|
|
remote_change_tag = models.CharField(
|
|
max_length=200,
|
|
blank=True,
|
|
help_text='CloudKit change tag for conflict detection'
|
|
)
|
|
sync_status = models.CharField(
|
|
max_length=20,
|
|
choices=SYNC_STATUS_CHOICES,
|
|
default='pending'
|
|
)
|
|
last_synced = models.DateTimeField(
|
|
null=True,
|
|
blank=True
|
|
)
|
|
last_error = models.TextField(
|
|
blank=True,
|
|
help_text='Last sync error message'
|
|
)
|
|
retry_count = models.PositiveSmallIntegerField(
|
|
default=0
|
|
)
|
|
|
|
# Metadata
|
|
created_at = models.DateTimeField(auto_now_add=True)
|
|
updated_at = models.DateTimeField(auto_now=True)
|
|
|
|
class Meta:
|
|
ordering = ['-updated_at']
|
|
unique_together = ['record_type', 'record_id']
|
|
verbose_name = 'CloudKit Sync State'
|
|
verbose_name_plural = 'CloudKit Sync States'
|
|
indexes = [
|
|
models.Index(fields=['sync_status', 'record_type']),
|
|
models.Index(fields=['record_type', 'last_synced']),
|
|
]
|
|
|
|
def __str__(self):
|
|
return f"{self.record_type}:{self.record_id} ({self.sync_status})"
|
|
|
|
def mark_synced(self, change_tag=''):
|
|
"""Mark record as successfully synced."""
|
|
from django.utils import timezone
|
|
self.sync_status = 'synced'
|
|
self.remote_change_tag = change_tag
|
|
self.last_synced = timezone.now()
|
|
self.last_error = ''
|
|
self.retry_count = 0
|
|
self.save()
|
|
|
|
def mark_failed(self, error_message):
|
|
"""Mark record as failed to sync."""
|
|
self.sync_status = 'failed'
|
|
self.last_error = error_message
|
|
self.retry_count += 1
|
|
self.save()
|
|
|
|
def mark_pending(self, new_hash=''):
|
|
"""Mark record as pending sync (e.g., after local change)."""
|
|
self.sync_status = 'pending'
|
|
if new_hash:
|
|
self.local_hash = new_hash
|
|
self.save()
|
|
|
|
|
|
class CloudKitSyncJob(models.Model):
|
|
"""
|
|
Record of a CloudKit sync job execution.
|
|
"""
|
|
STATUS_CHOICES = [
|
|
('pending', 'Pending'),
|
|
('running', 'Running'),
|
|
('completed', 'Completed'),
|
|
('completed_with_errors', 'Completed with Errors'),
|
|
('failed', 'Failed'),
|
|
('cancelled', 'Cancelled'),
|
|
]
|
|
|
|
configuration = models.ForeignKey(
|
|
CloudKitConfiguration,
|
|
on_delete=models.CASCADE,
|
|
related_name='sync_jobs'
|
|
)
|
|
status = models.CharField(
|
|
max_length=25,
|
|
choices=STATUS_CHOICES,
|
|
default='pending'
|
|
)
|
|
triggered_by = models.CharField(
|
|
max_length=50,
|
|
default='manual',
|
|
help_text='How the sync was triggered'
|
|
)
|
|
|
|
# Timing
|
|
started_at = models.DateTimeField(null=True, blank=True)
|
|
finished_at = models.DateTimeField(null=True, blank=True)
|
|
|
|
# Results
|
|
records_synced = models.PositiveIntegerField(default=0)
|
|
records_created = models.PositiveIntegerField(default=0)
|
|
records_updated = models.PositiveIntegerField(default=0)
|
|
records_deleted = models.PositiveIntegerField(default=0)
|
|
records_failed = models.PositiveIntegerField(default=0)
|
|
|
|
# Filter (optional - sync specific records)
|
|
sport_filter = models.ForeignKey(
|
|
'core.Sport',
|
|
on_delete=models.SET_NULL,
|
|
null=True,
|
|
blank=True,
|
|
help_text='Only sync this sport (all if blank)'
|
|
)
|
|
record_type_filter = models.CharField(
|
|
max_length=20,
|
|
blank=True,
|
|
help_text='Only sync this record type (all if blank)'
|
|
)
|
|
|
|
# Error tracking
|
|
error_message = models.TextField(blank=True)
|
|
|
|
# Progress tracking
|
|
current_record_type = models.CharField(
|
|
max_length=20,
|
|
blank=True,
|
|
help_text='Currently syncing record type'
|
|
)
|
|
sports_total = models.PositiveIntegerField(default=0)
|
|
sports_synced = models.PositiveIntegerField(default=0)
|
|
sports_failed = models.PositiveIntegerField(default=0)
|
|
teams_total = models.PositiveIntegerField(default=0)
|
|
teams_synced = models.PositiveIntegerField(default=0)
|
|
teams_failed = models.PositiveIntegerField(default=0)
|
|
stadiums_total = models.PositiveIntegerField(default=0)
|
|
stadiums_synced = models.PositiveIntegerField(default=0)
|
|
stadiums_failed = models.PositiveIntegerField(default=0)
|
|
conferences_total = models.PositiveIntegerField(default=0)
|
|
conferences_synced = models.PositiveIntegerField(default=0)
|
|
conferences_failed = models.PositiveIntegerField(default=0)
|
|
divisions_total = models.PositiveIntegerField(default=0)
|
|
divisions_synced = models.PositiveIntegerField(default=0)
|
|
divisions_failed = models.PositiveIntegerField(default=0)
|
|
team_aliases_total = models.PositiveIntegerField(default=0)
|
|
team_aliases_synced = models.PositiveIntegerField(default=0)
|
|
team_aliases_failed = models.PositiveIntegerField(default=0)
|
|
stadium_aliases_total = models.PositiveIntegerField(default=0)
|
|
stadium_aliases_synced = models.PositiveIntegerField(default=0)
|
|
stadium_aliases_failed = models.PositiveIntegerField(default=0)
|
|
games_total = models.PositiveIntegerField(default=0)
|
|
games_synced = models.PositiveIntegerField(default=0)
|
|
games_failed = models.PositiveIntegerField(default=0)
|
|
|
|
# Celery task ID
|
|
celery_task_id = models.CharField(
|
|
max_length=255,
|
|
blank=True
|
|
)
|
|
|
|
# Metadata
|
|
created_at = models.DateTimeField(auto_now_add=True)
|
|
updated_at = models.DateTimeField(auto_now=True)
|
|
|
|
class Meta:
|
|
ordering = ['-created_at']
|
|
verbose_name = 'CloudKit Sync Job'
|
|
verbose_name_plural = 'CloudKit Sync Jobs'
|
|
|
|
def __str__(self):
|
|
return f"Sync {self.configuration.name} - {self.created_at.strftime('%Y-%m-%d %H:%M')}"
|
|
|
|
@property
|
|
def duration(self):
|
|
if self.started_at and self.finished_at:
|
|
return self.finished_at - self.started_at
|
|
return None
|
|
|
|
@property
|
|
def duration_display(self):
|
|
duration = self.duration
|
|
if duration:
|
|
total_seconds = int(duration.total_seconds())
|
|
minutes, seconds = divmod(total_seconds, 60)
|
|
if minutes > 0:
|
|
return f"{minutes}m {seconds}s"
|
|
return f"{seconds}s"
|
|
return '-'
|
|
|
|
def get_progress(self):
|
|
"""Get progress data for API/display."""
|
|
total = (self.sports_total + self.conferences_total + self.divisions_total
|
|
+ self.teams_total + self.stadiums_total
|
|
+ self.team_aliases_total + self.stadium_aliases_total
|
|
+ self.games_total)
|
|
synced = (self.sports_synced + self.conferences_synced + self.divisions_synced
|
|
+ self.teams_synced + self.stadiums_synced
|
|
+ self.team_aliases_synced + self.stadium_aliases_synced
|
|
+ self.games_synced)
|
|
failed = (self.sports_failed + self.conferences_failed + self.divisions_failed
|
|
+ self.teams_failed + self.stadiums_failed
|
|
+ self.team_aliases_failed + self.stadium_aliases_failed
|
|
+ self.games_failed)
|
|
|
|
return {
|
|
'status': self.status,
|
|
'current_type': self.current_record_type,
|
|
'total': total,
|
|
'synced': synced,
|
|
'failed': failed,
|
|
'remaining': total - synced - failed,
|
|
'percent': round((synced + failed) / total * 100) if total > 0 else 0,
|
|
'sports': {
|
|
'total': self.sports_total,
|
|
'synced': self.sports_synced,
|
|
'failed': self.sports_failed,
|
|
'remaining': self.sports_total - self.sports_synced - self.sports_failed,
|
|
},
|
|
'conferences': {
|
|
'total': self.conferences_total,
|
|
'synced': self.conferences_synced,
|
|
'failed': self.conferences_failed,
|
|
'remaining': self.conferences_total - self.conferences_synced - self.conferences_failed,
|
|
},
|
|
'divisions': {
|
|
'total': self.divisions_total,
|
|
'synced': self.divisions_synced,
|
|
'failed': self.divisions_failed,
|
|
'remaining': self.divisions_total - self.divisions_synced - self.divisions_failed,
|
|
},
|
|
'teams': {
|
|
'total': self.teams_total,
|
|
'synced': self.teams_synced,
|
|
'failed': self.teams_failed,
|
|
'remaining': self.teams_total - self.teams_synced - self.teams_failed,
|
|
},
|
|
'stadiums': {
|
|
'total': self.stadiums_total,
|
|
'synced': self.stadiums_synced,
|
|
'failed': self.stadiums_failed,
|
|
'remaining': self.stadiums_total - self.stadiums_synced - self.stadiums_failed,
|
|
},
|
|
'team_aliases': {
|
|
'total': self.team_aliases_total,
|
|
'synced': self.team_aliases_synced,
|
|
'failed': self.team_aliases_failed,
|
|
'remaining': self.team_aliases_total - self.team_aliases_synced - self.team_aliases_failed,
|
|
},
|
|
'stadium_aliases': {
|
|
'total': self.stadium_aliases_total,
|
|
'synced': self.stadium_aliases_synced,
|
|
'failed': self.stadium_aliases_failed,
|
|
'remaining': self.stadium_aliases_total - self.stadium_aliases_synced - self.stadium_aliases_failed,
|
|
},
|
|
'games': {
|
|
'total': self.games_total,
|
|
'synced': self.games_synced,
|
|
'failed': self.games_failed,
|
|
'remaining': self.games_total - self.games_synced - self.games_failed,
|
|
},
|
|
}
|