Caching Strategy

import tif1

# Get configuration singleton
config = tif1.get_config()

# Configure memory cache size (number of items)
config.memory_cache_size = 200  # Store 200 most recent items

# Disable memory cache (use SQLite only)
config.memory_cache_enabled = False

# Configure per-type cache sizes
config.memory_cache_laps_size = 100
config.memory_cache_telemetry_size = 50
config.memory_cache_weather_size = 20

# Set memory cache size
export TIF1_MEMORY_CACHE_SIZE=200

# Disable memory cache
export TIF1_MEMORY_CACHE_ENABLED=false

# Set per-type sizes
export TIF1_MEMORY_CACHE_LAPS_SIZE=100
export TIF1_MEMORY_CACHE_TELEMETRY_SIZE=50

# Key format: {year}_{event}_{session_type}_{data_type}_{backend}
# Examples:
"2025_monaco_race_laps_pandas"
"2025_bahrain_qualifying_telemetry_polars"
"2024_silverstone_practice_1_weather_pandas"

# Key generation algorithm
def generate_cache_key(
    year: int,
    event: str,
    session_type: str,
    data_type: str,
    backend: str = "pandas"
) -> str:
    """Generate deterministic cache key."""
    # Normalize event name (lowercase, remove spaces)
    event_normalized = event.lower().replace(" ", "_")
    session_normalized = session_type.lower().replace(" ", "_")

    return f"{year}_{event_normalized}_{session_normalized}_{data_type}_{backend}"

import tif1

# First access - cache miss, fetches from SQLite or CDN
session = tif1.get_session(2025, "Monaco", "Race")
laps1 = session.laps  # ~50ms (SQLite) or ~2s (CDN)

# Second access - cache hit, returns from memory
laps2 = session.laps  # ~0.5ms (memory)

# Third access - still cached
laps3 = session.laps  # ~0.5ms (memory)

import tif1

config = tif1.get_config()
config.memory_cache_size = 2  # Very small cache for demonstration

# Load 3 sessions - third will evict first
session1 = tif1.get_session(2025, "Monaco", "Race")
laps1 = session1.laps  # Cached in slot 1

session2 = tif1.get_session(2025, "Bahrain", "Race")
laps2 = session2.laps  # Cached in slot 2

session3 = tif1.get_session(2025, "Silverstone", "Race")
laps3 = session3.laps  # Cached in slot 2, evicts session1

# Accessing session1 again requires SQLite/CDN fetch
laps1_again = session1.laps  # Cache miss, ~50ms (SQLite)

import tif1

# Monitor memory cache usage
cache = tif1.get_cache()
mem_stats = cache.get_memory_stats()

print(f"Items cached: {mem_stats['items']}")
print(f"Memory used: {mem_stats['size_mb']:.2f} MB")
print(f"Hit rate: {mem_stats['hit_rate']:.1%}")
print(f"Evictions: {mem_stats['evictions']}")

# Clear memory cache (keeps SQLite cache)
cache.clear_memory()

# Manually evict specific item
cache.evict_memory("2025_monaco_race_laps_pandas")

# Linux/macOS
~/.tif1/cache/tif1_cache.db

# Windows
C:\Users\<username>\.tif1\cache\tif1_cache.db

# Custom location
export TIF1_CACHE_DIR="/custom/cache/path"

import tif1
from pathlib import Path

config = tif1.get_config()

# Set custom cache directory
config.cache_dir = "/custom/cache/path"

# Or use Path object
config.cache_dir = Path.home() / "my_tif1_cache"

# Enable/disable SQLite cache
config.cache_enabled = True

# Set cache file name
config.cache_filename = "custom_cache.db"

# Configure SQLite performance options
config.cache_page_size = 4096  # SQLite page size (bytes)
config.cache_cache_size = 10000  # SQLite cache size (pages)
config.cache_wal_autocheckpoint = 1000  # WAL checkpoint interval

# Set compression level (1-22, higher = better compression, slower)
config.cache_compression_level = 3  # Default: 3 (good balance)

# Cache directory
export TIF1_CACHE_DIR="/custom/cache/path"

# Enable/disable cache
export TIF1_CACHE_ENABLED=true

# Compression level
export TIF1_CACHE_COMPRESSION_LEVEL=3

# SQLite performance tuning
export TIF1_CACHE_PAGE_SIZE=4096
export TIF1_CACHE_CACHE_SIZE=10000

-- Main cache table
CREATE TABLE cache (
    key TEXT PRIMARY KEY,              -- Unique cache key
    value BLOB NOT NULL,               -- Compressed Parquet data
    created_at TIMESTAMP NOT NULL,     -- Creation timestamp
    accessed_at TIMESTAMP NOT NULL,    -- Last access timestamp
    access_count INTEGER DEFAULT 1,    -- Number of accesses
    size_bytes INTEGER NOT NULL,       -- Uncompressed size
    compressed_size INTEGER NOT NULL,  -- Compressed size
    data_type TEXT NOT NULL,           -- Type: laps, telemetry, weather, etc.
    backend TEXT NOT NULL,             -- Backend: pandas or polars
    schema_version INTEGER NOT NULL,   -- Schema version for invalidation
    etag TEXT,                         -- CDN ETag for freshness checks
    last_modified TEXT                 -- CDN Last-Modified header
);

-- Index for LRU eviction (find oldest accessed entries)
CREATE INDEX idx_accessed_at ON cache(accessed_at);

-- Index for data type queries
CREATE INDEX idx_data_type ON cache(data_type);

-- Index for size-based queries
CREATE INDEX idx_size ON cache(size_bytes);

-- Index for creation time (TTL expiration)
CREATE INDEX idx_created_at ON cache(created_at);

-- Metadata table for cache statistics
CREATE TABLE cache_metadata (
    key TEXT PRIMARY KEY,
    value TEXT NOT NULL
);

-- Insert cache version
INSERT INTO cache_metadata (key, value) VALUES ('version', '1.0');
INSERT INTO cache_metadata (key, value) VALUES ('created_at', datetime('now'));

import tif1

# First access - fetches from CDN, stores in cache
session = tif1.get_session(2025, "Monaco", "Race")
laps = session.laps  # ~2.5s (cold - network fetch)

# Second access - reads from SQLite cache
session2 = tif1.get_session(2025, "Monaco", "Race")
laps2 = session2.laps  # ~50ms (warm - disk read)

# Third access - reads from memory cache
laps3 = session2.laps  # ~1ms (hot - memory read)

# Access from different process - reads from SQLite
# (memory cache is process-specific)
# In another Python process:
session3 = tif1.get_session(2025, "Monaco", "Race")
laps3 = session3.laps  # ~50ms (warm - disk read)

import tif1

cache = tif1.get_cache()

# Get comprehensive cache statistics
stats = cache.get_stats()

print(f"Total entries: {stats['total_entries']}")
print(f"Total size (uncompressed): {stats['total_size_mb']:.2f} MB")
print(f"Total size (compressed): {stats['compressed_size_mb']:.2f} MB")
print(f"Compression ratio: {stats['compression_ratio']:.1f}:1")
print(f"Space saved: {stats['space_saved_mb']:.2f} MB ({stats['space_saved_pct']:.1f}%)")
print(f"Hit rate: {stats['hit_rate']:.1%}")
print(f"Miss rate: {stats['miss_rate']:.1%}")
print(f"Total hits: {stats['total_hits']}")
print(f"Total misses: {stats['total_misses']}")
print(f"Average access time: {stats['avg_access_ms']:.1f}ms")

# Get per-data-type statistics
type_stats = cache.get_stats_by_type()
for data_type, stats in type_stats.items():
    print(f"\n{data_type}:")
    print(f"  Entries: {stats['count']}")
    print(f"  Size: {stats['size_mb']:.2f} MB")
    print(f"  Avg size: {stats['avg_size_mb']:.2f} MB")
    print(f"  Hit rate: {stats['hit_rate']:.1%}")

import tif1

cache = tif1.get_cache()

# Clear all cache entries
cache.clear()
print("All cache cleared")

# Clear specific session
cache.clear_session(2025, "Monaco", "Race")
print("Monaco 2025 Race cleared")

# Clear specific data type
cache.clear_type("telemetry")
print("All telemetry data cleared")

# Clear old entries (older than 30 days)
removed = cache.clear_old(days=30)
print(f"Removed {removed} entries older than 30 days")

# Clear by size (remove largest entries first)
removed = cache.clear_largest(count=10)
print(f"Removed 10 largest entries, freed {removed:.2f} MB")

# Clear least recently used entries
removed = cache.clear_lru(count=50)
print(f"Removed 50 least recently used entries")

# Clear entries matching pattern
removed = cache.clear_pattern("2024_*_practice_*")
print(f"Removed {removed} practice session entries from 2024")

import tif1

cache = tif1.get_cache()

# List all cached sessions
sessions = cache.list_sessions()
for session in sessions:
    print(f"{session['year']} {session['event']} {session['session_type']}")
    print(f"  Size: {session['size_mb']:.2f} MB")
    print(f"  Accessed: {session['accessed_at']}")
    print(f"  Access count: {session['access_count']}")

# Find largest cache entries
largest = cache.get_largest(limit=10)
for entry in largest:
    print(f"{entry['key']}: {entry['size_mb']:.2f} MB")

# Find least recently used entries
lru = cache.get_lru(limit=10)
for entry in lru:
    print(f"{entry['key']}: last accessed {entry['accessed_at']}")

# Find entries by data type
telemetry_entries = cache.get_by_type("telemetry")
print(f"Found {len(telemetry_entries)} telemetry entries")

# Search cache by pattern
monaco_entries = cache.search("*monaco*")
print(f"Found {len(monaco_entries)} Monaco entries")

import tif1
from concurrent.futures import ThreadPoolExecutor

# Multiple threads can safely access cache
def load_session(year, event):
    session = tif1.get_session(year, event, "Race")
    return session.laps

# Concurrent access from multiple threads
with ThreadPoolExecutor(max_workers=4) as executor:
    futures = [
        executor.submit(load_session, 2025, "Monaco"),
        executor.submit(load_session, 2025, "Bahrain"),
        executor.submit(load_session, 2025, "Silverstone"),
        executor.submit(load_session, 2025, "Spa"),
    ]

    results = [f.result() for f in futures]
    print(f"Loaded {len(results)} sessions concurrently")

import tif1

config = tif1.get_config()

# Increase SQLite cache size (more memory, faster queries)
config.cache_cache_size = 20000  # 20000 pages × 4KB = 80MB

# Increase page size (better for large BLOBs)
config.cache_page_size = 8192  # 8KB pages

# Adjust WAL checkpoint interval
config.cache_wal_autocheckpoint = 2000  # Checkpoint every 2000 pages

# Enable memory-mapped I/O (faster on 64-bit systems)
config.cache_mmap_size = 268435456  # 256MB mmap

# Adjust compression level (1-22)
config.cache_compression_level = 1  # Faster compression, less savings
config.cache_compression_level = 9  # Slower compression, more savings

import tif1

# First access - cold cache (network fetch)
# Flow: Memory miss → SQLite miss → CDN fetch → Store in SQLite → Store in memory
session = tif1.get_session(2025, "Monaco", "Race")
laps = session.laps  # ~2.5s (cold)

# Second access - warm cache (SQLite hit)
# Flow: Memory miss → SQLite hit → Decompress → Store in memory
session2 = tif1.get_session(2025, "Monaco", "Race")
laps2 = session2.laps  # ~50ms (warm)

# Third access - hot cache (memory hit)
# Flow: Memory hit → Return immediately
laps3 = session2.laps  # ~1ms (hot)

# Access from different process - warm cache
# (Memory cache is process-specific, but SQLite is shared)
# In another Python process:
import tif1
session3 = tif1.get_session(2025, "Monaco", "Race")
laps3 = session3.laps  # ~50ms (warm - SQLite hit)

import tif1

cache = tif1.get_cache()

# Get comprehensive statistics
stats = cache.get_stats()

print("=== Cache Statistics ===")
print(f"Total entries: {stats['total_entries']}")
print(f"Total size (uncompressed): {stats['total_size_mb']:.2f} MB")
print(f"Total size (compressed): {stats['compressed_size_mb']:.2f} MB")
print(f"Compression ratio: {stats['compression_ratio']:.1f}:1")
print(f"Space saved: {stats['space_saved_mb']:.2f} MB ({stats['space_saved_pct']:.1f}%)")
print(f"\nPerformance:")
print(f"Hit rate: {stats['hit_rate']:.1%}")
print(f"Miss rate: {stats['miss_rate']:.1%}")
print(f"Total hits: {stats['total_hits']}")
print(f"Total misses: {stats['total_misses']}")
print(f"Average access time: {stats['avg_access_ms']:.1f}ms")
print(f"\nMemory Cache:")
print(f"Memory entries: {stats['memory_entries']}")
print(f"Memory size: {stats['memory_size_mb']:.2f} MB")
print(f"Memory hit rate: {stats['memory_hit_rate']:.1%}")

# Get per-data-type breakdown
print("\n=== By Data Type ===")
type_stats = cache.get_stats_by_type()
for data_type, type_stat in type_stats.items():
    print(f"\n{data_type.upper()}:")
    print(f"  Entries: {type_stat['count']}")
    print(f"  Total size: {type_stat['size_mb']:.2f} MB")
    print(f"  Avg size: {type_stat['avg_size_mb']:.2f} MB")
    print(f"  Hit rate: {type_stat['hit_rate']:.1%}")
    print(f"  Avg access time: {type_stat['avg_access_ms']:.1f}ms")

# Get per-year breakdown
print("\n=== By Year ===")
year_stats = cache.get_stats_by_year()
for year, year_stat in year_stats.items():
    print(f"\n{year}:")
    print(f"  Sessions: {year_stat['session_count']}")
    print(f"  Total size: {year_stat['size_mb']:.2f} MB")
    print(f"  Most accessed: {year_stat['most_accessed_event']}")

# Export statistics to JSON
import json
with open("cache_stats.json", "w") as f:
    json.dump(stats, f, indent=2, default=str)
print("\nStatistics exported to cache_stats.json")

import tif1

cache = tif1.get_cache()

# 1. Clear all cache (nuclear option)
cache.clear()
print("All cache cleared")

# 2. Clear specific session
cache.clear_session(2025, "Monaco", "Race")
print("Monaco 2025 Race cleared")

# 3. Clear specific event (all sessions)
cache.clear_event(2025, "Monaco")
print("All Monaco 2025 sessions cleared")

# 4. Clear specific year
cache.clear_year(2024)
print("All 2024 data cleared")

# 5. Clear specific data type across all sessions
cache.clear_type("telemetry")
print("All telemetry data cleared")

# 6. Clear old entries (time-based)
removed = cache.clear_old(days=30)
print(f"Removed {removed} entries older than 30 days")

# 7. Clear by size (remove largest entries)
freed_mb = cache.clear_largest(count=10)
print(f"Removed 10 largest entries, freed {freed_mb:.2f} MB")

# 8. Clear least recently used entries
removed = cache.clear_lru(count=50)
print(f"Removed 50 least recently used entries")

# 9. Clear entries matching pattern (glob-style)
removed = cache.clear_pattern("2024_*_practice_*")
print(f"Removed {removed} practice session entries from 2024")

# 10. Clear entries below access threshold
removed = cache.clear_low_access(min_access_count=2)
print(f"Removed {removed} entries accessed less than 2 times")

# 11. Clear memory cache only (keep SQLite)
cache.clear_memory()
print("Memory cache cleared, SQLite cache preserved")

# 12. Smart cleanup (remove old + low-access + large)
removed = cache.smart_cleanup(
    max_age_days=60,
    min_access_count=2,
    max_size_mb=5000
)
print(f"Smart cleanup removed {removed} entries")

import tif1

cache = tif1.get_cache()

# List all cached sessions
print("=== Cached Sessions ===")
sessions = cache.list_sessions()
for session in sessions:
    print(f"{session['year']} {session['event']} {session['session_type']}")
    print(f"  Size: {session['size_mb']:.2f} MB (compressed: {session['compressed_mb']:.2f} MB)")
    print(f"  Created: {session['created_at']}")
    print(f"  Last accessed: {session['accessed_at']}")
    print(f"  Access count: {session['access_count']}")
    print(f"  Data types: {', '.join(session['data_types'])}")
    print()

# Find largest cache entries
print("=== Largest Entries ===")
largest = cache.get_largest(limit=10)
for i, entry in enumerate(largest, 1):
    print(f"{i}. {entry['key']}")
    print(f"   Size: {entry['size_mb']:.2f} MB")
    print(f"   Compressed: {entry['compressed_mb']:.2f} MB")
    print(f"   Ratio: {entry['compression_ratio']:.1f}:1")

# Find least recently used entries
print("\n=== Least Recently Used ===")
lru = cache.get_lru(limit=10)
for i, entry in enumerate(lru, 1):
    print(f"{i}. {entry['key']}")
    print(f"   Last accessed: {entry['accessed_at']}")
    print(f"   Access count: {entry['access_count']}")
    print(f"   Age: {entry['age_days']:.1f} days")

# Find entries by data type
print("\n=== Telemetry Entries ===")
telemetry_entries = cache.get_by_type("telemetry")
print(f"Found {len(telemetry_entries)} telemetry entries")
total_size = sum(e['size_mb'] for e in telemetry_entries)
print(f"Total size: {total_size:.2f} MB")

# Search cache by pattern
print("\n=== Monaco Entries ===")
monaco_entries = cache.search("*monaco*")
for entry in monaco_entries:
    print(f"  {entry['key']} - {entry['size_mb']:.2f} MB")

# Check if specific entry exists
exists = cache.exists(2025, "Monaco", "Race", "laps")
print(f"\nMonaco 2025 Race laps cached: {exists}")

# Get entry metadata
if exists:
    metadata = cache.get_metadata(2025, "Monaco", "Race", "laps")
    print(f"  Created: {metadata['created_at']}")
    print(f"  Size: {metadata['size_mb']:.2f} MB")
    print(f"  Access count: {metadata['access_count']}")

import tif1

# Disable cache for specific request (always fetch fresh)
session = tif1.get_session(
    2025,
    "Monaco",
    "Race",
    enable_cache=False  # Skip cache, always fetch from CDN
)

# This is useful when:
# - You know data has been updated on CDN
# - You're debugging cache issues
# - You need guaranteed fresh data
# - You're testing data pipeline changes

# Disable cache for specific data type
session = tif1.get_session(2025, "Monaco", "Race")
laps = session.laps  # Uses cache
telemetry = session.get_telemetry(enable_cache=False)  # Bypasses cache

# Disable cache globally (affects all requests)
config = tif1.get_config()
config.cache_enabled = False

# Now all requests bypass cache
session = tif1.get_session(2025, "Monaco", "Race")
laps = session.laps  # Fetches from CDN

# Re-enable cache
config.cache_enabled = True

# Cache entry includes schema version
# Old entry: schema_version=1
# New code: schema_version=2
# Result: Cache miss, fetch fresh data

# This happens automatically when:
# - tif1 library is updated
# - Data format changes on CDN
# - Column names or types change
# - New data fields are added

import tif1

config = tif1.get_config()

# Enable CDN freshness checks
config.cache_check_freshness = True

# Check frequency (seconds)
config.cache_freshness_interval = 3600  # Check every hour

# When enabled, the system:
# 1. Sends HEAD request to CDN with If-None-Match (ETag)
# 2. If CDN returns 304 Not Modified: Use cache
# 3. If CDN returns 200 OK: Fetch fresh data, update cache

# This adds ~50-100ms latency but ensures freshness

# Corruption detection happens when:
# - Decompression fails (corrupted zstd data)
# - Parquet parsing fails (corrupted format)
# - Checksum mismatch (if enabled)
# - SQLite integrity check fails

# When corruption is detected:
# 1. Log warning with entry details
# 2. Remove corrupted entry from cache
# 3. Fetch fresh data from CDN
# 4. Store new data in cache

# Enable checksum verification (adds ~5ms overhead)
config = tif1.get_config()
config.cache_verify_checksum = True

import tif1
from datetime import timedelta

config = tif1.get_config()

# Set global cache TTL (time to live)
config.cache_ttl = timedelta(days=7)

# Entries older than 7 days will be automatically refetched
# This is checked on every cache access

# Set per-data-type TTL
config.cache_ttl_laps = timedelta(days=30)  # Lap data rarely changes
config.cache_ttl_telemetry = timedelta(days=30)
config.cache_ttl_weather = timedelta(days=7)  # Weather might be updated
config.cache_ttl_messages = timedelta(days=7)

# Disable TTL (cache never expires)
config.cache_ttl = None

# Check TTL status
cache = tif1.get_cache()
expired = cache.get_expired_entries()
print(f"Found {len(expired)} expired entries")

# Manually remove expired entries
removed = cache.clear_expired()
print(f"Removed {removed} expired entries")

import tif1

cache = tif1.get_cache()

# Invalidate when new race data is available
def on_race_complete(year: int, event: str):
    """Called when race completes and data is available."""
    # Clear all cache for this event
    cache.clear_event(year, event)
    print(f"Cache cleared for {year} {event}")

    # Optionally warm cache with fresh data
    session = tif1.get_session(year, event, "Race")
    session.load()  # Fetch and cache all data

# Invalidate when season schedule changes
def on_schedule_update(year: int):
    """Called when season schedule is updated."""
    # Clear schedule cache
    cache.clear_pattern(f"{year}_*_schedule")
    print(f"Schedule cache cleared for {year}")

# Invalidate on library update
def on_library_update():
    """Called after tif1 library update."""
    # Clear all cache (schema might have changed)
    cache.clear()
    print("All cache cleared after library update")

import tif1

cache = tif1.get_cache()

# Invalidate specific session
cache.invalidate_session(2025, "Monaco", "Race")

# Invalidate specific data type for a session
cache.invalidate_session_data(2025, "Monaco", "Race", "telemetry")

# Invalidate all practice sessions
cache.invalidate_pattern("*_practice_*")

# Invalidate all data from specific year
cache.invalidate_year(2024)

# Invalidate all data for specific event
cache.invalidate_event(2025, "Monaco")

# Invalidate based on custom criteria
def should_invalidate(entry):
    """Custom invalidation logic."""
    # Invalidate if entry is large and old
    return entry['size_mb'] > 10 and entry['age_days'] > 30

removed = cache.invalidate_custom(should_invalidate)
print(f"Invalidated {removed} entries")

import tif1

# Cache entries include version information:
# - schema_version: Data structure version
# - library_version: tif1 library version
# - format_version: Serialization format version

# When library is updated:
# 1. Check cache entry versions
# 2. If versions mismatch: Invalidate entry
# 3. Fetch fresh data with new version
# 4. Store with updated version info

# Force cache version upgrade
cache = tif1.get_cache()
cache.upgrade_version()  # Invalidates all old-version entries

# Check cache version compatibility
compatible = cache.check_version_compatibility()
if not compatible:
    print("Cache version incompatible, clearing...")
    cache.clear()

import tif1
import asyncio
from datetime import datetime

async def warm_season(year: int, session_types: list[str] = None):
    """
    Pre-cache all races for a season.

    Args:
        year: Season year
        session_types: List of session types to cache (default: all)
    """
    if session_types is None:
        session_types = ["Practice 1", "Practice 2", "Practice 3",
                        "Qualifying", "Sprint", "Race"]

    # Get all events for the season
    events = tif1.get_events(year)
    print(f"Warming cache for {len(events)} events in {year}")

    tasks = []
    for event in events:
        for session_type in session_types:
            try:
                session = tif1.get_session(year, event, session_type)
                # Fetch all data types asynchronously
                tasks.append(session.load_async(
                    laps=True,
                    telemetry=True,
                    weather=True,
                    messages=True
                ))
            except Exception as e:
                print(f"Skipping {event} {session_type}: {e}")

    # Execute all fetches in parallel
    print(f"Fetching {len(tasks)} sessions...")
    start = datetime.now()
    results = await asyncio.gather(*tasks, return_exceptions=True)
    elapsed = (datetime.now() - start).total_seconds()

    # Count successes and failures
    successes = sum(1 for r in results if not isinstance(r, Exception))
    failures = len(results) - successes

    print(f"\nCache warming complete:")
    print(f"  Time: {elapsed:.1f}s")
    print(f"  Successes: {successes}")
    print(f"  Failures: {failures}")
    print(f"  Avg time per session: {elapsed/len(tasks):.2f}s")

    # Get cache statistics
    cache = tif1.get_cache()
    stats = cache.get_stats()
    print(f"  Cache size: {stats['total_size_mb']:.2f} MB")
    print(f"  Cache entries: {stats['total_entries']}")

# Run warming
asyncio.run(warm_season(2025))

# Warm multiple seasons
async def warm_multiple_seasons(years: list[int]):
    """Warm cache for multiple seasons."""
    for year in years:
        await warm_season(year)

asyncio.run(warm_multiple_seasons([2023, 2024, 2025]))

import tif1

def warm_event(year: int, event: str, session_types: list[str] = None):
    """
    Pre-cache all sessions for a specific event.

    Args:
        year: Season year
        event: Event name (e.g., "Monaco", "Silverstone")
        session_types: List of session types (default: all)
    """
    if session_types is None:
        session_types = ["Practice 1", "Practice 2", "Practice 3",
                        "Qualifying", "Sprint", "Race"]

    print(f"Warming cache for {year} {event}")

    for session_type in session_types:
        try:
            session = tif1.get_session(year, event, session_type)
            # Load all data
            session.load(laps=True, telemetry=True, weather=True, messages=True)
            print(f"  ✓ {session_type}")
        except Exception as e:
            print(f"  ✗ {session_type}: {e}")

    print(f"Cache warming complete for {event}")

# Warm specific events
warm_event(2025, "Monaco")
warm_event(2025, "Silverstone")
warm_event(2025, "Spa")

# Warm upcoming race weekend
def warm_upcoming_race():
    """Warm cache for the next race weekend."""
    # Get next event
    next_event = tif1.get_next_event()
    if next_event:
        warm_event(next_event['year'], next_event['name'])
    else:
        print("No upcoming events")

warm_upcoming_race()

import tif1

def warm_telemetry(year: int, event: str, session_type: str = "Race"):
    """
    Pre-cache telemetry for all drivers.

    This is useful when you know you'll need telemetry data
    but not necessarily lap times or weather.
    """
    print(f"Warming telemetry cache for {year} {event} {session_type}")

    session = tif1.get_session(year, event, session_type)

    # Get all drivers
    drivers = session.drivers
    print(f"Fetching telemetry for {len(drivers)} drivers...")

    # Fetch telemetry for each driver
    for driver in drivers:
        try:
            tel = session.get_driver_telemetry(driver)
            print(f"  ✓ {driver}: {len(tel)} samples")
        except Exception as e:
            print(f"  ✗ {driver}: {e}")

    print(f"Telemetry cache warming complete")

warm_telemetry(2025, "Monaco", "Race")

def warm_laps_only(year: int, event: str, session_type: str = "Race"):
    """Pre-cache only lap data (fastest warming)."""
    session = tif1.get_session(year, event, session_type)
    laps = session.laps
    print(f"Cached {len(laps)} laps for {event}")

def warm_weather_only(year: int, event: str, session_type: str = "Race"):
    """Pre-cache only weather data."""
    session = tif1.get_session(year, event, session_type)
    weather = session.weather
    print(f"Cached {len(weather)} weather samples for {event}")

import tif1

def warm_driver_data(year: int, driver: str, events: list[str] = None):
    """
    Pre-cache all data for a specific driver across multiple events.

    Args:
        year: Season year
        driver: Driver identifier (e.g., "VER", "HAM", "LEC")
        events: List of events (default: all events)
    """
    if events is None:
        events = [e['name'] for e in tif1.get_events(year)]

    print(f"Warming cache for driver {driver} in {year}")

    for event in events:
        try:
            session = tif1.get_session(year, event, "Race")

            # Get driver laps
            laps = session.get_driver_laps(driver)

            # Get driver telemetry
            tel = session.get_driver_telemetry(driver)

            print(f"  ✓ {event}: {len(laps)} laps, {len(tel)} telemetry samples")
        except Exception as e:
            print(f"  ✗ {event}: {e}")

    print(f"Driver cache warming complete for {driver}")

# Warm cache for specific drivers
warm_driver_data(2025, "VER")  # Verstappen
warm_driver_data(2025, "HAM")  # Hamilton
warm_driver_data(2025, "LEC")  # Leclerc

# Warm cache for all drivers in a race
def warm_all_drivers(year: int, event: str, session_type: str = "Race"):
    """Pre-cache data for all drivers in a race."""
    session = tif1.get_session(year, event, session_type)
    drivers = session.drivers

    print(f"Warming cache for {len(drivers)} drivers")

    for driver in drivers:
        try:
            laps = session.get_driver_laps(driver)
            tel = session.get_driver_telemetry(driver)
            print(f"  ✓ {driver}")
        except Exception as e:
            print(f"  ✗ {driver}: {e}")

warm_all_drivers(2025, "Monaco")

import tif1
import schedule
import time
from datetime import datetime

def scheduled_warm_upcoming():
    """Warm cache for upcoming race weekend."""
    print(f"[{datetime.now()}] Starting scheduled cache warming...")

    # Get next event
    next_event = tif1.get_next_event()
    if next_event:
        year = next_event['year']
        event = next_event['name']

        # Warm cache for all sessions
        for session_type in ["Practice 1", "Practice 2", "Practice 3",
                            "Qualifying", "Sprint", "Race"]:
            try:
                session = tif1.get_session(year, event, session_type)
                session.load(laps=True, telemetry=True, weather=True, messages=True)
                print(f"  ✓ {session_type}")
            except Exception as e:
                print(f"  ✗ {session_type}: {e}")

        print(f"Cache warming complete for {event}")
    else:
        print("No upcoming events")

# Schedule warming every day at 2 AM
schedule.every().day.at("02:00").do(scheduled_warm_upcoming)

# Schedule warming every Monday
schedule.every().monday.at("00:00").do(scheduled_warm_upcoming)

# Run scheduler
print("Cache warming scheduler started")
while True:
    schedule.run_pending()
    time.sleep(60)  # Check every minute

import tif1
import asyncio
from concurrent.futures import ThreadPoolExecutor, as_completed

async def parallel_warm_season(year: int, max_workers: int = 10):
    """
    Warm cache for entire season using parallel execution.

    Args:
        year: Season year
        max_workers: Maximum parallel workers (default: 10)
    """
    events = tif1.get_events(year)
    session_types = ["Practice 1", "Practice 2", "Practice 3",
                    "Qualifying", "Sprint", "Race"]

    # Create all tasks
    tasks = []
    for event in events:
        for session_type in session_types:
            tasks.append((year, event['name'], session_type))

    print(f"Warming {len(tasks)} sessions with {max_workers} workers...")

    async def warm_session(year, event, session_type):
        """Warm a single session."""
        try:
            session = tif1.get_session(year, event, session_type)
            await session.load_async(laps=True, telemetry=True,
                                    weather=True, messages=True)
            return f"✓ {event} {session_type}"
        except Exception as e:
            return f"✗ {event} {session_type}: {e}"

    # Execute in parallel with semaphore to limit concurrency
    semaphore = asyncio.Semaphore(max_workers)

    async def warm_with_limit(task):
        async with semaphore:
            return await warm_session(*task)

    # Run all tasks
    start = asyncio.get_event_loop().time()
    results = await asyncio.gather(*[warm_with_limit(t) for t in tasks])
    elapsed = asyncio.get_event_loop().time() - start

    # Print results
    for result in results:
        print(result)

    successes = sum(1 for r in results if r.startswith("✓"))
    print(f"\nCompleted in {elapsed:.1f}s")
    print(f"Success rate: {successes}/{len(tasks)} ({successes/len(tasks)*100:.1f}%)")

# Run parallel warming
asyncio.run(parallel_warm_season(2025, max_workers=10))

import tif1
from collections import Counter

def smart_warm_cache(min_access_count: int = 5):
    """
    Warm cache for frequently accessed sessions.

    Analyzes cache access patterns and pre-warms frequently
    accessed sessions that are not currently cached.
    """
    cache = tif1.get_cache()

    # Get access statistics
    stats = cache.get_access_stats()

    # Find frequently accessed sessions
    frequent = [
        s for s in stats
        if s['access_count'] >= min_access_count
    ]

    print(f"Found {len(frequent)} frequently accessed sessions")

    # Warm cache for these sessions
    for session_info in frequent:
        year = session_info['year']
        event = session_info['event']
        session_type = session_info['session_type']

        # Check if currently cached
        if not cache.exists(year, event, session_type):
            print(f"Warming {year} {event} {session_type}...")
            try:
                session = tif1.get_session(year, event, session_type)
                session.load(laps=True, telemetry=True,
                           weather=True, messages=True)
                print(f"  ✓ Cached")
            except Exception as e:
                print(f"  ✗ Error: {e}")
        else:
            print(f"  ✓ Already cached: {year} {event} {session_type}")

smart_warm_cache(min_access_count=5)

# Calculate memory requirements for different cache sizes

# Small cache (100 items, typical sessions)
small_cache_mb = 100 * 20  # 2 GB

# Medium cache (200 items)
medium_cache_mb = 200 * 20  # 4 GB

# Large cache (500 items)
large_cache_mb = 500 * 20  # 10 GB

# Full season (24 races × 5 sessions)
full_season_mb = 24 * 5 * 20  # 2.4 GB

print(f"Small cache: {small_cache_mb / 1024:.1f} GB")
print(f"Medium cache: {medium_cache_mb / 1024:.1f} GB")
print(f"Large cache: {large_cache_mb / 1024:.1f} GB")
print(f"Full season: {full_season_mb / 1024:.1f} GB")

# Estimate cache growth

# Per race weekend (5 sessions)
per_weekend_mb = 5 * 6  # 30 MB compressed

# Per season (24 races)
per_season_mb = 24 * per_weekend_mb  # 720 MB

# Multiple seasons
seasons = 5
total_mb = seasons * per_season_mb  # 3.6 GB

print(f"Per weekend: {per_weekend_mb} MB")
print(f"Per season: {per_season_mb} MB")
print(f"{seasons} seasons: {total_mb / 1024:.1f} GB")

import tif1
import time
from contextlib import contextmanager

@contextmanager
def timer(name: str):
    """Context manager for timing operations."""
    start = time.perf_counter()
    yield
    elapsed = (time.perf_counter() - start) * 1000
    print(f"{name}: {elapsed:.2f}ms")

# Profile cache operations
cache = tif1.get_cache()

# Profile cache lookup
with timer("Cache lookup"):
    exists = cache.exists(2025, "Monaco", "Race", "laps")

# Profile cache read
with timer("Cache read"):
    session = tif1.get_session(2025, "Monaco", "Race")
    laps = session.laps

# Profile cache write
with timer("Cache write"):
    cache.clear_session(2025, "Monaco", "Race")
    session = tif1.get_session(2025, "Monaco", "Race")
    laps = session.laps

# Profile memory vs SQLite
cache.clear_memory()  # Clear memory cache

with timer("SQLite read"):
    laps1 = session.laps  # From SQLite

with timer("Memory read"):
    laps2 = session.laps  # From memory

# Profile compression
import pandas as pd
import io

df = session.laps

with timer("Parquet serialization"):
    buffer = io.BytesIO()
    df.to_parquet(buffer)
    parquet_bytes = buffer.getvalue()

with timer("Zstd compression"):
    import zstandard as zstd
    compressor = zstd.ZstdCompressor(level=3)
    compressed = compressor.compress(parquet_bytes)

print(f"Original size: {len(parquet_bytes) / 1024:.2f} KB")
print(f"Compressed size: {len(compressed) / 1024:.2f} KB")
print(f"Ratio: {len(parquet_bytes) / len(compressed):.1f}:1")

import tif1
import time
import json
from datetime import datetime

class CacheMonitor:
    """Monitor cache performance metrics."""

    def __init__(self):
        self.cache = tif1.get_cache()
        self.metrics = {
            'hits': 0,
            'misses': 0,
            'total_access_time_ms': 0,
            'access_count': 0,
        }

    def record_access(self, hit: bool, access_time_ms: float):
        """Record a cache access."""
        if hit:
            self.metrics['hits'] += 1
        else:
            self.metrics['misses'] += 1

        self.metrics['total_access_time_ms'] += access_time_ms
        self.metrics['access_count'] += 1

    def get_stats(self):
        """Get current statistics."""
        total = self.metrics['hits'] + self.metrics['misses']
        hit_rate = self.metrics['hits'] / total if total > 0 else 0
        avg_time = (self.metrics['total_access_time_ms'] /
                   self.metrics['access_count'] if self.metrics['access_count'] > 0 else 0)

        return {
            'hit_rate': hit_rate,
            'miss_rate': 1 - hit_rate,
            'total_accesses': total,
            'avg_access_time_ms': avg_time,
            'timestamp': datetime.now().isoformat(),
        }

    def export_metrics(self, filename: str):
        """Export metrics to JSON file."""
        stats = self.get_stats()
        cache_stats = self.cache.get_stats()

        combined = {
            'monitor_stats': stats,
            'cache_stats': cache_stats,
        }

        with open(filename, 'w') as f:
            json.dump(combined, f, indent=2, default=str)

        print(f"Metrics exported to {filename}")

# Use monitor
monitor = CacheMonitor()

# Wrap cache access
def monitored_get_session(year, event, session_type):
    """Get session with monitoring."""
    start = time.perf_counter()

    # Check if cached
    cache = tif1.get_cache()
    hit = cache.exists(year, event, session_type)

    # Get session
    session = tif1.get_session(year, event, session_type)
    laps = session.laps

    # Record metrics
    elapsed_ms = (time.perf_counter() - start) * 1000
    monitor.record_access(hit, elapsed_ms)

    return session

# Use monitored access
session = monitored_get_session(2025, "Monaco", "Race")

# Get and export stats
stats = monitor.get_stats()
print(f"Hit rate: {stats['hit_rate']:.1%}")
print(f"Avg access time: {stats['avg_access_time_ms']:.2f}ms")

monitor.export_metrics("cache_metrics.json")

import tif1

cache = tif1.get_cache()

# Get detailed size information
size_info = cache.get_size_info()

print("=== Cache Size Information ===")
print(f"Total entries: {size_info['total_entries']}")
print(f"Uncompressed size: {size_info['uncompressed_mb']:.2f} MB")
print(f"Compressed size: {size_info['compressed_mb']:.2f} MB")
print(f"Disk usage: {size_info['disk_mb']:.2f} MB")
print(f"Compression ratio: {size_info['compression_ratio']:.1f}:1")
print(f"Space saved: {size_info['space_saved_mb']:.2f} MB ({size_info['space_saved_pct']:.1f}%)")

# Get size by data type
print("\n=== Size by Data Type ===")
type_sizes = cache.get_size_by_type()
for data_type, size_mb in sorted(type_sizes.items(), key=lambda x: x[1], reverse=True):
    print(f"{data_type}: {size_mb:.2f} MB")

# Get size by year
print("\n=== Size by Year ===")
year_sizes = cache.get_size_by_year()
for year, size_mb in sorted(year_sizes.items(), reverse=True):
    print(f"{year}: {size_mb:.2f} MB")

# List cached sessions with sizes
print("\n=== Cached Sessions (Top 10 by Size) ===")
sessions = cache.list_sessions(sort_by='size', limit=10)
for i, session in enumerate(sessions, 1):
    print(f"{i}. {session['year']} {session['event']} {session['session_type']}")
    print(f"   Size: {session['size_mb']:.2f} MB (compressed: {session['compressed_mb']:.2f} MB)")
    print(f"   Created: {session['created_at']}")
    print(f"   Accessed: {session['accessed_at']} ({session['access_count']} times)")

# Check disk space availability
import shutil
cache_dir = cache.get_cache_dir()
disk_usage = shutil.disk_usage(cache_dir)

print(f"\n=== Disk Space ===")
print(f"Total: {disk_usage.total / (1024**3):.2f} GB")
print(f"Used: {disk_usage.used / (1024**3):.2f} GB")
print(f"Free: {disk_usage.free / (1024**3):.2f} GB")
print(f"Cache usage: {size_info['disk_mb'] / 1024:.2f} GB ({size_info['disk_mb'] / (disk_usage.total / 1024**2) * 100:.2f}% of disk)")