chore: upgrade all Python packages and fix linting issues

- Update all dependencies to latest versions (fastmcp, httpx, packaging, etc.)
- Downgrade click from yanked 8.2.2 to stable 8.1.7
- Fix code formatting and linting issues with ruff
- Most tests passing (2 test failures in dependency resolver need investigation)
This commit is contained in:
Ryan Malloy 2025-08-15 20:23:14 -06:00
parent 503ea589f1
commit 8b43927493
34 changed files with 2276 additions and 1593 deletions

View file

@ -5,7 +5,7 @@ import logging
import random
import time
from datetime import datetime, timedelta
from typing import Any, Dict, List, Optional
from typing import Any
import httpx
@ -50,7 +50,7 @@ class PyPIStatsClient:
self._cache: dict[str, dict[str, Any]] = {}
self._cache_ttl = 86400 # 24 hours (increased for resilience)
self._fallback_cache_ttl = 604800 # 7 days for fallback data
# Track API health for smart fallback decisions
self._api_health = {
"last_success": None,
@ -106,31 +106,33 @@ class PyPIStatsClient:
)
return f"{endpoint}:{package_name}:{param_str}"
def _is_cache_valid(self, cache_entry: dict[str, Any], fallback: bool = False) -> bool:
def _is_cache_valid(
self, cache_entry: dict[str, Any], fallback: bool = False
) -> bool:
"""Check if cache entry is still valid.
Args:
cache_entry: Cache entry to validate
fallback: Whether to use fallback cache TTL (longer for resilience)
"""
ttl = self._fallback_cache_ttl if fallback else self._cache_ttl
return time.time() - cache_entry.get("timestamp", 0) < ttl
def _should_use_fallback(self) -> bool:
"""Determine if fallback mechanisms should be used based on API health."""
if not self.fallback_enabled:
return False
# Use fallback if we've had multiple consecutive failures
if self._api_health["consecutive_failures"] >= 3:
return True
# Use fallback if last success was more than 1 hour ago
if self._api_health["last_success"]:
time_since_success = time.time() - self._api_health["last_success"]
if time_since_success > 3600: # 1 hour
return True
return False
async def _make_request(self, url: str) -> dict[str, Any]:
@ -152,7 +154,9 @@ class PyPIStatsClient:
for attempt in range(self.max_retries + 1):
try:
logger.debug(f"Making request to {url} (attempt {attempt + 1}/{self.max_retries + 1})")
logger.debug(
f"Making request to {url} (attempt {attempt + 1}/{self.max_retries + 1})"
)
response = await self._client.get(url)
@ -171,16 +175,25 @@ class PyPIStatsClient:
elif response.status_code == 429:
retry_after = response.headers.get("Retry-After")
retry_after_int = int(retry_after) if retry_after else None
self._update_api_failure(f"Rate limit exceeded (retry after {retry_after_int}s)")
self._update_api_failure(
f"Rate limit exceeded (retry after {retry_after_int}s)"
)
raise RateLimitError(retry_after_int)
elif response.status_code >= 500:
error_msg = f"Server error: HTTP {response.status_code}"
self._update_api_failure(error_msg)
# For 502/503/504 errors, continue retrying
if response.status_code in [502, 503, 504] and attempt < self.max_retries:
last_exception = PyPIServerError(response.status_code, error_msg)
logger.warning(f"Retryable server error {response.status_code}, attempt {attempt + 1}")
if (
response.status_code in [502, 503, 504]
and attempt < self.max_retries
):
last_exception = PyPIServerError(
response.status_code, error_msg
)
logger.warning(
f"Retryable server error {response.status_code}, attempt {attempt + 1}"
)
else:
raise PyPIServerError(response.status_code, error_msg)
else:
@ -205,7 +218,9 @@ class PyPIStatsClient:
# Only retry certain server errors
if e.status_code in [502, 503, 504] and attempt < self.max_retries:
last_exception = e
logger.warning(f"Retrying server error {e.status_code}, attempt {attempt + 1}")
logger.warning(
f"Retrying server error {e.status_code}, attempt {attempt + 1}"
)
else:
raise
except Exception as e:
@ -216,7 +231,7 @@ class PyPIStatsClient:
# Calculate exponential backoff with jitter
if attempt < self.max_retries:
base_delay = self.retry_delay * (2 ** attempt)
base_delay = self.retry_delay * (2**attempt)
jitter = random.uniform(0.1, 0.3) * base_delay # Add 10-30% jitter
delay = base_delay + jitter
logger.debug(f"Waiting {delay:.2f}s before retry...")
@ -227,21 +242,25 @@ class PyPIStatsClient:
raise last_exception
else:
raise NetworkError("All retry attempts failed with unknown error")
def _update_api_failure(self, error_msg: str) -> None:
"""Update API health tracking on failure."""
self._api_health["consecutive_failures"] += 1
self._api_health["last_error"] = error_msg
logger.debug(f"API failure count: {self._api_health['consecutive_failures']}, error: {error_msg}")
def _generate_fallback_recent_downloads(self, package_name: str, period: str = "month") -> dict[str, Any]:
logger.debug(
f"API failure count: {self._api_health['consecutive_failures']}, error: {error_msg}"
)
def _generate_fallback_recent_downloads(
self, package_name: str, period: str = "month"
) -> dict[str, Any]:
"""Generate fallback download statistics when API is unavailable.
This provides estimated download counts based on package popularity patterns
to ensure the system remains functional during API outages.
"""
logger.warning(f"Generating fallback download data for {package_name}")
# Base estimates for popular packages (these are conservative estimates)
popular_packages = {
"requests": {"day": 1500000, "week": 10500000, "month": 45000000},
@ -270,39 +289,50 @@ class PyPIStatsClient:
"pandas": {"day": 200000, "week": 1400000, "month": 6000000},
"sqlalchemy": {"day": 90000, "week": 630000, "month": 2700000},
}
# Get estimates for known packages or generate based on package name characteristics
if package_name.lower() in popular_packages:
estimates = popular_packages[package_name.lower()]
else:
# Generate estimates based on common package patterns
if any(keyword in package_name.lower() for keyword in ["test", "dev", "debug"]):
if any(
keyword in package_name.lower() for keyword in ["test", "dev", "debug"]
):
# Development/testing packages - lower usage
base_daily = random.randint(100, 1000)
elif any(keyword in package_name.lower() for keyword in ["aws", "google", "microsoft", "azure"]):
elif any(
keyword in package_name.lower()
for keyword in ["aws", "google", "microsoft", "azure"]
):
# Cloud provider packages - higher usage
base_daily = random.randint(10000, 50000)
elif any(keyword in package_name.lower() for keyword in ["http", "request", "client", "api"]):
elif any(
keyword in package_name.lower()
for keyword in ["http", "request", "client", "api"]
):
# HTTP/API packages - moderate to high usage
base_daily = random.randint(5000, 25000)
elif any(keyword in package_name.lower() for keyword in ["data", "pandas", "numpy", "scipy"]):
elif any(
keyword in package_name.lower()
for keyword in ["data", "pandas", "numpy", "scipy"]
):
# Data science packages - high usage
base_daily = random.randint(15000, 75000)
else:
# Generic packages - moderate usage
base_daily = random.randint(1000, 10000)
estimates = {
"day": base_daily,
"week": base_daily * 7,
"month": base_daily * 30,
}
# Add some realistic variation (±20%)
variation = random.uniform(0.8, 1.2)
for key in estimates:
estimates[key] = int(estimates[key] * variation)
return {
"data": {
"last_day": estimates["day"],
@ -314,42 +344,48 @@ class PyPIStatsClient:
"source": "fallback_estimates",
"note": "Estimated data due to API unavailability. Actual values may differ.",
}
def _generate_fallback_overall_downloads(self, package_name: str, mirrors: bool = False) -> dict[str, Any]:
def _generate_fallback_overall_downloads(
self, package_name: str, mirrors: bool = False
) -> dict[str, Any]:
"""Generate fallback time series data when API is unavailable."""
logger.warning(f"Generating fallback time series data for {package_name}")
# Generate 180 days of synthetic time series data
time_series = []
base_date = datetime.now() - timedelta(days=180)
# Get base daily estimate from recent downloads fallback
recent_fallback = self._generate_fallback_recent_downloads(package_name)
base_daily = recent_fallback["data"]["last_day"]
for i in range(180):
current_date = base_date + timedelta(days=i)
# Add weekly and seasonal patterns
day_of_week = current_date.weekday()
# Lower downloads on weekends
week_factor = 0.7 if day_of_week >= 5 else 1.0
# Add some growth trend (packages generally grow over time)
growth_factor = 1.0 + (i / 180) * 0.3 # 30% growth over 180 days
# Add random daily variation
daily_variation = random.uniform(0.7, 1.3)
daily_downloads = int(base_daily * week_factor * growth_factor * daily_variation)
daily_downloads = int(
base_daily * week_factor * growth_factor * daily_variation
)
category = "with_mirrors" if mirrors else "without_mirrors"
time_series.append({
"category": category,
"date": current_date.strftime("%Y-%m-%d"),
"downloads": daily_downloads,
})
time_series.append(
{
"category": category,
"date": current_date.strftime("%Y-%m-%d"),
"downloads": daily_downloads,
}
)
return {
"data": time_series,
"package": package_name,
@ -385,16 +421,24 @@ class PyPIStatsClient:
if self._is_cache_valid(cache_entry):
logger.debug(f"Using cached recent downloads for: {normalized_name}")
return cache_entry["data"]
elif self._should_use_fallback() and self._is_cache_valid(cache_entry, fallback=True):
logger.info(f"Using extended cache (fallback mode) for: {normalized_name}")
elif self._should_use_fallback() and self._is_cache_valid(
cache_entry, fallback=True
):
logger.info(
f"Using extended cache (fallback mode) for: {normalized_name}"
)
cache_entry["data"]["note"] = "Extended cache data due to API issues"
return cache_entry["data"]
# Check if we should use fallback immediately
if self._should_use_fallback():
logger.warning(f"API health poor, using fallback data for: {normalized_name}")
fallback_data = self._generate_fallback_recent_downloads(normalized_name, period)
logger.warning(
f"API health poor, using fallback data for: {normalized_name}"
)
fallback_data = self._generate_fallback_recent_downloads(
normalized_name, period
)
# Cache fallback data with extended TTL
self._cache[cache_key] = {"data": fallback_data, "timestamp": time.time()}
return fallback_data
@ -418,28 +462,39 @@ class PyPIStatsClient:
except (PyPIServerError, NetworkError) as e:
logger.error(f"API request failed for {normalized_name}: {e}")
# Try to use stale cache data if available
if use_cache and cache_key in self._cache:
cache_entry = self._cache[cache_key]
logger.warning(f"Using stale cache data for {normalized_name} due to API failure")
logger.warning(
f"Using stale cache data for {normalized_name} due to API failure"
)
cache_entry["data"]["note"] = f"Stale cache data due to API error: {e}"
return cache_entry["data"]
# Last resort: generate fallback data
if self.fallback_enabled:
logger.warning(f"Generating fallback data for {normalized_name} due to API failure")
fallback_data = self._generate_fallback_recent_downloads(normalized_name, period)
logger.warning(
f"Generating fallback data for {normalized_name} due to API failure"
)
fallback_data = self._generate_fallback_recent_downloads(
normalized_name, period
)
# Cache fallback data
self._cache[cache_key] = {"data": fallback_data, "timestamp": time.time()}
self._cache[cache_key] = {
"data": fallback_data,
"timestamp": time.time(),
}
return fallback_data
# If fallback is disabled, re-raise the original exception
raise
except Exception as e:
logger.error(f"Unexpected error fetching recent downloads for {normalized_name}: {e}")
logger.error(
f"Unexpected error fetching recent downloads for {normalized_name}: {e}"
)
raise
async def get_overall_downloads(
@ -469,16 +524,24 @@ class PyPIStatsClient:
if self._is_cache_valid(cache_entry):
logger.debug(f"Using cached overall downloads for: {normalized_name}")
return cache_entry["data"]
elif self._should_use_fallback() and self._is_cache_valid(cache_entry, fallback=True):
logger.info(f"Using extended cache (fallback mode) for: {normalized_name}")
elif self._should_use_fallback() and self._is_cache_valid(
cache_entry, fallback=True
):
logger.info(
f"Using extended cache (fallback mode) for: {normalized_name}"
)
cache_entry["data"]["note"] = "Extended cache data due to API issues"
return cache_entry["data"]
# Check if we should use fallback immediately
if self._should_use_fallback():
logger.warning(f"API health poor, using fallback data for: {normalized_name}")
fallback_data = self._generate_fallback_overall_downloads(normalized_name, mirrors)
logger.warning(
f"API health poor, using fallback data for: {normalized_name}"
)
fallback_data = self._generate_fallback_overall_downloads(
normalized_name, mirrors
)
# Cache fallback data with extended TTL
self._cache[cache_key] = {"data": fallback_data, "timestamp": time.time()}
return fallback_data
@ -502,28 +565,39 @@ class PyPIStatsClient:
except (PyPIServerError, NetworkError) as e:
logger.error(f"API request failed for {normalized_name}: {e}")
# Try to use stale cache data if available
if use_cache and cache_key in self._cache:
cache_entry = self._cache[cache_key]
logger.warning(f"Using stale cache data for {normalized_name} due to API failure")
logger.warning(
f"Using stale cache data for {normalized_name} due to API failure"
)
cache_entry["data"]["note"] = f"Stale cache data due to API error: {e}"
return cache_entry["data"]
# Last resort: generate fallback data
if self.fallback_enabled:
logger.warning(f"Generating fallback data for {normalized_name} due to API failure")
fallback_data = self._generate_fallback_overall_downloads(normalized_name, mirrors)
logger.warning(
f"Generating fallback data for {normalized_name} due to API failure"
)
fallback_data = self._generate_fallback_overall_downloads(
normalized_name, mirrors
)
# Cache fallback data
self._cache[cache_key] = {"data": fallback_data, "timestamp": time.time()}
self._cache[cache_key] = {
"data": fallback_data,
"timestamp": time.time(),
}
return fallback_data
# If fallback is disabled, re-raise the original exception
raise
except Exception as e:
logger.error(f"Unexpected error fetching overall downloads for {normalized_name}: {e}")
logger.error(
f"Unexpected error fetching overall downloads for {normalized_name}: {e}"
)
raise
def clear_cache(self):