aiecs 1.0.1__py3-none-any.whl → 1.7.6__py3-none-any.whl

aiecs/tools/apisource/monitoring/metrics.py

@@ -0,0 +1,330 @@
+"""
+Detailed Metrics and Health Monitoring for API Providers
+
+This module provides comprehensive performance tracking including:
+- Response time percentiles
+- Data volume statistics
+- Error type distribution
+- Rate limiting events
+- Cache hit rates
+- Overall health scoring
+"""
+
+import logging
+from collections import defaultdict
+from datetime import datetime
+from threading import Lock
+from typing import Any, Dict, List, Optional, cast
+
+logger = logging.getLogger(__name__)
+
+# Type definitions for metrics structure
+MetricsDict = Dict[str, Any]
+RequestsMetrics = Dict[str, int]
+PerformanceMetrics = Dict[str, Any]
+DataVolumeMetrics = Dict[str, Any]
+ErrorsMetrics = Dict[str, Any]
+RateLimitingMetrics = Dict[str, Any]
+TimestampsMetrics = Dict[str, Optional[str]]
+
+
+class DetailedMetrics:
+    """
+    Tracks detailed performance metrics for API providers.
+
+    Provides comprehensive monitoring including response times, data volumes,
+    error patterns, and overall health scoring.
+    """
+
+    def __init__(self, max_response_times: int = 100):
+        """
+        Initialize metrics tracker.
+
+        Args:
+            max_response_times: Maximum number of response times to keep in memory
+        """
+        self.max_response_times = max_response_times
+        self.lock = Lock()
+
+        # Request metrics with proper type annotations
+        self.metrics: Dict[str, MetricsDict] = {
+            "requests": {
+                "total": 0,
+                "successful": 0,
+                "failed": 0,
+                "cached": 0,
+            },
+            "performance": {
+                "response_times": [],  # Last N response times
+                "avg_response_time_ms": 0.0,
+                "p50_response_time_ms": 0.0,
+                "p95_response_time_ms": 0.0,
+                "p99_response_time_ms": 0.0,
+                "min_response_time_ms": 0.0,
+                "max_response_time_ms": 0.0,
+            },
+            "data_volume": {
+                "total_records_fetched": 0,
+                "total_bytes_transferred": 0,
+                "avg_records_per_request": 0.0,
+                "avg_bytes_per_request": 0.0,
+            },
+            "errors": {
+                "by_type": defaultdict(int),  # {error_type: count}
+                "recent_errors": [],  # Last 10 errors with details
+            },
+            "rate_limiting": {
+                "throttled_requests": 0,
+                "total_wait_time_ms": 0.0,
+                "avg_wait_time_ms": 0.0,
+            },
+            "timestamps": {
+                "first_request": None,
+                "last_request": None,
+                "last_success": None,
+                "last_failure": None,
+            },
+        }
+
+    def record_request(
+        self,
+        success: bool,
+        response_time_ms: float,
+        record_count: int = 0,
+        bytes_transferred: int = 0,
+        cached: bool = False,
+        error_type: Optional[str] = None,
+        error_message: Optional[str] = None,
+    ):
+        """
+        Record a request with its metrics.
+
+        Args:
+            success: Whether the request was successful
+            response_time_ms: Response time in milliseconds
+            record_count: Number of records returned
+            bytes_transferred: Bytes transferred in the response
+            cached: Whether the response was cached
+            error_type: Type of error if failed (e.g., 'timeout', 'auth', 'rate_limit')
+            error_message: Error message if failed
+        """
+        with self.lock:
+            now = datetime.utcnow().isoformat()
+
+            # Get typed references to nested dictionaries
+            requests = cast(RequestsMetrics, self.metrics["requests"])
+            timestamps = cast(TimestampsMetrics, self.metrics["timestamps"])
+            performance = cast(PerformanceMetrics, self.metrics["performance"])
+            data_volume = cast(DataVolumeMetrics, self.metrics["data_volume"])
+            errors = cast(ErrorsMetrics, self.metrics["errors"])
+
+            # Update request counts
+            requests["total"] += 1
+            if success:
+                requests["successful"] += 1
+                timestamps["last_success"] = now
+            else:
+                requests["failed"] += 1
+                timestamps["last_failure"] = now
+
+            if cached:
+                requests["cached"] += 1
+
+            # Update timestamps
+            if timestamps["first_request"] is None:
+                timestamps["first_request"] = now
+            timestamps["last_request"] = now
+
+            # Update performance metrics
+            response_times = cast(List[float], performance["response_times"])
+            response_times.append(response_time_ms)
+            if len(response_times) > self.max_response_times:
+                response_times.pop(0)
+
+            # Calculate percentiles
+            self._calculate_percentiles()
+
+            # Update data volume metrics
+            data_volume["total_records_fetched"] += record_count
+            data_volume["total_bytes_transferred"] += bytes_transferred
+
+            total_requests = requests["total"]
+            if total_requests > 0:
+                data_volume["avg_records_per_request"] = data_volume["total_records_fetched"] / total_requests
+                data_volume["avg_bytes_per_request"] = data_volume["total_bytes_transferred"] / total_requests
+
+            # Record errors
+            if not success and error_type:
+                by_type = cast(Dict[str, int], errors["by_type"])
+                by_type[error_type] += 1
+
+                error_entry = {
+                    "type": error_type,
+                    "message": error_message or "Unknown error",
+                    "timestamp": now,
+                    "response_time_ms": response_time_ms,
+                }
+
+                recent_errors = cast(List[Dict[str, Any]], errors["recent_errors"])
+                recent_errors.append(error_entry)
+                if len(recent_errors) > 10:
+                    recent_errors.pop(0)
+
+    def record_rate_limit_wait(self, wait_time_ms: float):
+        """
+        Record a rate limit wait event.
+
+        Args:
+            wait_time_ms: Time waited in milliseconds
+        """
+        with self.lock:
+            rate_limiting = cast(RateLimitingMetrics, self.metrics["rate_limiting"])
+            rate_limiting["throttled_requests"] += 1
+            rate_limiting["total_wait_time_ms"] += wait_time_ms
+
+            throttled = rate_limiting["throttled_requests"]
+            if throttled > 0:
+                rate_limiting["avg_wait_time_ms"] = rate_limiting["total_wait_time_ms"] / throttled
+
+    def _calculate_percentiles(self):
+        """Calculate response time percentiles"""
+        performance = cast(PerformanceMetrics, self.metrics["performance"])
+        response_times = cast(List[float], performance["response_times"])
+        times = sorted(response_times)
+        if not times:
+            return
+
+        n = len(times)
+        performance["avg_response_time_ms"] = sum(times) / n
+        performance["min_response_time_ms"] = times[0]
+        performance["max_response_time_ms"] = times[-1]
+        performance["p50_response_time_ms"] = times[n // 2]
+        performance["p95_response_time_ms"] = times[int(n * 0.95)]
+        performance["p99_response_time_ms"] = times[min(int(n * 0.99), n - 1)]
+
+    def _calculate_health_score_unlocked(self) -> float:
+        """
+        Calculate health score without acquiring lock (internal use only).
+        Must be called while holding self.lock.
+        """
+        requests = cast(RequestsMetrics, self.metrics["requests"])
+        performance = cast(PerformanceMetrics, self.metrics["performance"])
+        errors = cast(ErrorsMetrics, self.metrics["errors"])
+
+        total = requests["total"]
+        if total == 0:
+            return 1.0
+
+        # Success rate score (40%)
+        success_rate = requests["successful"] / total
+        success_score = success_rate * 0.4
+
+        # Performance score (30%)
+        avg_time = cast(float, performance["avg_response_time_ms"])
+        # Assume < 200ms is excellent, > 2000ms is poor
+        if avg_time < 200:
+            performance_score = 0.3
+        elif avg_time > 2000:
+            performance_score = 0.0
+        else:
+            performance_score = max(0, min(1, (2000 - avg_time) / 1800)) * 0.3
+
+        # Cache hit rate score (20%)
+        cache_rate = requests["cached"] / total
+        cache_score = cache_rate * 0.2
+
+        # Error diversity score (10%) - fewer error types is better
+        by_type = cast(Dict[str, int], errors["by_type"])
+        error_types = len(by_type)
+        error_score = max(0, (5 - error_types) / 5) * 0.1
+
+        return success_score + performance_score + cache_score + error_score
+
+    def get_health_score(self) -> float:
+        """
+        Calculate overall health score (0-1).
+
+        The health score considers:
+        - Success rate (40%)
+        - Performance (30%)
+        - Cache hit rate (20%)
+        - Error diversity (10%)
+
+        Returns:
+            Health score between 0 and 1
+        """
+        with self.lock:
+            return self._calculate_health_score_unlocked()
+
+    def get_stats(self) -> Dict[str, Any]:
+        """
+        Get all metrics as a dictionary.
+
+        Returns:
+            Complete metrics dictionary
+        """
+        with self.lock:
+            # Convert defaultdict to regular dict for JSON serialization
+            requests_dict: Dict[str, int] = dict(self.metrics["requests"])  # type: ignore[arg-type]
+            performance_dict: Dict[str, Any] = dict(self.metrics["performance"])  # type: ignore[arg-type]
+            data_volume_dict: Dict[str, Any] = dict(self.metrics["data_volume"])  # type: ignore[arg-type]
+            errors_by_type: Dict[str, int] = dict(self.metrics["errors"]["by_type"])  # type: ignore[arg-type]
+            recent_errors_list: List[Dict[str, Any]] = list(self.metrics["errors"]["recent_errors"])  # type: ignore[arg-type]
+            rate_limiting_dict: Dict[str, Any] = dict(self.metrics["rate_limiting"])  # type: ignore[arg-type]
+            timestamps_dict: Dict[str, Optional[str]] = dict(self.metrics["timestamps"])  # type: ignore[arg-type]
+
+            stats: Dict[str, Any] = {
+                "requests": requests_dict,
+                "performance": performance_dict,
+                "data_volume": data_volume_dict,
+                "errors": {
+                    "by_type": errors_by_type,
+                    "recent_errors": recent_errors_list,
+                },
+                "rate_limiting": rate_limiting_dict,
+                "timestamps": timestamps_dict,
+                "health_score": self.get_health_score(),
+            }
+
+            # Remove response_times array to keep output clean
+            stats["performance"] = {k: v for k, v in stats["performance"].items() if k != "response_times"}
+
+            return stats
+
+    def get_summary(self) -> Dict[str, Any]:
+        """
+        Get a concise summary of key metrics.
+
+        Returns:
+            Summary dictionary with key metrics
+        """
+        with self.lock:
+            requests = cast(RequestsMetrics, self.metrics["requests"])
+            performance = cast(PerformanceMetrics, self.metrics["performance"])
+            errors = cast(ErrorsMetrics, self.metrics["errors"])
+
+            total = requests["total"]
+            if total == 0:
+                return {"status": "no_activity", "health_score": 1.0}
+
+            success_rate = requests["successful"] / total
+            cache_hit_rate = requests["cached"] / total
+            # Use unlocked version to avoid deadlock
+            health_score = self._calculate_health_score_unlocked()
+
+            return {
+                "status": "healthy" if health_score > 0.7 else "degraded",
+                "health_score": round(health_score, 3),
+                "total_requests": total,
+                "success_rate": round(success_rate, 3),
+                "cache_hit_rate": round(cache_hit_rate, 3),
+                "avg_response_time_ms": round(cast(float, performance["avg_response_time_ms"]), 2),
+                "p95_response_time_ms": round(cast(float, performance["p95_response_time_ms"]), 2),
+                "total_errors": requests["failed"],
+                "error_types": len(cast(Dict[str, int], errors["by_type"])),
+            }
+
+    def reset(self):
+        """Reset all metrics"""
+        with self.lock:
+            self.__init__(self.max_response_times)

aiecs/tools/apisource/providers/__init__.py

@@ -0,0 +1,112 @@
+"""
+API Providers Module
+
+Contains all API provider implementations for the APISource tool.
+"""
+
+from aiecs.tools.apisource.providers.base import BaseAPIProvider, RateLimiter
+from aiecs.tools.apisource.providers.fred import FREDProvider
+from aiecs.tools.apisource.providers.worldbank import WorldBankProvider
+from aiecs.tools.apisource.providers.newsapi import NewsAPIProvider
+from aiecs.tools.apisource.providers.census import CensusProvider
+
+import logging
+from typing import Dict, List, Optional, Type, Any
+
+logger = logging.getLogger(__name__)
+
+# Global provider registry
+PROVIDER_REGISTRY: Dict[str, Type[BaseAPIProvider]] = {}
+PROVIDER_INSTANCES: Dict[str, BaseAPIProvider] = {}
+
+
+def register_provider(provider_class: Type[BaseAPIProvider]):
+    """
+    Register a provider class.
+
+    Args:
+        provider_class: Provider class to register
+    """
+    # Instantiate to get name
+    temp_instance = provider_class()
+    provider_name = temp_instance.name
+
+    PROVIDER_REGISTRY[provider_name] = provider_class
+    logger.debug(f"Registered provider: {provider_name}")
+
+
+def get_provider(name: str, config: Optional[Dict] = None) -> BaseAPIProvider:
+    """
+    Get a provider instance by name.
+
+    Args:
+        name: Provider name
+        config: Optional configuration for the provider
+
+    Returns:
+        Provider instance
+
+    Raises:
+        ValueError: If provider is not registered
+    """
+    if name not in PROVIDER_REGISTRY:
+        raise ValueError(f"Provider '{name}' not found. " f"Available providers: {', '.join(PROVIDER_REGISTRY.keys())}")
+
+    # Return cached instance or create new one with config
+    if config is None and name in PROVIDER_INSTANCES:
+        return PROVIDER_INSTANCES[name]
+
+    provider_instance = PROVIDER_REGISTRY[name](config)
+
+    if config is None:
+        PROVIDER_INSTANCES[name] = provider_instance
+
+    return provider_instance
+
+
+def list_providers() -> List[Dict[str, Any]]:
+    """
+    List all registered providers.
+
+    Returns:
+        List of provider metadata dictionaries
+    """
+    providers = []
+    for name, provider_class in PROVIDER_REGISTRY.items():
+        try:
+            # Get or create instance to access metadata
+            provider = get_provider(name)
+            providers.append(provider.get_metadata())
+        except Exception as e:
+            logger.warning(f"Failed to get metadata for provider {name}: {e}")
+            providers.append(
+                {
+                    "name": name,
+                    "description": "Provider metadata unavailable",
+                    "operations": [],
+                    "error": str(e),
+                }
+            )
+
+    return providers
+
+
+# Auto-register all providers
+register_provider(FREDProvider)
+register_provider(WorldBankProvider)
+register_provider(NewsAPIProvider)
+register_provider(CensusProvider)
+
+
+__all__ = [
+    "BaseAPIProvider",
+    "RateLimiter",
+    "FREDProvider",
+    "WorldBankProvider",
+    "NewsAPIProvider",
+    "CensusProvider",
+    "register_provider",
+    "get_provider",
+    "list_providers",
+    "PROVIDER_REGISTRY",
+]