greenmining 0.1.10__py3-none-any.whl → 0.1.12__py3-none-any.whl

greenmining/gsf_patterns.py

@@ -554,9 +554,667 @@ GSF_PATTERNS = {
         "description": "Process multiple items together instead of one at a time",
         "sci_impact": "Reduces per-item overhead and improves throughput",
     },
+    # ==================== RESOURCE OPTIMIZATION (Expanded from VU Amsterdam) ====================
+    "resource_limits": {
+        "name": "Resource Limits & Constraints",
+        "category": "resource",
+        "keywords": [
+            "limit",
+            "limits",
+            "constraint",
+            "constraints",
+            "quota",
+            "cap",
+            "memory limit",
+            "cpu limit",
+            "resource limit",
+            "timeout",
+            "max threads",
+            "thread limit",
+        ],
+        "description": "Set hard limits on resource consumption to prevent waste",
+        "sci_impact": "Prevents resource exhaustion and runaway processes",
+    },
+    "dynamic_resource_allocation": {
+        "name": "Dynamic Resource Allocation",
+        "category": "resource",
+        "keywords": [
+            "dynamic allocation",
+            "vpa",
+            "vertical pod autoscaling",
+            "custom metric",
+            "scaling policy",
+            "predictive scaling",
+        ],
+        "description": "Dynamically adjust resource allocation based on actual usage",
+        "sci_impact": "Eliminates over-provisioning and improves efficiency",
+    },
+    # ==================== CACHING & DATA EFFICIENCY (Expanded) ====================
+    "multi_level_caching": {
+        "name": "Multi-Level Caching",
+        "category": "caching",
+        "keywords": [
+            "multi level cache",
+            "l1 cache",
+            "l2 cache",
+            "cache layer",
+            "distributed cache",
+            "local cache",
+            "cache hierarchy",
+        ],
+        "description": "Implement multiple caching levels for optimal efficiency",
+        "sci_impact": "Reduces redundant computation and network traffic",
+    },
+    "cache_invalidation": {
+        "name": "Cache Invalidation Strategy",
+        "category": "caching",
+        "keywords": [
+            "cache invalidation",
+            "cache invalidate",
+            "ttl",
+            "time to live",
+            "eviction",
+            "cache expir",
+            "cache refresh",
+            "stale while revalidate",
+        ],
+        "description": "Implement efficient cache invalidation to balance freshness and efficiency",
+        "sci_impact": "Optimizes cache hit rates and reduces unnecessary refreshes",
+    },
+    "data_deduplication": {
+        "name": "Data Deduplication",
+        "category": "data",
+        "keywords": [
+            "deduplicate",
+            "deduplication",
+            "redundancy",
+            "duplicate",
+            "unique",
+            "eliminate redundant",
+        ],
+        "description": "Eliminate redundant data storage and processing",
+        "sci_impact": "Reduces storage footprint and processing overhead",
+    },
+    "efficient_serialization": {
+        "name": "Efficient Serialization",
+        "category": "data",
+        "keywords": [
+            "protobuf",
+            "protocol buffer",
+            "avro",
+            "messagepack",
+            "serialization",
+            "binary format",
+            "efficient format",
+        ],
+        "description": "Use efficient binary serialization formats",
+        "sci_impact": "Reduces network bandwidth and serialization overhead",
+    },
+    "pagination": {
+        "name": "Pagination & Lazy Loading",
+        "category": "data",
+        "keywords": [
+            "pagination",
+            "paginate",
+            "page size",
+            "limit offset",
+            "lazy load",
+            "on demand",
+            "defer load",
+        ],
+        "description": "Load data incrementally instead of all at once",
+        "sci_impact": "Reduces memory usage and initial load time",
+    },
+    # ==================== EVENT-DRIVEN & ASYNCHRONOUS (Expanded) ====================
+    "event_driven_architecture": {
+        "name": "Event-Driven Architecture",
+        "category": "async",
+        "keywords": [
+            "event driven",
+            "event sourcing",
+            "event listener",
+            "kafka",
+            "rabbitmq",
+            "message broker",
+            "pub sub",
+            "publish subscribe",
+            "webhook",
+        ],
+        "description": "Use events instead of polling to reduce unnecessary CPU cycles",
+        "sci_impact": "Eliminates wasteful polling and enables reactive processing",
+    },
+    "eliminate_polling": {
+        "name": "Eliminate Polling",
+        "category": "async",
+        "keywords": [
+            "poll",
+            "polling",
+            "reduce poll",
+            "stop poll",
+            "websocket",
+            "sse",
+            "server sent event",
+            "push notification",
+        ],
+        "description": "Replace polling with push-based notifications",
+        "sci_impact": "Dramatically reduces unnecessary CPU and network usage",
+    },
+    "reactive_streams": {
+        "name": "Reactive Streams",
+        "category": "async",
+        "keywords": [
+            "reactive",
+            "reactor",
+            "rxjava",
+            "completablefuture",
+            "webflux",
+            "project reactor",
+            "coroutine",
+            "suspend",
+        ],
+        "description": "Use reactive programming for non-blocking I/O",
+        "sci_impact": "Improves thread utilization and reduces blocking operations",
+    },
+    # ==================== DATABASE OPTIMIZATION (Expanded) ====================
+    "query_optimization": {
+        "name": "Query Optimization",
+        "category": "database",
+        "keywords": [
+            "query optim",
+            "explain plan",
+            "query plan",
+            "n+1",
+            "n plus 1",
+            "eager load",
+            "batch query",
+            "select star",
+        ],
+        "description": "Optimize database queries to reduce CPU and I/O",
+        "sci_impact": "Reduces database server load and query execution time",
+    },
+    "prepared_statements": {
+        "name": "Prepared Statements",
+        "category": "database",
+        "keywords": [
+            "prepared statement",
+            "statement cache",
+            "statement pool",
+            "parameterized query",
+            "bind variable",
+        ],
+        "description": "Use prepared statements for repeated queries",
+        "sci_impact": "Reduces query parsing overhead and improves security",
+    },
+    "database_views": {
+        "name": "Database Views & Materialized Views",
+        "category": "database",
+        "keywords": [
+            "database view",
+            "materialized view",
+            "view",
+            "precompute",
+            "aggregate table",
+        ],
+        "description": "Use views to precompute complex queries",
+        "sci_impact": "Reduces redundant computation for common queries",
+    },
+    # ==================== ALGORITHM & CODE EFFICIENCY ====================
+    "algorithm_optimization": {
+        "name": "Algorithm Optimization",
+        "category": "code",
+        "keywords": [
+            "algorithm optim",
+            "complexity",
+            "big o",
+            "o(n)",
+            "memoiz",
+            "dynamic programming",
+            "efficient algorithm",
+        ],
+        "description": "Use efficient algorithms to reduce CPU cycles",
+        "sci_impact": "Reduces computational overhead and execution time",
+    },
+    "code_efficiency": {
+        "name": "Code Efficiency & Lean Code",
+        "category": "code",
+        "keywords": [
+            "dead code",
+            "unused code",
+            "remove code",
+            "efficient code",
+            "lean code",
+            "minimize allocation",
+            "object pool",
+        ],
+        "description": "Write efficient, minimal code that uses fewer resources",
+        "sci_impact": "Reduces memory pressure and garbage collection overhead",
+    },
+    "reduce_reflection": {
+        "name": "Reduce Reflection Usage",
+        "category": "code",
+        "keywords": [
+            "reflection",
+            "avoid reflection",
+            "introspection",
+            "dynamic lookup",
+            "hot path",
+        ],
+        "description": "Avoid reflection in performance-critical code paths",
+        "sci_impact": "Reduces CPU overhead in frequently executed code",
+    },
+    "garbage_collection_tuning": {
+        "name": "Garbage Collection Tuning",
+        "category": "code",
+        "keywords": [
+            "garbage collection",
+            "gc tune",
+            "gc pause",
+            "gc optim",
+            "memory management",
+            "heap size",
+        ],
+        "description": "Optimize garbage collection settings for efficiency",
+        "sci_impact": "Reduces GC overhead and improves application responsiveness",
+    },
+    # ==================== NETWORK OPTIMIZATION (Expanded) ====================
+    "grpc_optimization": {
+        "name": "gRPC for Service Communication",
+        "category": "network",
+        "keywords": [
+            "grpc",
+            "http2",
+            "h2",
+            "multiplexing",
+            "protocol buffer",
+            "efficient protocol",
+        ],
+        "description": "Use gRPC instead of REST for high-traffic inter-service communication",
+        "sci_impact": "Reduces network overhead through binary protocols and multiplexing",
+    },
+    "request_batching": {
+        "name": "Request Batching & Aggregation",
+        "category": "network",
+        "keywords": [
+            "batch request",
+            "request batch",
+            "aggregate request",
+            "coalesce",
+            "reduce call",
+            "bulk request",
+        ],
+        "description": "Aggregate multiple requests into batches",
+        "sci_impact": "Reduces network overhead and context switching",
+    },
+    "graphql_optimization": {
+        "name": "GraphQL for Selective Retrieval",
+        "category": "network",
+        "keywords": [
+            "graphql",
+            "selective field",
+            "field selection",
+            "overfetch",
+            "underfetch",
+        ],
+        "description": "Use GraphQL to retrieve only required fields",
+        "sci_impact": "Eliminates over-fetching and reduces payload size",
+    },
+    "api_gateway": {
+        "name": "API Gateway Pattern",
+        "category": "network",
+        "keywords": [
+            "api gateway",
+            "gateway",
+            "aggregation gateway",
+            "bff",
+            "backend for frontend",
+        ],
+        "description": "Use API gateway for request aggregation and routing",
+        "sci_impact": "Reduces client-server roundtrips and enables caching",
+    },
+    "circuit_breaker_advanced": {
+        "name": "Circuit Breaker & Bulkhead",
+        "category": "network",
+        "keywords": [
+            "circuit breaker",
+            "bulkhead",
+            "isolation",
+            "fault tolerance",
+            "retry backoff",
+            "exponential backoff",
+        ],
+        "description": "Prevent cascading failures and isolate resources",
+        "sci_impact": "Avoids wasted energy on failing requests and protects resources",
+    },
+    "keep_alive_connections": {
+        "name": "Keep-Alive Connections",
+        "category": "network",
+        "keywords": [
+            "keep alive",
+            "persistent connection",
+            "connection reuse",
+            "http keep-alive",
+            "tcp keep-alive",
+        ],
+        "description": "Maintain persistent connections to avoid handshake overhead",
+        "sci_impact": "Reduces TCP/TLS handshake energy consumption",
+    },
+    # ==================== MONITORING & OBSERVABILITY ====================
+    "energy_monitoring": {
+        "name": "Energy-Aware Monitoring",
+        "category": "monitoring",
+        "keywords": [
+            "rapl",
+            "power consumption",
+            "energy monitor",
+            "carbon footprint",
+            "energy metric",
+            "watt",
+            "power usage",
+        ],
+        "description": "Monitor and measure energy consumption",
+        "sci_impact": "Enables identification of energy optimization opportunities",
+    },
+    "performance_profiling": {
+        "name": "Performance Profiling",
+        "category": "monitoring",
+        "keywords": [
+            "profil",
+            "flame graph",
+            "cpu profile",
+            "memory profile",
+            "perf",
+            "benchmark",
+        ],
+        "description": "Profile application performance to identify bottlenecks",
+        "sci_impact": "Enables targeted optimization of resource-intensive operations",
+    },
+    "apm_monitoring": {
+        "name": "Application Performance Monitoring",
+        "category": "monitoring",
+        "keywords": [
+            "apm",
+            "application performance",
+            "latency tracking",
+            "throughput",
+            "response time",
+            "slow query",
+        ],
+        "description": "Monitor application performance metrics",
+        "sci_impact": "Identifies performance degradation and optimization opportunities",
+    },
+    # ==================== AI/ML OPTIMIZATION (Expanded from VU Amsterdam) ====================
+    "model_pruning": {
+        "name": "Model Pruning",
+        "category": "ai",
+        "keywords": [
+            "prune",
+            "pruning",
+            "model prune",
+            "weight pruning",
+            "sparse model",
+        ],
+        "description": "Remove unnecessary model weights to reduce size",
+        "sci_impact": "Reduces inference time and memory requirements",
+    },
+    "model_quantization": {
+        "name": "Model Quantization",
+        "category": "ai",
+        "keywords": [
+            "quantiz",
+            "quantization",
+            "int8",
+            "fp16",
+            "mixed precision",
+            "lower precision",
+        ],
+        "description": "Reduce model precision to improve efficiency",
+        "sci_impact": "Dramatically reduces memory and compute requirements",
+    },
+    "knowledge_distillation": {
+        "name": "Knowledge Distillation",
+        "category": "ai",
+        "keywords": [
+            "distill",
+            "distillation",
+            "knowledge distill",
+            "teacher student",
+            "model compress",
+        ],
+        "description": "Compress large models into smaller ones",
+        "sci_impact": "Maintains accuracy while reducing model size and energy",
+    },
+    "early_stopping": {
+        "name": "Early Stopping",
+        "category": "ai",
+        "keywords": [
+            "early stop",
+            "early stopping",
+            "stop training",
+            "convergence",
+            "plateau",
+        ],
+        "description": "Stop training when model performance plateaus",
+        "sci_impact": "Prevents wasted training energy after convergence",
+    },
+    "efficient_ml_framework": {
+        "name": "Efficient ML Framework",
+        "category": "ai",
+        "keywords": [
+            "tensorflow lite",
+            "onnx",
+            "onnx runtime",
+            "tflite",
+            "efficient framework",
+            "optimized runtime",
+        ],
+        "description": "Use ML frameworks optimized for efficiency",
+        "sci_impact": "Reduces inference overhead and energy consumption",
+    },
+    "edge_inference": {
+        "name": "Edge Inference",
+        "category": "ai",
+        "keywords": [
+            "edge inference",
+            "edge computing",
+            "edge ai",
+            "on device",
+            "local inference",
+        ],
+        "description": "Run inference on edge devices to reduce network transfers",
+        "sci_impact": "Eliminates cloud round-trip and reduces data center load",
+    },
+    "batch_inference_optimization": {
+        "name": "Batch Inference Optimization",
+        "category": "ai",
+        "keywords": [
+            "batch inference",
+            "batch predict",
+            "inference batch",
+            "batch size",
+            "throughput optim",
+        ],
+        "description": "Optimize batch size for efficient inference",
+        "sci_impact": "Maximizes GPU utilization and throughput",
+    },
+    "model_caching": {
+        "name": "Model Prediction Caching",
+        "category": "ai",
+        "keywords": [
+            "prediction cache",
+            "inference cache",
+            "model cache",
+            "cache prediction",
+        ],
+        "description": "Cache predictions for repeated inputs",
+        "sci_impact": "Avoids redundant inference computations",
+    },
+    # ==================== MICROSERVICES-SPECIFIC PATTERNS ====================
+    "service_decomposition": {
+        "name": "Service Decomposition & Right-Sizing",
+        "category": "microservices",
+        "keywords": [
+            "service decomp",
+            "decomposition",
+            "right size service",
+            "service boundary",
+            "ddd",
+            "domain driven",
+        ],
+        "description": "Decompose services appropriately to minimize overhead",
+        "sci_impact": "Reduces inter-service communication overhead",
+    },
+    "service_colocation": {
+        "name": "Service Co-location",
+        "category": "microservices",
+        "keywords": [
+            "co-location",
+            "colocation",
+            "colocate",
+            "data locality",
+            "same host",
+            "affinity",
+        ],
+        "description": "Co-locate related services to reduce network hops",
+        "sci_impact": "Minimizes network latency and energy consumption",
+    },
+    "graceful_shutdown": {
+        "name": "Graceful Shutdown",
+        "category": "microservices",
+        "keywords": [
+            "graceful shutdown",
+            "drain connection",
+            "shutdown hook",
+            "termination handler",
+            "sigterm",
+        ],
+        "description": "Implement graceful service shutdown procedures",
+        "sci_impact": "Prevents wasted work and enables efficient decommissioning",
+    },
+    "service_mesh_optimization": {
+        "name": "Service Mesh Optimization",
+        "category": "microservices",
+        "keywords": [
+            "service mesh",
+            "istio",
+            "linkerd",
+            "envoy proxy",
+            "sidecar",
+            "mesh config",
+        ],
+        "description": "Optimize service mesh configuration for efficiency",
+        "sci_impact": "Reduces proxy overhead and improves observability efficiency",
+    },
+    # ==================== INFRASTRUCTURE PATTERNS ====================
+    "alpine_containers": {
+        "name": "Minimal Container Images",
+        "category": "infrastructure",
+        "keywords": [
+            "alpine",
+            "lightweight image",
+            "minimal image",
+            "distroless",
+            "slim image",
+            "reduce image size",
+        ],
+        "description": "Use minimal base images for containers",
+        "sci_impact": "Reduces storage, network transfer, and attack surface",
+    },
+    "infrastructure_as_code": {
+        "name": "Infrastructure as Code Optimization",
+        "category": "infrastructure",
+        "keywords": [
+            "terraform",
+            "cloudformation",
+            "infrastructure code",
+            "iac",
+            "infra automation",
+        ],
+        "description": "Automate infrastructure management for consistency",
+        "sci_impact": "Enables consistent resource optimization and right-sizing",
+    },
+    "renewable_energy_regions": {
+        "name": "Renewable Energy Regions",
+        "category": "infrastructure",
+        "keywords": [
+            "renewable energy",
+            "green region",
+            "sustainable region",
+            "clean energy",
+            "carbon aware",
+        ],
+        "description": "Deploy in regions with renewable energy",
+        "sci_impact": "Reduces carbon intensity of operations",
+    },
+    "pod_disruption_budget": {
+        "name": "Pod Disruption Budget",
+        "category": "infrastructure",
+        "keywords": [
+            "pod disruption",
+            "pdb",
+            "disruption budget",
+            "availability",
+            "maintenance",
+        ],
+        "description": "Configure pod disruption budgets for efficient updates",
+        "sci_impact": "Enables efficient rolling updates without over-provisioning",
+    },
+    # ==================== GENERAL OPTIMIZATION PATTERNS ====================
+    "feature_flags": {
+        "name": "Feature Flags for Resource Control",
+        "category": "general",
+        "keywords": [
+            "feature flag",
+            "feature toggle",
+            "flag",
+            "toggle",
+            "conditional feature",
+        ],
+        "description": "Use feature flags to control expensive features",
+        "sci_impact": "Enables selective activation of resource-intensive features",
+    },
+    "incremental_processing": {
+        "name": "Incremental Processing",
+        "category": "general",
+        "keywords": [
+            "incremental",
+            "delta",
+            "diff",
+            "change data capture",
+            "cdc",
+            "streaming",
+        ],
+        "description": "Process only changed data instead of full datasets",
+        "sci_impact": "Dramatically reduces processing overhead for large datasets",
+    },
+    "precomputation": {
+        "name": "Precomputation & Pre-aggregation",
+        "category": "general",
+        "keywords": [
+            "precompute",
+            "pre-aggregate",
+            "preaggregate",
+            "compute once",
+            "precalculate",
+        ],
+        "description": "Precompute expensive operations ahead of time",
+        "sci_impact": "Avoids redundant computation at query time",
+    },
+    "approximation_algorithms": {
+        "name": "Approximation Algorithms",
+        "category": "general",
+        "keywords": [
+            "approximate",
+            "approximation",
+            "heuristic",
+            "estimation",
+            "sampling",
+        ],
+        "description": "Use approximation for non-critical calculations",
+        "sci_impact": "Reduces computational complexity with acceptable accuracy",
+    },
 }
 
-# Green software keywords (comprehensive list from all GSF patterns)
+# Green software keywords (comprehensive list from all GSF patterns + VU Amsterdam research)
 GREEN_KEYWORDS = [
     # Core sustainability terms
     "energy",
@@ -572,6 +1230,7 @@ GREEN_KEYWORDS = [
     "eco",
     "environment",
     "renewable",
+    "clean energy",
     # Performance & optimization
     "optimize",
     "optimization",
@@ -586,6 +1245,10 @@ GREEN_KEYWORDS = [
     "latency",
     "throughput",
     "utilization",
+    "improve",
+    "improvement",
+    "enhance",
+    "enhancement",
     # Caching & storage
     "cache",
     "caching",
@@ -601,10 +1264,17 @@ GREEN_KEYWORDS = [
     "gzip",
     "brotli",
     "zstd",
+    "deflate",
     "deduplicate",
+    "deduplication",
     "dedup",
     "archive",
     "retention",
+    "ttl",
+    "time to live",
+    "eviction",
+    "invalidation",
+    "invalidate",
     # Resource management
     "pool",
     "pooling",
@@ -612,12 +1282,20 @@ GREEN_KEYWORDS = [
     "thread pool",
     "worker pool",
     "resource pool",
+    "object pool",
     "reuse",
     "recycle",
     "cleanup",
     "garbage",
     "remove unused",
     "delete unused",
+    "limit",
+    "limits",
+    "constraint",
+    "quota",
+    "cap",
+    "timeout",
+    "idle",
     # Async & concurrency
     "async",
     "asynchronous",
@@ -630,6 +1308,13 @@ GREEN_KEYWORDS = [
     "job",
     "batch",
     "batching",
+    "concurrent",
+    "parallel",
+    "reactive",
+    "event-driven",
+    "event driven",
+    "webhook",
+    "callback",
     # Scaling & sizing
     "scale",
     "scaling",
@@ -641,12 +1326,16 @@ GREEN_KEYWORDS = [
     "vertical",
     "right-size",
     "rightsize",
+    "right size",
     "downsize",
     "upsize",
     "hpa",
+    "vpa",
     "keda",
     "demand",
     "load",
+    "replica",
+    "replicas",
     # Cloud native
     "serverless",
     "lambda",
@@ -661,7 +1350,9 @@ GREEN_KEYWORDS = [
     "microservice",
     "service mesh",
     "istio",
+    "linkerd",
     "envoy",
+    "sidecar",
     # Network optimization
     "network",
     "bandwidth",
@@ -677,9 +1368,19 @@ GREEN_KEYWORDS = [
     "throttle",
     "backpressure",
     "circuit breaker",
+    "bulkhead",
     "shed",
     "reduce request",
     "minimize request",
+    "http2",
+    "h2",
+    "grpc",
+    "protobuf",
+    "protocol buffer",
+    "multiplexing",
+    "keep-alive",
+    "keep alive",
+    "persistent connection",
     # Data & database
     "index",
     "indexing",
@@ -687,11 +1388,20 @@ GREEN_KEYWORDS = [
     "optimize query",
     "query optimization",
     "partition",
+    "partitioning",
     "shard",
+    "sharding",
     "database",
     "db",
     "sql",
     "nosql",
+    "n+1",
+    "n plus 1",
+    "eager load",
+    "prepared statement",
+    "materialized view",
+    "explain",
+    "slow query",
     # Web optimization
     "minify",
     "minification",
@@ -710,9 +1420,12 @@ GREEN_KEYWORDS = [
     "avif",
     "image",
     "responsive",
+    "pagination",
+    "paginate",
     # AI/ML
     "model",
     "quantiz",
+    "quantization",
     "prune",
     "pruning",
     "distill",
@@ -720,12 +1433,21 @@ GREEN_KEYWORDS = [
     "inference",
     "training",
     "pretrain",
+    "pre-trained",
     "transfer learning",
     "gpu",
     "tpu",
     "accelerator",
     "mobilenet",
     "efficientnet",
+    "edge ai",
+    "tflite",
+    "tensorflow lite",
+    "onnx",
+    "fp16",
+    "int8",
+    "mixed precision",
+    "early stop",
     # Code & architecture
     "stateless",
     "decouple",
@@ -739,6 +1461,15 @@ GREEN_KEYWORDS = [
     "dead code",
     "unused",
     "deprecate",
+    "algorithm",
+    "complexity",
+    "big o",
+    "memoize",
+    "memoization",
+    "reflection",
+    "garbage collection",
+    "gc",
+    "memory leak",
     # Security (energy-related)
     "ddos",
     "vulnerability",
@@ -750,7 +1481,17 @@ GREEN_KEYWORDS = [
     "sla",
     "metric",
     "monitor",
+    "monitoring",
     "observability",
+    "observ",
+    "telemetry",
+    "apm",
+    "profiling",
+    "profile",
+    "flame graph",
+    "benchmark",
+    "rapl",
+    "watt",
     # Specific technologies
     "redis",
     "memcached",
@@ -763,6 +1504,58 @@ GREEN_KEYWORDS = [
     "graviton",
     "arm",
     "architecture",
+    "kafka",
+    "rabbitmq",
+    "message broker",
+    "pub/sub",
+    "publish",
+    "subscribe",
+    # Microservices specific
+    "service",
+    "decomposition",
+    "boundary",
+    "colocation",
+    "co-location",
+    "locality",
+    "graceful shutdown",
+    "drain",
+    "api gateway",
+    "graphql",
+    "bff",
+    # Infrastructure
+    "alpine",
+    "distroless",
+    "lightweight",
+    "minimal",
+    "terraform",
+    "cloudformation",
+    "iac",
+    "infrastructure as code",
+    "pdb",
+    "disruption budget",
+    # General optimization
+    "feature flag",
+    "toggle",
+    "incremental",
+    "delta",
+    "streaming",
+    "precompute",
+    "pre-aggregate",
+    "approximate",
+    "approximation",
+    "heuristic",
+    "sampling",
+    # Additional VU Amsterdam terms
+    "dvfs",
+    "frequency scaling",
+    "power state",
+    "dynamic allocation",
+    "adaptive",
+    "predictive",
+    "forecasting",
+    "workload",
+    "overhead",
+    "footprint",
 ]
 
 
@@ -796,7 +1589,7 @@ def is_green_aware(commit_message: str) -> bool:
         commit_message: The commit message to analyze
 
     Returns:
-        True if commit contains green keywords
+        True if commit has green keywords
     """
     message_lower = commit_message.lower()
     return any(keyword.lower() in message_lower for keyword in GREEN_KEYWORDS)