greenmining 0.1.4__py3-none-any.whl

greenmining/gsf_patterns.py

@@ -0,0 +1,802 @@
+"""
+Green Software Foundation Patterns
+Official patterns from https://patterns.greensoftware.foundation/
+
+Categories: Cloud (40+ patterns), Web (15+ patterns), AI/ML (10+ patterns)
+Total: 65+ official GSF patterns
+"""
+
+GSF_PATTERNS = {
+    # ==================== CLOUD PATTERNS (40+) ====================
+    "cache_static_data": {
+        "name": "Cache Static Data",
+        "category": "cloud",
+        "keywords": ["cache", "caching", "static", "cdn", "redis", "memcache", "cached"],
+        "description": "Cache static content to reduce server load and network transfers",
+        "sci_impact": "Reduces energy by minimizing redundant compute and network operations",
+    },
+    "choose_region_closest": {
+        "name": "Choose Region Closest to Users",
+        "category": "cloud",
+        "keywords": ["region", "closest", "proximity", "latency", "location", "geographic"],
+        "description": "Deploy in regions closest to users to reduce network distance",
+        "sci_impact": "Less energy for network transmission, lower latency",
+    },
+    "compress_stored_data": {
+        "name": "Compress Stored Data",
+        "category": "cloud",
+        "keywords": ["compress", "compression", "stored", "storage", "gzip", "zstd"],
+        "description": "Compress data at rest to reduce storage footprint",
+        "sci_impact": "Lower embodied carbon from reduced storage infrastructure",
+    },
+    "compress_transmitted_data": {
+        "name": "Compress Transmitted Data",
+        "category": "cloud",
+        "keywords": ["compress", "transmission", "gzip", "brotli", "network"],
+        "description": "Compress data before transmission over network",
+        "sci_impact": "Reduces network energy consumption and transfer time",
+    },
+    "containerize_workload": {
+        "name": "Containerize Your Workload",
+        "category": "cloud",
+        "keywords": ["container", "docker", "kubernetes", "containerize", "pod"],
+        "description": "Use containers for better resource utilization",
+        "sci_impact": "Improved resource density and efficiency",
+    },
+    "delete_unused_storage": {
+        "name": "Delete Unused Storage Resources",
+        "category": "cloud",
+        "keywords": ["delete", "remove", "unused", "storage", "cleanup", "orphan"],
+        "description": "Remove storage resources that are no longer needed",
+        "sci_impact": "Eliminates wasted energy and embodied carbon",
+    },
+    "encrypt_necessary": {
+        "name": "Encrypt What Is Necessary",
+        "category": "cloud",
+        "keywords": ["encrypt", "encryption", "tls", "ssl", "crypto"],
+        "description": "Only encrypt data that requires protection",
+        "sci_impact": "Reduces CPU overhead from unnecessary encryption",
+    },
+    "evaluate_cpu_architectures": {
+        "name": "Evaluate Other CPU Architectures",
+        "category": "cloud",
+        "keywords": ["cpu", "architecture", "arm", "graviton", "processor"],
+        "description": "Consider ARM and other efficient CPU architectures",
+        "sci_impact": "Some architectures offer better energy efficiency",
+    },
+    "service_mesh": {
+        "name": "Evaluate Using a Service Mesh",
+        "category": "cloud",
+        "keywords": ["service mesh", "istio", "linkerd", "envoy"],
+        "description": "Use service mesh for efficient service-to-service communication",
+        "sci_impact": "Optimizes network traffic and observability overhead",
+    },
+    "tls_termination": {
+        "name": "Evaluate TLS Termination",
+        "category": "cloud",
+        "keywords": ["tls", "termination", "ssl", "offload"],
+        "description": "Terminate TLS at edge to reduce encryption overhead",
+        "sci_impact": "Reduces CPU load on backend services",
+    },
+    "stateless_design": {
+        "name": "Implement Stateless Design",
+        "category": "cloud",
+        "keywords": ["stateless", "state", "session", "horizontal"],
+        "description": "Design services without server-side state",
+        "sci_impact": "Enables better scaling and resource utilization",
+    },
+    "match_slo": {
+        "name": "Match SLO Requirements",
+        "category": "cloud",
+        "keywords": ["slo", "sla", "service level", "objective"],
+        "description": "Don't over-engineer beyond required SLO",
+        "sci_impact": "Avoids wasted resources on unnecessary performance",
+    },
+    "match_vm_utilization": {
+        "name": "Match VM Utilization Requirements",
+        "category": "cloud",
+        "keywords": ["vm", "virtual machine", "instance", "size", "utilization"],
+        "description": "Right-size VMs to actual workload needs",
+        "sci_impact": "Eliminates idle resources and wasted energy",
+    },
+    "minimize_environments": {
+        "name": "Minimize Deployed Environments",
+        "category": "cloud",
+        "keywords": ["environment", "dev", "staging", "ephemeral"],
+        "description": "Reduce number of permanent test/dev environments",
+        "sci_impact": "Less idle infrastructure consuming energy",
+    },
+    "optimize_storage_utilization": {
+        "name": "Optimize Storage Resource Utilization",
+        "category": "cloud",
+        "keywords": ["storage", "optimize", "utilization", "efficiency"],
+        "description": "Maximize usage of provisioned storage",
+        "sci_impact": "Better hardware efficiency, less waste",
+    },
+    "optimize_avg_cpu": {
+        "name": "Optimize Average CPU Utilization",
+        "category": "cloud",
+        "keywords": ["cpu", "utilization", "average", "optimize"],
+        "description": "Increase average CPU utilization across fleet",
+        "sci_impact": "More work per unit of energy consumed",
+    },
+    "optimize_peak_cpu": {
+        "name": "Optimize Peak CPU Utilization",
+        "category": "cloud",
+        "keywords": ["cpu", "peak", "spike", "burst"],
+        "description": "Handle peaks efficiently without over-provisioning",
+        "sci_impact": "Reduces need for excess capacity",
+    },
+    "queue_non_urgent": {
+        "name": "Queue Non-Urgent Requests",
+        "category": "cloud",
+        "keywords": ["queue", "async", "batch", "defer", "background"],
+        "description": "Queue non-urgent work for batch processing",
+        "sci_impact": "Better resource utilization and scheduling",
+    },
+    "reduce_network_traversal": {
+        "name": "Reduce Network Traversal Between VMs",
+        "category": "cloud",
+        "keywords": ["network", "traversal", "hop", "latency", "cross-az"],
+        "description": "Minimize network hops between services",
+        "sci_impact": "Less network equipment and energy per request",
+    },
+    "reduce_transmitted_data": {
+        "name": "Reduce Transmitted Data",
+        "category": "cloud",
+        "keywords": ["reduce", "data", "transmission", "payload", "bandwidth"],
+        "description": "Send only necessary data over network",
+        "sci_impact": "Lower network energy consumption",
+    },
+    "remove_unused_assets": {
+        "name": "Remove Unused Assets",
+        "category": "cloud",
+        "keywords": ["remove", "unused", "asset", "cleanup", "garbage"],
+        "description": "Delete unused code, images, dependencies",
+        "sci_impact": "Reduces storage and deployment footprint",
+    },
+    "scale_down_k8s": {
+        "name": "Scale Down Kubernetes Workloads",
+        "category": "cloud",
+        "keywords": ["scale down", "kubernetes", "k8s", "downscale", "hpa"],
+        "description": "Automatically scale down idle K8s workloads",
+        "sci_impact": "Eliminates idle resource consumption",
+    },
+    "scale_down_apps": {
+        "name": "Scale Down Unused Applications",
+        "category": "cloud",
+        "keywords": ["scale", "down", "idle", "suspend", "hibernate"],
+        "description": "Scale down or shut down idle applications",
+        "sci_impact": "Reduces energy from idle services",
+    },
+    "scale_with_load": {
+        "name": "Scale Infrastructure with User Load",
+        "category": "cloud",
+        "keywords": ["autoscale", "scale", "elastic", "demand", "load"],
+        "description": "Automatically scale based on actual demand",
+        "sci_impact": "Matches resources to need, eliminates waste",
+    },
+    "k8s_event_scaling": {
+        "name": "Scale Kubernetes Workloads Based on Events",
+        "category": "cloud",
+        "keywords": ["keda", "event", "scale", "kubernetes", "queue"],
+        "description": "Scale based on queue depth or event metrics",
+        "sci_impact": "Precise scaling to actual workload",
+    },
+    "scale_logical_components": {
+        "name": "Scale Logical Components Independently",
+        "category": "cloud",
+        "keywords": ["microservice", "component", "independent", "decouple"],
+        "description": "Scale services independently based on their load",
+        "sci_impact": "Avoids scaling entire application for one bottleneck",
+    },
+    "scan_vulnerabilities": {
+        "name": "Scan for Vulnerabilities",
+        "category": "cloud",
+        "keywords": ["security", "vulnerability", "scan", "cve"],
+        "description": "Regular security scans to prevent breaches",
+        "sci_impact": "Prevents energy waste from attacks and remediation",
+    },
+    "retention_policy": {
+        "name": "Set Retention Policy on Storage",
+        "category": "cloud",
+        "keywords": ["retention", "policy", "lifecycle", "expire", "ttl"],
+        "description": "Automatically delete old data per policy",
+        "sci_impact": "Reduces storage footprint over time",
+    },
+    "shed_lower_priority": {
+        "name": "Shed Lower Priority Traffic",
+        "category": "cloud",
+        "keywords": ["shed", "priority", "throttle", "rate limit", "backpressure"],
+        "description": "Drop low-priority requests under load",
+        "sci_impact": "Protects resources for important work",
+    },
+    "time_shift_cron": {
+        "name": "Time-Shift Kubernetes Cron Jobs",
+        "category": "cloud",
+        "keywords": ["time shift", "cron", "schedule", "carbon aware", "renewable"],
+        "description": "Run batch jobs when renewable energy is available",
+        "sci_impact": "Uses cleaner energy sources",
+    },
+    "async_not_sync": {
+        "name": "Use Async Instead of Sync",
+        "category": "cloud",
+        "keywords": ["async", "asynchronous", "non-blocking", "await"],
+        "description": "Use asynchronous processing patterns",
+        "sci_impact": "Better CPU utilization, less idle time",
+    },
+    "circuit_breaker": {
+        "name": "Use Circuit Breaker",
+        "category": "cloud",
+        "keywords": ["circuit breaker", "fault tolerance", "resilience", "fallback"],
+        "description": "Prevent cascading failures with circuit breakers",
+        "sci_impact": "Avoids wasted energy on failing requests",
+    },
+    "cloud_native_security": {
+        "name": "Use Cloud Native Security Tools",
+        "category": "cloud",
+        "keywords": ["security", "native", "managed", "cloud"],
+        "description": "Use cloud provider's native security services",
+        "sci_impact": "More efficient than running own security infrastructure",
+    },
+    "compiled_languages": {
+        "name": "Use Compiled Languages",
+        "category": "cloud",
+        "keywords": ["compile", "compiled", "rust", "go", "c++"],
+        "description": "Consider compiled languages for CPU-intensive tasks",
+        "sci_impact": "Better runtime performance and energy efficiency",
+    },
+    "ddos_protection": {
+        "name": "Use DDoS Protection",
+        "category": "cloud",
+        "keywords": ["ddos", "protection", "shield", "waf"],
+        "description": "Protect against distributed denial of service",
+        "sci_impact": "Prevents energy waste from malicious traffic",
+    },
+    "energy_efficient_hardware": {
+        "name": "Use Energy Efficient Hardware",
+        "category": "cloud",
+        "keywords": ["energy", "efficient", "hardware", "green", "sustainable"],
+        "description": "Choose hardware optimized for energy efficiency",
+        "sci_impact": "Direct reduction in energy consumption",
+    },
+    # ==================== WEB PATTERNS (15+) ====================
+    "avoid_chaining_requests": {
+        "name": "Avoid Chaining Critical Requests",
+        "category": "web",
+        "keywords": ["chain", "critical", "request", "waterfall", "sequential"],
+        "description": "Avoid serialized network requests that block rendering",
+        "sci_impact": "Faster page load, less energy for rendering",
+    },
+    "avoid_excessive_dom": {
+        "name": "Avoid Excessive DOM Size",
+        "category": "web",
+        "keywords": ["dom", "size", "excessive", "tree", "nodes"],
+        "description": "Keep DOM tree small and shallow",
+        "sci_impact": "Reduces memory and rendering energy",
+    },
+    "avoid_tracking_unnecessary": {
+        "name": "Avoid Tracking Unnecessary Data",
+        "category": "web",
+        "keywords": ["tracking", "analytics", "telemetry", "unnecessary"],
+        "description": "Don't track data you won't use",
+        "sci_impact": "Reduces network and storage overhead",
+    },
+    "defer_offscreen_images": {
+        "name": "Defer Offscreen Images",
+        "category": "web",
+        "keywords": ["defer", "offscreen", "image", "lazy load", "viewport"],
+        "description": "Load images only when they enter viewport",
+        "sci_impact": "Reduces initial load energy and bandwidth",
+    },
+    "deprecate_gifs": {
+        "name": "Deprecate GIFs",
+        "category": "web",
+        "keywords": ["gif", "video", "webm", "mp4", "animation"],
+        "description": "Use video formats instead of animated GIFs",
+        "sci_impact": "Video formats are much more efficient than GIF",
+    },
+    "enable_text_compression": {
+        "name": "Enable Text Compression",
+        "category": "web",
+        "keywords": ["compress", "text", "gzip", "brotli", "minify"],
+        "description": "Compress HTML, CSS, JS before transmission",
+        "sci_impact": "Reduces bandwidth and transfer energy",
+    },
+    "keep_request_counts_low": {
+        "name": "Keep Request Counts Low",
+        "category": "web",
+        "keywords": ["request", "count", "http", "reduce", "combine"],
+        "description": "Minimize number of HTTP requests",
+        "sci_impact": "Lower network overhead per page load",
+    },
+    "minify_web_assets": {
+        "name": "Minify Web Assets",
+        "category": "web",
+        "keywords": ["minify", "minification", "uglify", "compress"],
+        "description": "Remove whitespace and unnecessary code",
+        "sci_impact": "Smaller files, less bandwidth",
+    },
+    "minimize_main_thread": {
+        "name": "Minimize Main Thread Work",
+        "category": "web",
+        "keywords": ["main thread", "worker", "offload", "javascript"],
+        "description": "Offload work from main thread to web workers",
+        "sci_impact": "Better CPU utilization, faster page loads",
+    },
+    "properly_sized_images": {
+        "name": "Properly Sized Images",
+        "category": "web",
+        "keywords": ["image", "size", "resize", "responsive", "srcset"],
+        "description": "Serve images at correct display size",
+        "sci_impact": "Avoids transferring and processing oversized images",
+    },
+    "remove_unused_css": {
+        "name": "Remove Unused CSS",
+        "category": "web",
+        "keywords": ["css", "unused", "purge", "tree shake"],
+        "description": "Eliminate CSS that isn't used on page",
+        "sci_impact": "Smaller CSS files, faster parsing",
+    },
+    "serve_modern_image_formats": {
+        "name": "Serve Images in Modern Formats",
+        "category": "web",
+        "keywords": ["webp", "avif", "image", "format", "modern"],
+        "description": "Use WebP or AVIF instead of JPEG/PNG",
+        "sci_impact": "Modern formats are significantly more efficient",
+    },
+    "server_side_rendering": {
+        "name": "Use Server-Side Rendering",
+        "category": "web",
+        "keywords": ["ssr", "server side", "rendering", "nextjs"],
+        "description": "Pre-render pages on server when appropriate",
+        "sci_impact": "Reduces client-side computation energy",
+    },
+    # ==================== AI/ML PATTERNS (10+) ====================
+    "compress_ml_models": {
+        "name": "Compress ML Models for Inference",
+        "category": "ai",
+        "keywords": ["compress", "model", "quantiz", "prune", "distill"],
+        "description": "Reduce model size through quantization, pruning, distillation",
+        "sci_impact": "Dramatically reduces inference energy and memory",
+    },
+    "efficient_format_training": {
+        "name": "Efficient Format for Model Training",
+        "category": "ai",
+        "keywords": ["format", "training", "tfrecord", "parquet", "efficient"],
+        "description": "Use efficient data formats for training",
+        "sci_impact": "Faster I/O, less storage, quicker training",
+    },
+    "energy_efficient_ai_edge": {
+        "name": "Energy Efficient AI at Edge",
+        "category": "ai",
+        "keywords": ["edge", "ai", "inference", "local", "device"],
+        "description": "Run inference on edge devices when possible",
+        "sci_impact": "Eliminates network transfer, uses local compute",
+    },
+    "energy_efficient_framework": {
+        "name": "Energy Efficient Framework",
+        "category": "ai",
+        "keywords": ["framework", "tensorflow", "pytorch", "efficient"],
+        "description": "Choose ML frameworks optimized for efficiency",
+        "sci_impact": "Different frameworks have different energy profiles",
+    },
+    "energy_efficient_models": {
+        "name": "Energy Efficient Models",
+        "category": "ai",
+        "keywords": ["model", "efficient", "mobilenet", "efficientnet"],
+        "description": "Use models designed for efficiency (e.g., MobileNet)",
+        "sci_impact": "Purpose-built efficient architectures",
+    },
+    "leverage_sustainable_regions_ai": {
+        "name": "Leverage Sustainable Regions for AI",
+        "category": "ai",
+        "keywords": ["region", "sustainable", "renewable", "carbon"],
+        "description": "Train models in regions with clean energy",
+        "sci_impact": "Lower carbon intensity for training",
+    },
+    "pretrained_transfer_learning": {
+        "name": "Pre-trained Transfer Learning",
+        "category": "ai",
+        "keywords": ["transfer", "learning", "pretrain", "fine-tune"],
+        "description": "Start from pre-trained models instead of training from scratch",
+        "sci_impact": "Avoids massive energy cost of full training",
+    },
+    "right_hardware_ai": {
+        "name": "Right Hardware Type for AI",
+        "category": "ai",
+        "keywords": ["hardware", "gpu", "tpu", "accelerator", "ai"],
+        "description": "Use appropriate hardware (GPU/TPU) for AI workloads",
+        "sci_impact": "Specialized hardware is more energy efficient",
+    },
+    "serverless_ml": {
+        "name": "Serverless Model Development",
+        "category": "ai",
+        "keywords": ["serverless", "ml", "sagemaker", "vertex", "lambda"],
+        "description": "Use serverless platforms for ML development",
+        "sci_impact": "Pay-per-use, no idle resources",
+    },
+    # ==================== GENERAL PATTERNS ====================
+    "autoscaling": {
+        "name": "Autoscaling",
+        "category": "cloud",
+        "keywords": ["autoscal", "scale", "elastic", "horizontal scaling", "hpa", "scaling policy"],
+        "description": "Automatically scale resources based on demand",
+        "sci_impact": "Reduces carbon emissions by matching resources to actual demand",
+    },
+    "optimize_storage": {
+        "name": "Optimize Storage",
+        "category": "cloud",
+        "keywords": ["compress", "deduplicate", "archive", "storage tier", "lifecycle policy"],
+        "description": "Reduce storage footprint through compression and deduplication",
+        "sci_impact": "Lower embodied carbon through reduced storage infrastructure",
+    },
+    "serverless": {
+        "name": "Use Serverless",
+        "category": "cloud",
+        "keywords": ["serverless", "lambda", "function", "faas", "cloud function"],
+        "description": "Use serverless computing for event-driven workloads",
+        "sci_impact": "Pay-per-use model ensures zero idle resource consumption",
+    },
+    "right_sizing": {
+        "name": "Right-size Resources",
+        "category": "cloud",
+        "keywords": [
+            "right-size",
+            "rightsize",
+            "downsize",
+            "optimize instance",
+            "resource optimization",
+        ],
+        "description": "Match compute resources to actual workload requirements",
+        "sci_impact": "Eliminates over-provisioning and reduces wasted energy",
+    },
+    # WEB PATTERNS
+    "lazy_loading": {
+        "name": "Lazy Loading",
+        "category": "web",
+        "keywords": ["lazy load", "lazy", "defer", "async", "on-demand"],
+        "description": "Load resources only when needed",
+        "sci_impact": "Reduces unnecessary data transfer and processing",
+    },
+    "minimize_data_transfer": {
+        "name": "Minimize Data Transfer",
+        "category": "web",
+        "keywords": ["minif", "compress", "gzip", "brotli", "optimize payload", "reduce bundle"],
+        "description": "Reduce size of data transferred over network",
+        "sci_impact": "Lower network energy consumption and faster load times",
+    },
+    "optimize_images": {
+        "name": "Optimize Images",
+        "category": "web",
+        "keywords": ["image optim", "webp", "responsive image", "lazy image", "compress image"],
+        "description": "Use efficient image formats and sizing",
+        "sci_impact": "Reduces bandwidth and storage requirements",
+    },
+    "reduce_http_requests": {
+        "name": "Reduce HTTP Requests",
+        "category": "web",
+        "keywords": ["bundle", "sprite", "inline", "combine", "reduce request"],
+        "description": "Minimize number of network requests",
+        "sci_impact": "Lower network overhead and latency",
+    },
+    # AI/ML PATTERNS
+    "model_optimization": {
+        "name": "Optimize ML Models",
+        "category": "ai",
+        "keywords": ["quantiz", "prune", "distill", "model compress", "optimize model"],
+        "description": "Reduce model size and complexity",
+        "sci_impact": "Dramatically reduces compute and memory requirements",
+    },
+    "batch_inference": {
+        "name": "Batch Inference",
+        "category": "ai",
+        "keywords": ["batch inference", "batch predict", "batch processing"],
+        "description": "Process ML predictions in batches",
+        "sci_impact": "More efficient GPU utilization",
+    },
+    # NETWORKING PATTERNS
+    "connection_pooling": {
+        "name": "Connection Pooling",
+        "category": "networking",
+        "keywords": ["connection pool", "pool", "reuse connection", "persistent connection"],
+        "description": "Reuse network connections instead of creating new ones",
+        "sci_impact": "Reduces connection overhead and resource consumption",
+    },
+    "rate_limiting": {
+        "name": "Rate Limiting",
+        "category": "networking",
+        "keywords": ["rate limit", "throttle", "backpressure", "circuit breaker"],
+        "description": "Control request rate to prevent resource exhaustion",
+        "sci_impact": "Prevents waste from processing excessive or malicious traffic",
+    },
+    # DATABASE PATTERNS
+    "database_indexing": {
+        "name": "Database Indexing",
+        "category": "database",
+        "keywords": ["index", "query optim", "explain", "slow query"],
+        "description": "Optimize database queries with proper indexing",
+        "sci_impact": "Reduces compute cycles for data retrieval",
+    },
+    "data_partitioning": {
+        "name": "Data Partitioning",
+        "category": "database",
+        "keywords": ["partition", "shard", "segment", "distribute data"],
+        "description": "Distribute data across multiple nodes",
+        "sci_impact": "Improves query efficiency and resource utilization",
+    },
+    # GENERAL PATTERNS
+    "async_processing": {
+        "name": "Asynchronous Processing",
+        "category": "general",
+        "keywords": ["async", "asynchronous", "non-blocking", "event-driven", "queue"],
+        "description": "Process tasks asynchronously to improve resource utilization",
+        "sci_impact": "Better CPU utilization and reduced idle time",
+    },
+    "resource_pooling": {
+        "name": "Resource Pooling",
+        "category": "general",
+        "keywords": ["thread pool", "worker pool", "object pool", "resource pool"],
+        "description": "Reuse expensive resources instead of creating new ones",
+        "sci_impact": "Reduces initialization overhead and memory pressure",
+    },
+    "memoization": {
+        "name": "Memoization",
+        "category": "general",
+        "keywords": ["memoize", "memoization", "cache result", "remember"],
+        "description": "Cache results of expensive function calls",
+        "sci_impact": "Avoids redundant computation",
+    },
+    "batch_processing": {
+        "name": "Batch Processing",
+        "category": "general",
+        "keywords": ["batch", "bulk", "batch insert", "batch update"],
+        "description": "Process multiple items together instead of one at a time",
+        "sci_impact": "Reduces per-item overhead and improves throughput",
+    },
+}
+
+# Green software keywords (comprehensive list from all GSF patterns)
+GREEN_KEYWORDS = [
+    # Core sustainability terms
+    "energy",
+    "power",
+    "carbon",
+    "emission",
+    "footprint",
+    "sustainability",
+    "sustainable",
+    "green",
+    "efficient",
+    "efficiency",
+    "eco",
+    "environment",
+    "renewable",
+    # Performance & optimization
+    "optimize",
+    "optimization",
+    "optimise",
+    "optimisation",
+    "performance",
+    "performant",
+    "fast",
+    "faster",
+    "speed",
+    "speedup",
+    "latency",
+    "throughput",
+    "utilization",
+    # Caching & storage
+    "cache",
+    "caching",
+    "cached",
+    "redis",
+    "memcache",
+    "memcached",
+    "cdn",
+    "storage",
+    "compress",
+    "compression",
+    "compressed",
+    "gzip",
+    "brotli",
+    "zstd",
+    "deduplicate",
+    "dedup",
+    "archive",
+    "retention",
+    # Resource management
+    "pool",
+    "pooling",
+    "connection pool",
+    "thread pool",
+    "worker pool",
+    "resource pool",
+    "reuse",
+    "recycle",
+    "cleanup",
+    "garbage",
+    "remove unused",
+    "delete unused",
+    # Async & concurrency
+    "async",
+    "asynchronous",
+    "non-blocking",
+    "await",
+    "promise",
+    "queue",
+    "background",
+    "worker",
+    "job",
+    "batch",
+    "batching",
+    # Scaling & sizing
+    "scale",
+    "scaling",
+    "autoscale",
+    "autoscaling",
+    "elastic",
+    "elasticity",
+    "horizontal",
+    "vertical",
+    "right-size",
+    "rightsize",
+    "downsize",
+    "upsize",
+    "hpa",
+    "keda",
+    "demand",
+    "load",
+    # Cloud native
+    "serverless",
+    "lambda",
+    "function",
+    "faas",
+    "cloud function",
+    "container",
+    "docker",
+    "kubernetes",
+    "k8s",
+    "pod",
+    "microservice",
+    "service mesh",
+    "istio",
+    "envoy",
+    # Network optimization
+    "network",
+    "bandwidth",
+    "transmission",
+    "transfer",
+    "latency",
+    "proximity",
+    "region",
+    "closest",
+    "edge",
+    "cdn",
+    "rate limit",
+    "throttle",
+    "backpressure",
+    "circuit breaker",
+    "shed",
+    "reduce request",
+    "minimize request",
+    # Data & database
+    "index",
+    "indexing",
+    "query",
+    "optimize query",
+    "query optimization",
+    "partition",
+    "shard",
+    "database",
+    "db",
+    "sql",
+    "nosql",
+    # Web optimization
+    "minify",
+    "minification",
+    "uglify",
+    "tree shake",
+    "bundle",
+    "lazy load",
+    "defer",
+    "offscreen",
+    "viewport",
+    "dom",
+    "ssr",
+    "server side rendering",
+    "hydration",
+    "webp",
+    "avif",
+    "image",
+    "responsive",
+    # AI/ML
+    "model",
+    "quantiz",
+    "prune",
+    "pruning",
+    "distill",
+    "distillation",
+    "inference",
+    "training",
+    "pretrain",
+    "transfer learning",
+    "gpu",
+    "tpu",
+    "accelerator",
+    "mobilenet",
+    "efficientnet",
+    # Code & architecture
+    "stateless",
+    "decouple",
+    "independent",
+    "component",
+    "compiled",
+    "rust",
+    "go",
+    "c++",
+    "refactor",
+    "dead code",
+    "unused",
+    "deprecate",
+    # Security (energy-related)
+    "ddos",
+    "vulnerability",
+    "scan",
+    "security",
+    "waf",
+    # Monitoring & SLO
+    "slo",
+    "sla",
+    "metric",
+    "monitor",
+    "observability",
+    # Specific technologies
+    "redis",
+    "memcached",
+    "nginx",
+    "envoy",
+    "linkerd",
+    "karpenter",
+    "cluster autoscaler",
+    "spot instance",
+    "graviton",
+    "arm",
+    "architecture",
+]
+
+
+def get_pattern_by_keywords(commit_message: str) -> list:
+    """
+    Match commit message against GSF patterns.
+
+    Args:
+        commit_message: The commit message to analyze
+
+    Returns:
+        List of matched pattern names
+    """
+    message_lower = commit_message.lower()
+    matched_patterns = []
+
+    for _pattern_id, pattern in GSF_PATTERNS.items():
+        for keyword in pattern["keywords"]:
+            if keyword.lower() in message_lower:
+                matched_patterns.append(pattern["name"])
+                break
+
+    return matched_patterns
+
+
+def is_green_aware(commit_message: str) -> bool:
+    """
+    Check if commit shows green software awareness.
+
+    Args:
+        commit_message: The commit message to analyze
+
+    Returns:
+        True if commit contains green keywords
+    """
+    message_lower = commit_message.lower()
+    return any(keyword.lower() in message_lower for keyword in GREEN_KEYWORDS)