scirs2-ndimage 0.4.2

//! Processing Strategies and Algorithm Management
//!
//! This module handles processing strategies, algorithm steps, and strategy selection
//! for the AI-driven adaptive processing system.

use scirs2_core::ndarray::Array1;
use std::collections::HashMap;

use super::config::{AlgorithmType, ImagePattern};
use super::learning::PerformanceMetrics;

/// Processing Strategy
#[derive(Debug, Clone)]
pub struct ProcessingStrategy {
    /// Algorithm sequence
    pub algorithm_sequence: Vec<AlgorithmStep>,
    /// Parameter settings
    pub parameters: HashMap<String, f64>,
    /// Expected performance
    pub expected_performance: PerformanceMetrics,
    /// Confidence level
    pub confidence: f64,
    /// Usage count (for popularity-based selection)
    pub usage_count: usize,
    /// Success rate
    pub success_rate: f64,
}

/// Algorithm Step
#[derive(Debug, Clone)]
pub struct AlgorithmStep {
    /// Algorithm type
    pub algorithm: AlgorithmType,
    /// Parameters for this step
    pub parameters: HashMap<String, f64>,
    /// Expected contribution to quality
    pub quality_contribution: f64,
    /// Computational cost
    pub computational_cost: f64,
}

/// Processing Algorithm Variants
#[derive(Debug, Clone)]
pub enum ProcessingAlgorithm {
    AdaptiveGaussianFilter,
    IntelligentEdgeDetection,
    AIEnhancedMedianFilter,
    SmartBilateralFilter,
    ContextAwareNoiseReduction,
    AdaptiveMorphology,
    IntelligentSegmentation,
    AIFeatureExtraction,
}

/// Adaptation Strategy
#[derive(Debug, Clone)]
pub struct AdaptationStrategy {
    /// Strategy name
    pub name: String,
    /// Adaptation parameters
    pub parameters: Array1<f64>,
    /// Adaptation speed
    pub speed: f64,
    /// Effectiveness score
    pub effectiveness: f64,
}

/// Performance Record
#[derive(Debug, Clone)]
pub struct PerformanceRecord {
    /// Timestamp
    pub timestamp: u64,
    /// Input characteristics
    pub input_characteristics: Array1<f64>,
    /// Applied strategy
    pub strategy_used: ProcessingStrategy,
    /// Achieved metrics
    pub achievedmetrics: PerformanceMetrics,
    /// Context information
    pub context: String,
}

impl ProcessingStrategy {
    /// Create a new processing strategy
    pub fn new() -> Self {
        Self {
            algorithm_sequence: Vec::new(),
            parameters: HashMap::new(),
            expected_performance: PerformanceMetrics {
                speed: 0.0,
                quality: 0.0,
                memory_usage: 0.0,
                energy_consumption: 0.0,
                user_satisfaction: None,
            },
            confidence: 0.0,
            usage_count: 0,
            success_rate: 0.0,
        }
    }

    /// Add an algorithm step to the strategy
    pub fn add_step(&mut self, step: AlgorithmStep) {
        self.algorithm_sequence.push(step);
    }

    /// Update strategy performance based on results
    pub fn update_performance(&mut self, metrics: &PerformanceMetrics, success: bool) {
        self.usage_count += 1;

        if success {
            // Update success rate
            let total_attempts = self.usage_count as f64;
            let previous_successes = (total_attempts - 1.0) * self.success_rate;
            self.success_rate = (previous_successes + 1.0) / total_attempts;

            // Update expected performance with exponential moving average
            let alpha = 0.1; // Learning rate
            self.expected_performance.speed =
                (1.0 - alpha) * self.expected_performance.speed + alpha * metrics.speed;
            self.expected_performance.quality =
                (1.0 - alpha) * self.expected_performance.quality + alpha * metrics.quality;
            self.expected_performance.memory_usage = (1.0 - alpha)
                * self.expected_performance.memory_usage
                + alpha * metrics.memory_usage;
        }
    }

    /// Calculate strategy score for selection
    pub fn calculate_score(&self, pattern: &ImagePattern) -> f64 {
        // Combine multiple factors for strategy selection
        let performance_score = self.expected_performance.quality;
        let confidence_score = self.confidence;
        let usage_score = (self.usage_count as f64).ln().max(0.0) / 10.0; // Log of usage count
        let success_score = self.success_rate;

        // Weighted combination
        0.4 * performance_score + 0.3 * confidence_score + 0.1 * usage_score + 0.2 * success_score
    }
}

impl AlgorithmStep {
    /// Create a new algorithm step
    pub fn new(algorithm: AlgorithmType) -> Self {
        Self {
            algorithm,
            parameters: HashMap::new(),
            quality_contribution: 0.0,
            computational_cost: 1.0,
        }
    }

    /// Set a parameter for this step
    pub fn set_parameter(&mut self, name: &str, value: f64) {
        self.parameters.insert(name.to_string(), value);
    }

    /// Get estimated execution time
    pub fn estimated_time(&self) -> f64 {
        // Base time multiplied by computational cost
        let base_time = match self.algorithm {
            AlgorithmType::GaussianFilter => 0.1,
            AlgorithmType::MedianFilter => 0.3,
            AlgorithmType::BilateralFilter => 0.5,
            AlgorithmType::EdgeDetection => 0.2,
            AlgorithmType::MorphologyOperation => 0.25,
            AlgorithmType::QuantumProcessing => 1.0,
            AlgorithmType::NeuromorphicProcessing => 0.8,
            AlgorithmType::ConsciousnessSimulation => 1.5,
            AlgorithmType::AdvancedFusion => 1.2,
            AlgorithmType::CustomAI => 0.7,
        };

        base_time * self.computational_cost
    }
}