hakorune/src/mir/loop_pattern_detection/condition_var_analyzer.rs

//! Phase 170-D-impl-2: Condition Variable Analyzer Box
//!
//! Pure functions for analyzing condition AST nodes and determining
//! variable scopes based on LoopScopeShape information.
//!
//! This Box extracts all variable references from condition expressions
//! and determines whether they are from the loop parameter, outer scope,
//! or loop body scope.
//!
//! # Design Philosophy
//!
//! - **Pure functions**: No side effects, only analysis
//! - **Composable**: Can be used independently or as part of LoopConditionScopeBox
//! - **Fail-Fast**: Defaults to conservative classification (LoopBodyLocal)

use crate::ast::ASTNode;
use crate::mir::join_ir::lowering::loop_scope_shape::LoopScopeShape;
use crate::mir::BasicBlockId;
use std::collections::HashSet;

/// Extract all variable names from an AST expression
///
/// Recursively traverses the AST node and collects all Variable references.
/// Handles: Variables, UnaryOp, BinaryOp, MethodCall, FieldAccess, Index, If
///
/// # Arguments
///
/// * `node` - AST node to analyze
///
/// # Returns
///
/// HashSet of all variable names found in the expression
///
/// # Example
///
/// For expression `(i < 10) && (ch != ' ')`, returns `{"i", "ch"}`
pub fn extract_all_variables(node: &ASTNode) -> HashSet<String> {
    let mut vars = HashSet::new();
    extract_vars_recursive(node, &mut vars);
    vars
}

/// Internal recursive helper for variable extraction
fn extract_vars_recursive(node: &ASTNode, vars: &mut HashSet<String>) {
    match node {
        ASTNode::Variable { name, .. } => {
            vars.insert(name.clone());
        }
        ASTNode::UnaryOp { operand, .. } => {
            extract_vars_recursive(operand, vars);
        }
        ASTNode::BinaryOp { left, right, .. } => {
            extract_vars_recursive(left, vars);
            extract_vars_recursive(right, vars);
        }
        ASTNode::MethodCall {
            object, arguments, ..
        } => {
            extract_vars_recursive(object, vars);
            for arg in arguments {
                extract_vars_recursive(arg, vars);
            }
        }
        ASTNode::FieldAccess { object, .. } => {
            extract_vars_recursive(object, vars);
        }
        ASTNode::Index { target, index, .. } => {
            extract_vars_recursive(target, vars);
            extract_vars_recursive(index, vars);
        }
        ASTNode::If {
            condition,
            then_body,
            else_body,
            ..
        } => {
            extract_vars_recursive(condition, vars);
            for stmt in then_body {
                extract_vars_recursive(stmt, vars);
            }
            if let Some(else_body) = else_body {
                for stmt in else_body {
                    extract_vars_recursive(stmt, vars);
                }
            }
        }
        _ => {} // Skip literals, constants, etc.
    }
}

/// Determine if a variable is from the outer scope
///
/// # Phase 170-D-impl-2: Simple heuristic
///
/// A variable is "outer local" if:
/// 1. It appears in LoopScopeShape.pinned (loop parameters or outer variables)
/// 2. It appears in LoopScopeShape.variable_definitions as being defined
///    in the header block (NOT in body/latch/exit)
///
/// # Arguments
///
/// * `var_name` - Name of the variable to check
/// * `scope` - Optional LoopScopeShape with variable definition information
///
/// # Returns
///
/// - `true` if variable is definitively from outer scope
/// - `false` if unknown, from body scope, or no scope info available
///
/// # Notes
///
/// This is a simplified implementation for Phase 170-D.
/// Future versions may include:
/// - Dominance tree analysis
/// - More sophisticated scope inference
pub fn is_outer_scope_variable(
    var_name: &str,
    scope: Option<&LoopScopeShape>,
) -> bool {
    match scope {
        None => false, // No scope info → assume body-local
        Some(scope) => {
            // Check 1: Is it a pinned variable (loop parameter or passed-in)?
            if scope.pinned.contains(var_name) {
                return true;
            }

            // Check 2: Does it appear in variable_definitions and is it from header only?
            if let Some(def_blocks) = scope.variable_definitions.get(var_name) {
                // Simplified heuristic: if defined ONLY in header block → outer scope
                // Phase 170-D: This is conservative - we only accept variables
                // that are EXCLUSIVELY defined in the header (before loop enters)
                if def_blocks.len() == 1 && def_blocks.contains(&scope.header) {
                    return true;
                }
            }

            false
        }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    // Helper: Create a Variable node
    fn var_node(name: &str) -> ASTNode {
        ASTNode::Variable {
            name: name.to_string(),
            span: crate::ast::Span::unknown(),
        }
    }

    // Helper: Create a BinaryOp node
    fn binop_node(left: ASTNode, right: ASTNode) -> ASTNode {
        ASTNode::BinaryOp {
            operator: crate::ast::BinaryOperator::Add, // Placeholder operator
            left: Box::new(left),
            right: Box::new(right),
            span: crate::ast::Span::unknown(),
        }
    }

    // Helper: Create a UnaryOp node
    fn unary_node(operand: ASTNode) -> ASTNode {
        ASTNode::UnaryOp {
            operator: crate::ast::UnaryOperator::Not,
            operand: Box::new(operand),
            span: crate::ast::Span::unknown(),
        }
    }

    #[test]
    fn test_extract_single_variable() {
        let node = var_node("x");
        let vars = extract_all_variables(&node);
        assert_eq!(vars.len(), 1);
        assert!(vars.contains("x"));
    }

    #[test]
    fn test_extract_multiple_variables() {
        let node = binop_node(var_node("x"), var_node("y"));
        let vars = extract_all_variables(&node);
        assert_eq!(vars.len(), 2);
        assert!(vars.contains("x"));
        assert!(vars.contains("y"));
    }

    #[test]
    fn test_extract_deduplicated_variables() {
        let node = binop_node(var_node("x"), var_node("x"));
        let vars = extract_all_variables(&node);
        assert_eq!(vars.len(), 1); // HashSet deduplicates
        assert!(vars.contains("x"));
    }

    #[test]
    fn test_extract_nested_variables() {
        // Create (x + y) + z structure
        let inner = binop_node(var_node("x"), var_node("y"));
        let outer = binop_node(inner, var_node("z"));
        let vars = extract_all_variables(&outer);

        assert_eq!(vars.len(), 3);
        assert!(vars.contains("x"));
        assert!(vars.contains("y"));
        assert!(vars.contains("z"));
    }

    #[test]
    fn test_extract_with_unary_op() {
        // Create !(x) structure
        let node = unary_node(var_node("x"));
        let vars = extract_all_variables(&node);

        assert_eq!(vars.len(), 1);
        assert!(vars.contains("x"));
    }

    #[test]
    fn test_extract_no_variables_from_literal() {
        let node = ASTNode::Literal {
            value: crate::ast::LiteralValue::Integer(42),
            span: crate::ast::Span::unknown(),
        };
        let vars = extract_all_variables(&node);

        assert!(vars.is_empty());
    }

    #[test]
    fn test_is_outer_scope_variable_with_no_scope() {
        let result = is_outer_scope_variable("x", None);
        assert!(!result); // No scope → assume body-local
    }

    #[test]
    fn test_is_outer_scope_variable_pinned() {
        use std::collections::{BTreeMap, BTreeSet};

        let mut pinned = BTreeSet::new();
        pinned.insert("len".to_string());

        let scope = LoopScopeShape {
            header: BasicBlockId(0),
            body: BasicBlockId(1),
            latch: BasicBlockId(2),
            exit: BasicBlockId(3),
            pinned,
            carriers: BTreeSet::new(),
            body_locals: BTreeSet::new(),
            exit_live: BTreeSet::new(),
            progress_carrier: None,
            variable_definitions: BTreeMap::new(),
        };

        assert!(is_outer_scope_variable("len", Some(&scope)));
        assert!(!is_outer_scope_variable("unknown", Some(&scope)));
    }

    #[test]
    fn test_is_outer_scope_variable_from_header_only() {
        use std::collections::{BTreeMap, BTreeSet};

        let mut variable_definitions = BTreeMap::new();
        let mut header_only = BTreeSet::new();
        header_only.insert(BasicBlockId(0)); // Only in header

        variable_definitions.insert("start".to_string(), header_only);

        let scope = LoopScopeShape {
            header: BasicBlockId(0),
            body: BasicBlockId(1),
            latch: BasicBlockId(2),
            exit: BasicBlockId(3),
            pinned: BTreeSet::new(),
            carriers: BTreeSet::new(),
            body_locals: BTreeSet::new(),
            exit_live: BTreeSet::new(),
            progress_carrier: None,
            variable_definitions,
        };

        assert!(is_outer_scope_variable("start", Some(&scope)));
    }

    #[test]
    fn test_is_outer_scope_variable_from_body() {
        use std::collections::{BTreeMap, BTreeSet};

        let mut variable_definitions = BTreeMap::new();
        let mut header_and_body = BTreeSet::new();
        header_and_body.insert(BasicBlockId(0)); // header
        header_and_body.insert(BasicBlockId(1)); // body

        variable_definitions.insert("ch".to_string(), header_and_body);

        let scope = LoopScopeShape {
            header: BasicBlockId(0),
            body: BasicBlockId(1),
            latch: BasicBlockId(2),
            exit: BasicBlockId(3),
            pinned: BTreeSet::new(),
            carriers: BTreeSet::new(),
            body_locals: BTreeSet::new(),
            exit_live: BTreeSet::new(),
            progress_carrier: None,
            variable_definitions,
        };

        // Variable defined in body (in addition to header) → NOT outer-only
        assert!(!is_outer_scope_variable("ch", Some(&scope)));
    }
}