lattix 0.8.1

//! Personalized PageRank (PPR).
//!
//! Computes node importance relative to a specific seed entity,
//! measuring proximity in the graph's link structure.
//!
//! The power iteration runs over unique neighbor nodes, so parallel triples
//! do not act as implicit edge weights.

use crate::{EntityId, KnowledgeGraph};
use std::collections::HashMap;

/// PPR configuration.
#[derive(Debug, Clone, Copy)]
pub struct PprConfig {
    /// Damping factor (probability of following a link vs teleporting back to seed).
    /// Typically 0.85.
    pub damping: f64,
    /// Maximum iterations before stopping.
    pub max_iterations: usize,
    /// Convergence tolerance (L1 norm of score changes).
    pub tolerance: f64,
}

impl Default for PprConfig {
    fn default() -> Self {
        Self {
            damping: 0.85,
            max_iterations: 100,
            tolerance: 1e-6,
        }
    }
}

/// Compute personalized PageRank from a seed entity.
///
/// Returns scores keyed by entity ID. Higher scores indicate
/// entities closer/more connected to the seed in the graph's link structure.
///
/// Returns an empty map if the graph is empty or the seed entity is not found.
///
/// # Example
///
/// ```
/// use lattix::{KnowledgeGraph, Triple};
/// use lattix::algo::ppr::{personalized_pagerank, PprConfig};
///
/// let mut kg = KnowledgeGraph::new();
/// kg.add_triple(Triple::new("Alice", "knows", "Bob"));
/// kg.add_triple(Triple::new("Bob", "knows", "Carol"));
///
/// let scores = personalized_pagerank(&kg, "Alice", PprConfig::default());
/// assert!(scores.contains_key("Alice"));
/// assert!(scores.contains_key("Bob"));
/// ```
#[must_use]
#[allow(clippy::cast_precision_loss)]
pub fn personalized_pagerank(
    kg: &KnowledgeGraph,
    seed: &str,
    config: PprConfig,
) -> HashMap<EntityId, f64> {
    let graph = kg.as_petgraph();
    let n = graph.node_count();
    if n == 0 {
        return HashMap::new();
    }

    // Find the seed node index
    let seed_id = crate::EntityId::from(seed);
    let seed_idx = match kg.get_node_index(&seed_id) {
        Some(idx) => idx.index(),
        None => return HashMap::new(),
    };

    let mut personalization = vec![0.0; n];
    personalization[seed_idx] = 1.0;

    let adjacency: Vec<Vec<_>> = graph
        .node_indices()
        .map(|idx| {
            crate::algo::unique_neighbors_directed(graph, idx, petgraph::Direction::Outgoing)
                .into_iter()
                .map(|n| n.index())
                .collect()
        })
        .collect();

    let mut scores = personalization.clone();
    let mut next = vec![0.0; n];

    for _ in 0..config.max_iterations {
        for (idx, value) in next.iter_mut().enumerate() {
            *value = (1.0 - config.damping) * personalization[idx];
        }

        let dangling: f64 = adjacency
            .iter()
            .enumerate()
            .filter(|(_, neighbors)| neighbors.is_empty())
            .map(|(idx, _)| scores[idx])
            .sum();
        let dangling_share = config.damping * dangling;
        for (idx, value) in next.iter_mut().enumerate() {
            *value += dangling_share * personalization[idx];
        }

        for (src, neighbors) in adjacency.iter().enumerate() {
            if neighbors.is_empty() {
                continue;
            }
            let share = config.damping * scores[src] / neighbors.len() as f64;
            for &dst in neighbors {
                next[dst] += share;
            }
        }

        let diff: f64 = scores
            .iter()
            .zip(next.iter())
            .map(|(old, new)| (old - new).abs())
            .sum();
        std::mem::swap(&mut scores, &mut next);
        if diff < config.tolerance {
            break;
        }
    }

    let mut result = HashMap::with_capacity(n);
    for (idx, score) in scores.into_iter().enumerate() {
        let entity = &graph[petgraph::graph::NodeIndex::new(idx)];
        result.insert(entity.id.clone(), score);
    }
    result
}

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

    #[test]
    fn ppr_seed_scores_highest() {
        let mut kg = KnowledgeGraph::new();
        kg.add_triple(Triple::new("A", "rel", "B"));
        kg.add_triple(Triple::new("B", "rel", "C"));
        kg.add_triple(Triple::new("C", "rel", "A"));

        let scores = personalized_pagerank(&kg, "A", PprConfig::default());

        let a = *scores.get("A").unwrap();
        let b = *scores.get("B").unwrap();
        let c = *scores.get("C").unwrap();

        // Seed should have the highest score in PPR
        assert!(a > b, "Seed A ({a}) should score higher than B ({b})");
        assert!(a > c, "Seed A ({a}) should score higher than C ({c})");
    }

    #[test]
    fn ppr_missing_seed_returns_empty() {
        let mut kg = KnowledgeGraph::new();
        kg.add_triple(Triple::new("A", "rel", "B"));

        let scores = personalized_pagerank(&kg, "Z", PprConfig::default());
        assert!(scores.is_empty());
    }

    #[test]
    fn ppr_empty_graph() {
        let kg = KnowledgeGraph::new();
        let scores = personalized_pagerank(&kg, "A", PprConfig::default());
        assert!(scores.is_empty());
    }

    #[test]
    fn ppr_scores_sum_to_one() {
        let mut kg = KnowledgeGraph::new();
        kg.add_triple(Triple::new("A", "rel", "B"));
        kg.add_triple(Triple::new("B", "rel", "C"));
        kg.add_triple(Triple::new("C", "rel", "A"));
        kg.add_triple(Triple::new("A", "rel", "D"));

        let scores = personalized_pagerank(&kg, "A", PprConfig::default());
        let total: f64 = scores.values().sum();

        assert!(
            (total - 1.0).abs() < 1e-6,
            "Scores should sum to 1.0, got {total}",
        );
    }
}