|
|
|
@ -7,11 +7,10 @@
|
|
|
|
*/
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
use std::collections::BTreeMap;
|
|
|
|
use std::collections::BTreeMap;
|
|
|
|
use std::mem;
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
use approx::AbsDiffEq;
|
|
|
|
#[cfg(feature = "rayon")]
|
|
|
|
|
|
|
|
use graph::prelude::{page_rank, CsrLayout, DirectedCsrGraph, GraphBuilder, PageRankConfig};
|
|
|
|
use miette::Result;
|
|
|
|
use miette::Result;
|
|
|
|
use nalgebra::{Dynamic, OMatrix, U1};
|
|
|
|
|
|
|
|
use smartstring::{LazyCompact, SmartString};
|
|
|
|
use smartstring::{LazyCompact, SmartString};
|
|
|
|
|
|
|
|
|
|
|
|
use crate::algo::AlgoImpl;
|
|
|
|
use crate::algo::AlgoImpl;
|
|
|
|
@ -34,20 +33,49 @@ impl AlgoImpl for PageRank {
|
|
|
|
algo: &'a MagicAlgoApply,
|
|
|
|
algo: &'a MagicAlgoApply,
|
|
|
|
stores: &'a BTreeMap<MagicSymbol, InMemRelation>,
|
|
|
|
stores: &'a BTreeMap<MagicSymbol, InMemRelation>,
|
|
|
|
out: &'a InMemRelation,
|
|
|
|
out: &'a InMemRelation,
|
|
|
|
poison: Poison,
|
|
|
|
_poison: Poison,
|
|
|
|
) -> Result<()> {
|
|
|
|
) -> Result<()> {
|
|
|
|
let edges = algo.relation(0)?;
|
|
|
|
let edges = algo.relation(0)?;
|
|
|
|
let undirected = algo.bool_option("undirected", Some(false))?;
|
|
|
|
let undirected = algo.bool_option("undirected", Some(false))?;
|
|
|
|
let theta = algo.unit_interval_option("theta", Some(0.8))? as f32;
|
|
|
|
let theta = algo.unit_interval_option("theta", Some(0.8))? as f32;
|
|
|
|
let epsilon = algo.unit_interval_option("epsilon", Some(0.05))? as f32;
|
|
|
|
let epsilon = algo.unit_interval_option("epsilon", Some(0.05))? as f32;
|
|
|
|
let iterations = algo.pos_integer_option("iterations", Some(20))?;
|
|
|
|
let iterations = algo.pos_integer_option("iterations", Some(10))?;
|
|
|
|
|
|
|
|
|
|
|
|
let (graph, indices, _) = edges.convert_edge_to_graph(undirected, tx, stores)?;
|
|
|
|
let (graph, indices, _) = edges.convert_edge_to_graph(undirected, tx, stores)?;
|
|
|
|
let res = pagerank(&graph, theta, epsilon, iterations, poison)?;
|
|
|
|
|
|
|
|
for (idx, score) in res.iter().enumerate() {
|
|
|
|
#[cfg(feature = "rayon")]
|
|
|
|
out.put(
|
|
|
|
{
|
|
|
|
Tuple(vec![indices[idx].clone(), DataValue::from(*score as f64)]),
|
|
|
|
let graph: DirectedCsrGraph<u32> = GraphBuilder::new()
|
|
|
|
0,
|
|
|
|
.csr_layout(CsrLayout::Sorted)
|
|
|
|
|
|
|
|
.edges(
|
|
|
|
|
|
|
|
graph
|
|
|
|
|
|
|
|
.iter()
|
|
|
|
|
|
|
|
.enumerate()
|
|
|
|
|
|
|
|
.flat_map(|(fr, ls)| ls.iter().map(move |to| (fr as u32, *to as u32))),
|
|
|
|
|
|
|
|
)
|
|
|
|
|
|
|
|
.build();
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
let (ranks, _n_run, _) = page_rank(
|
|
|
|
|
|
|
|
&graph,
|
|
|
|
|
|
|
|
PageRankConfig::new(iterations, epsilon as f64, theta),
|
|
|
|
);
|
|
|
|
);
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
for (idx, score) in ranks.iter().enumerate() {
|
|
|
|
|
|
|
|
out.put(
|
|
|
|
|
|
|
|
Tuple(vec![indices[idx].clone(), DataValue::from(*score as f64)]),
|
|
|
|
|
|
|
|
0,
|
|
|
|
|
|
|
|
);
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
}
|
|
|
|
|
|
|
|
#[cfg(not(feature = "rayon"))]
|
|
|
|
|
|
|
|
{
|
|
|
|
|
|
|
|
let res = pagerank(&graph, theta, epsilon, iterations, _poison)?;
|
|
|
|
|
|
|
|
for (idx, score) in res.iter().enumerate() {
|
|
|
|
|
|
|
|
out.put(
|
|
|
|
|
|
|
|
Tuple(vec![indices[idx].clone(), DataValue::from(*score as f64)]),
|
|
|
|
|
|
|
|
0,
|
|
|
|
|
|
|
|
);
|
|
|
|
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
Ok(())
|
|
|
|
Ok(())
|
|
|
|
}
|
|
|
|
}
|
|
|
|
@ -62,6 +90,7 @@ impl AlgoImpl for PageRank {
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
#[cfg(not(feature = "rayon"))]
|
|
|
|
fn pagerank(
|
|
|
|
fn pagerank(
|
|
|
|
edges: &[Vec<usize>],
|
|
|
|
edges: &[Vec<usize>],
|
|
|
|
theta: f32,
|
|
|
|
theta: f32,
|
|
|
|
@ -69,6 +98,9 @@ fn pagerank(
|
|
|
|
iterations: usize,
|
|
|
|
iterations: usize,
|
|
|
|
poison: Poison,
|
|
|
|
poison: Poison,
|
|
|
|
) -> Result<OMatrix<f32, Dynamic, U1>> {
|
|
|
|
) -> Result<OMatrix<f32, Dynamic, U1>> {
|
|
|
|
|
|
|
|
use approx::AbsDiffEq;
|
|
|
|
|
|
|
|
use nalgebra::{Dynamic, OMatrix, U1};
|
|
|
|
|
|
|
|
|
|
|
|
let init_val = (1. - theta) / edges.len() as f32;
|
|
|
|
let init_val = (1. - theta) / edges.len() as f32;
|
|
|
|
let mut g_mat = OMatrix::<f32, Dynamic, Dynamic>::repeat(edges.len(), edges.len(), init_val);
|
|
|
|
let mut g_mat = OMatrix::<f32, Dynamic, Dynamic>::repeat(edges.len(), edges.len(), init_val);
|
|
|
|
let n = edges.len();
|
|
|
|
let n = edges.len();
|
|
|
|
@ -90,7 +122,7 @@ fn pagerank(
|
|
|
|
let scale_target = (n as f32).sqrt();
|
|
|
|
let scale_target = (n as f32).sqrt();
|
|
|
|
let mut last_pi_vec = pi_vec.clone();
|
|
|
|
let mut last_pi_vec = pi_vec.clone();
|
|
|
|
for _ in 0..iterations {
|
|
|
|
for _ in 0..iterations {
|
|
|
|
mem::swap(&mut pi_vec, &mut last_pi_vec);
|
|
|
|
std::mem::swap(&mut pi_vec, &mut last_pi_vec);
|
|
|
|
pi_vec = g_mat.tr_mul(&last_pi_vec);
|
|
|
|
pi_vec = g_mat.tr_mul(&last_pi_vec);
|
|
|
|
pi_vec.normalize_mut();
|
|
|
|
pi_vec.normalize_mut();
|
|
|
|
let f = pi_vec.norm() / scale_target;
|
|
|
|
let f = pi_vec.norm() / scale_target;
|
|
|
|
|
Ziyang Hu
2 years ago