graphersal 0.2.0

In-memory graph database and traversal engine modelled on Apache TinkerPop Gremlin, with vector search for GraphRAG, a Rhai DSL, optimizer, persistence and GraphSON/GraphML I/O
Documentation 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36

use crate::graph::GraphStorage;
use crate::traversal::optimizer::plan::Plan;
use crate::traversal::step::GraphStep;

/// A rule that inspects and transforms traversal steps.
#[allow(dead_code)]
pub(crate) trait OptimizationRule<'g, S: GraphStorage + 'g>: Send + Sync {
    /// Human-readable identifier for the rule.
    fn name(&self) -> &'static str;

    /// Attempts to apply the optimization rule to the plan in-place.
    /// Returns `true` if steps were modified, `false` otherwise.
    fn apply(&self, plan: &mut Plan<'g, S>) -> bool;

    /// [`Self::apply`] with what the optimizer knows about the run (the storage's property
    /// indexes, whether this level is the root). Rules that need it override this one; the
    /// optimizer always calls it.
    fn apply_in(
        &self,
        plan: &mut Plan<'g, S>,
        context: &crate::traversal::optimizer::OptimizerContext<'_>,
    ) -> bool {
        let _ = context;
        self.apply(plan)
    }

    /// Applies the rule to bare steps, discarding logical provenance.
    ///
    /// Only for callers that have no plan to maintain, such as rule unit tests.
    fn apply_to_steps(&self, steps: &mut Vec<GraphStep<'g, S>>) -> bool {
        let mut plan = Plan::from_steps(std::mem::take(steps), true);
        let modified = self.apply(&mut plan);
        *steps = plan.into_steps();
        modified
    }
}