ohnomore/transform/
mod.rs

1use udgraph::graph::Sentence;
2
3#[allow(clippy::len_without_is_empty)]
4pub trait DependencyGraph {
5    fn dependents<'a>(&'a self, idx: usize) -> Box<dyn Iterator<Item = (usize, String)> + 'a>;
6
7    fn token(&self, idx: usize) -> &dyn Token;
8
9    fn token_mut(&mut self, idx: usize) -> &mut dyn TokenMut;
10
11    fn len(&self) -> usize;
12}
13
14impl DependencyGraph for Sentence {
15    fn dependents<'a>(&'a self, idx: usize) -> Box<dyn Iterator<Item = (usize, String)> + 'a> {
16        Box::new(self.dep_graph().dependents(idx).map(|triple| {
17            (
18                triple.dependent(),
19                triple
20                    .relation()
21                    .expect("Edge without a dependency relation")
22                    .to_owned(),
23            )
24        }))
25    }
26
27    fn token(&self, idx: usize) -> &dyn Token {
28        self[idx]
29            .token()
30            .expect("The root node was used as a token")
31    }
32
33    fn token_mut(&mut self, idx: usize) -> &mut dyn TokenMut {
34        self[idx]
35            .token_mut()
36            .expect("The root node was used as a token")
37    }
38
39    fn len(&self) -> usize {
40        self.len()
41    }
42}
43
44pub trait TokenMut: Token {
45    fn set_lemma(&mut self, lemma: Option<String>);
46}
47
48pub trait Token {
49    fn form(&self) -> &str;
50    fn lemma(&self) -> &str;
51    fn upos(&self) -> &str;
52    fn xpos(&self) -> &str;
53}
54
55impl Token for udgraph::token::Token {
56    fn form(&self) -> &str {
57        self.form()
58    }
59
60    fn lemma(&self) -> &str {
61        self.lemma().unwrap_or("_")
62    }
63
64    fn upos(&self) -> &str {
65        self.upos().unwrap()
66    }
67
68    fn xpos(&self) -> &str {
69        self.xpos().unwrap()
70    }
71}
72
73impl TokenMut for udgraph::token::Token {
74    fn set_lemma(&mut self, lemma: Option<String>) {
75        self.set_lemma(lemma);
76    }
77}
78
79pub trait Transform: Sync {
80    fn transform(&self, graph: &dyn DependencyGraph, node: usize) -> String;
81}
82
83/// A list of `Transform`s.
84pub struct Transforms(pub Vec<Box<dyn Transform>>);
85
86impl Transforms {
87    /// Transform a graph using the transformation list.
88    ///
89    /// This method applies the transformations to the given graph. Each
90    /// transform is fully applied to the graph before the next transform,
91    /// to ensure that dependencies between transforms are correctly handled.
92    pub fn transform(&self, graph: &mut dyn DependencyGraph) {
93        for t in &self.0 {
94            for idx in 1..graph.len() {
95                let lemma = t.as_ref().transform(graph, idx);
96                graph.token_mut(idx).set_lemma(Some(lemma));
97            }
98        }
99    }
100}
101
102pub mod delemmatization;
103
104pub mod lemmatization;
105
106pub mod misc;
107
108mod named_entity;
109
110mod svp;
111
112#[cfg(test)]
113pub(crate) mod test_helpers;