Skip to content

Offline, privacy-first grammar checker. Fast, open-source, Rust-powered

fix: hunspell dictionary builder now emits added word metadata

Elijah Potterb4f7c342344d
diff --git a/harper-core/affixes.json b/harper-core/affixes.json
index 578ac65f5..063e5b8f0 100644
--- a/harper-core/affixes.json
+++ b/harper-core/affixes.json
@@ -12,7 +12,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"L": {
@@ -27,7 +28,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"E": {
@@ -42,7 +44,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"Y": {
@@ -57,7 +60,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"U": {
@@ -72,7 +76,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"H": {
@@ -92,7 +97,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"T": {
@@ -122,7 +128,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"R": {
@@ -152,7 +159,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"C": {
@@ -167,7 +175,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"V": {
@@ -187,7 +196,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"N": {
@@ -212,7 +222,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"A": {
@@ -227,7 +238,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"Z": {
@@ -257,7 +269,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"P": {
@@ -282,7 +295,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"M": {
@@ -297,7 +311,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"F": {
@@ -312,7 +327,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"B": {
@@ -337,7 +353,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"S": {
@@ -367,7 +384,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": true
}
},
"D": {
@@ -397,7 +415,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"G": {
@@ -417,7 +436,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"Q": {
@@ -432,7 +452,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"O": {
@@ -447,7 +468,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"I": {
@@ -462,7 +484,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"X": {
@@ -487,7 +510,8 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
},
"J": {
@@ -507,8 +531,9 @@
],
"adds_metadata": {
"kind": null,
- "tense": null
+ "tense": null,
+ "plural": null
}
}
}
-}
\ No newline at end of file
+}
diff --git a/harper-core/src/spell/full_dictionary.rs b/harper-core/src/spell/full_dictionary.rs
index c264176ac..beb96597f 100644
--- a/harper-core/src/spell/full_dictionary.rs
+++ b/harper-core/src/spell/full_dictionary.rs
@@ -1,11 +1,11 @@
-use hashbrown::HashSet;
+use hashbrown::{HashMap, HashSet};
use once_cell::sync::Lazy;
use smallvec::{SmallVec, ToSmallVec};
use super::dictionary::Dictionary;
use super::hunspell::{parse_default_attribute_list, parse_default_word_list};
use super::seq_to_normalized;
-use crate::CharString;
+use crate::{CharString, WordMetadata};
/// A full, fat dictionary.
/// All of the elements are stored in-memory.
@@ -23,7 +23,7 @@ pub struct FullDictionary {
/// that has a word whose index is that length.
word_len_starts: Vec<usize>,
/// All English words
- word_set: HashSet<CharString>,
+ word_map: HashMap<CharString, WordMetadata>
}
fn uncached_inner_new() -> FullDictionary {
@@ -31,17 +31,18 @@ fn uncached_inner_new() -> FullDictionary {
let attr_list = parse_default_attribute_list();
// There will be at _least_ this number of words
- let mut words = Vec::with_capacity(word_list.len());
+ let mut word_map = HashMap::with_capacity(word_list.len());
- attr_list.expand_marked_words(word_list, &mut words);
+ attr_list.expand_marked_words(word_list, &mut word_map);
+ let mut words: Vec<CharString> = word_map.iter().map(|(v, _)| v.clone()).collect();
words.sort();
words.dedup();
FullDictionary {
- word_set: HashSet::from_iter(words.iter().cloned()),
+ word_map,
word_len_starts: FullDictionary::create_len_starts(&mut words),
- words,
+ words
}
}
@@ -52,7 +53,7 @@ impl FullDictionary {
Self {
words: Vec::new(),
word_len_starts: Vec::new(),
- word_set: HashSet::new(),
+ word_map: HashMap::new()
}
}
@@ -65,21 +66,26 @@ impl FullDictionary {
/// Appends words to the dictionary.
/// It is significantly faster to append many words with one call than many
/// distinct calls to this function.
- pub fn extend_words(&mut self, words: impl IntoIterator<Item = impl AsRef<[char]>>) {
- let init_size = self.words.len();
- self.words
- .extend(words.into_iter().map(|v| v.as_ref().to_smallvec()));
- self.word_set
- .extend(self.words[init_size..].iter().cloned());
+ pub fn extend_words(
+ &mut self,
+ words: impl IntoIterator<Item = (impl AsRef<[char]>, WordMetadata)>
+ ) {
+ let pairs: Vec<_> = words
+ .into_iter()
+ .map(|(v, m)| (v.as_ref().to_smallvec(), m))
+ .collect();
+
+ self.words.extend(pairs.iter().map(|(v, _)| v.clone()));
self.word_len_starts = Self::create_len_starts(&mut self.words);
+ self.word_map.extend(pairs);
}
/// Append a single word to the dictionary.
///
/// If you are appending many words, consider using [`Self::extend_words`]
/// instead.
- pub fn append_word(&mut self, word: impl AsRef<[char]>) {
- self.extend_words(std::iter::once(word.as_ref()))
+ pub fn append_word(&mut self, word: impl AsRef<[char]>, metadata: WordMetadata) {
+ self.extend_words(std::iter::once((word.as_ref(), metadata)))
}
/// Create a lookup table for finding words of a specific length in a word
@@ -132,7 +138,7 @@ impl Dictionary for FullDictionary {
let normalized = seq_to_normalized(word);
let lowercase: SmallVec<_> = normalized.iter().flat_map(|c| c.to_lowercase()).collect();
- self.word_set.contains(normalized.as_ref()) || self.word_set.contains(&lowercase)
+ self.word_map.contains_key(normalized.as_ref()) || self.word_map.contains_key(&lowercase)
}
}
diff --git a/harper-core/src/spell/hunspell/affix_replacement.rs b/harper-core/src/spell/hunspell/affix_replacement.rs
index 17744a590..dcdab2b96 100644
--- a/harper-core/src/spell/hunspell/affix_replacement.rs
+++ b/harper-core/src/spell/hunspell/affix_replacement.rs
@@ -1,12 +1,13 @@
use serde::{Deserialize, Serialize};
-use super::{matcher::Matcher, Error};
+use super::matcher::Matcher;
+use super::Error;
#[derive(Debug, Clone)]
pub struct AffixReplacement {
pub remove: Vec<char>,
pub add: Vec<char>,
- pub condition: Matcher,
+ pub condition: Matcher
}
impl AffixReplacement {
@@ -14,18 +15,18 @@ impl AffixReplacement {
HumanReadableAffixReplacement {
remove: self.remove.iter().collect(),
add: self.add.iter().collect(),
- condition: self.condition.to_string(),
+ condition: self.condition.to_string()
}
}
}
-/// A version of [`AffixReplacement`] that can be serialized to JSON (or whatever) and maintain the
-/// nice Regex syntax of the inner [`Matcher`].
+/// A version of [`AffixReplacement`] that can be serialized to JSON (or
+/// whatever) and maintain the nice Regex syntax of the inner [`Matcher`].
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct HumanReadableAffixReplacement {
pub remove: String,
pub add: String,
- pub condition: String,
+ pub condition: String
}
impl HumanReadableAffixReplacement {
@@ -33,7 +34,7 @@ impl HumanReadableAffixReplacement {
Ok(AffixReplacement {
remove: self.remove.chars().collect(),
add: self.add.chars().collect(),
- condition: Matcher::parse(&self.condition)?,
+ condition: Matcher::parse(&self.condition)?
})
}
}
diff --git a/harper-core/src/spell/hunspell/attribute_list.rs b/harper-core/src/spell/hunspell/attribute_list.rs
index 0626663c4..953b26c0d 100644
--- a/harper-core/src/spell/hunspell/attribute_list.rs
+++ b/harper-core/src/spell/hunspell/attribute_list.rs
@@ -2,16 +2,16 @@ use hashbrown::HashMap;
use serde::{Deserialize, Serialize};
use smallvec::ToSmallVec;
+use super::affix_replacement::AffixReplacement;
use super::expansion::{Expansion, HumanReadableExpansion};
use super::word_list::MarkedWord;
use super::Error;
-use super::{affix_replacement::AffixReplacement, matcher::Matcher};
use crate::{CharString, Span, WordMetadata};
#[derive(Debug, Clone)]
pub struct AttributeList {
/// Key = Affix Flag
- affixes: HashMap<char, Expansion>,
+ affixes: HashMap<char, Expansion>
}
impl AttributeList {
@@ -21,7 +21,7 @@ impl AttributeList {
.affixes
.iter()
.map(|(affix, exp)| (*affix, exp.to_human_readable()))
- .collect(),
+ .collect()
}
}
@@ -30,24 +30,26 @@ impl AttributeList {
/// Will append to the given `dest`;
///
/// In the future, I want to make this function cleaner and faster.
- pub fn expand_marked_word(&self, word: MarkedWord, dest: &mut Vec<CharString>) {
+ pub fn expand_marked_word(
+ &self,
+ word: MarkedWord,
+ dest: &mut HashMap<CharString, WordMetadata>
+ ) {
dest.reserve(word.attributes.len() + 1);
- let start_len = dest.len();
-
for attr in &word.attributes {
let Some(expansion) = self.affixes.get(attr) else {
continue;
};
- let mut new_words = Vec::new();
+ let mut new_words = HashMap::new();
for replacement in &expansion.replacements {
- new_words.extend(Self::apply_replacement(
- replacement,
- &word.letters,
- expansion.suffix,
- ))
+ if let Some(replaced) =
+ Self::apply_replacement(replacement, &word.letters, expansion.suffix)
+ {
+ new_words.insert(replaced, expansion.adds_metadata);
+ }
}
if expansion.cross_product {
@@ -62,31 +64,21 @@ impl AttributeList {
}
}
- let cross_product_words = Vec::new();
-
- for new_word in new_words {
+ for (new_word, _metadata) in new_words {
self.expand_marked_word(
MarkedWord {
letters: new_word,
- attributes: opp_attr.clone(),
+ attributes: opp_attr.clone()
},
- dest,
+ dest
);
}
-
- dest.extend_from_slice(&cross_product_words);
} else {
- dest.extend_from_slice(&new_words);
+ dest.extend(new_words.into_iter());
}
}
- dest.push(word.letters);
-
- let mut split = dest.split_off(start_len);
- split.sort();
- split.dedup();
-
- dest.append(&mut split);
+ dest.insert(word.letters, WordMetadata::default());
}
/// Expand an iterator of marked words into strings.
@@ -95,7 +87,7 @@ impl AttributeList {
pub fn expand_marked_words(
&self,
words: impl IntoIterator<Item = MarkedWord>,
- dest: &mut Vec<CharString>,
+ dest: &mut HashMap<CharString, WordMetadata>
) {
for word in words {
self.expand_marked_word(word, dest);
@@ -105,7 +97,7 @@ impl AttributeList {
fn apply_replacement(
replacement: &AffixReplacement,
letters: &[char],
- suffix: bool,
+ suffix: bool
) -> Option<CharString> {
if replacement.condition.len() > letters.len() {
return None;
@@ -160,7 +152,7 @@ impl AttributeList {
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct HumanReadableAttributeList {
- affixes: HashMap<char, HumanReadableExpansion>,
+ affixes: HashMap<char, HumanReadableExpansion>
}
impl HumanReadableAttributeList {
diff --git a/harper-core/src/spell/hunspell/expansion.rs b/harper-core/src/spell/hunspell/expansion.rs
index 21a64fdf6..46a86f7f5 100644
--- a/harper-core/src/spell/hunspell/expansion.rs
+++ b/harper-core/src/spell/hunspell/expansion.rs
@@ -1,19 +1,18 @@
use serde::{Deserialize, Serialize};
+use super::affix_replacement::{AffixReplacement, HumanReadableAffixReplacement};
+use super::Error;
use crate::WordMetadata;
-use super::{
- affix_replacement::{AffixReplacement, HumanReadableAffixReplacement},
- Error,
-};
-
#[derive(Debug, Clone)]
pub struct Expansion {
/// If `!true`, this is a prefix
pub suffix: bool,
pub cross_product: bool,
pub replacements: Vec<AffixReplacement>,
- pub adds_metadata: WordMetadata,
+ /// When the expansion is applied, the resulting word will have fields
+ /// overwritten with the none-null properties from here.
+ pub adds_metadata: WordMetadata
}
impl Expansion {
@@ -26,7 +25,7 @@ impl Expansion {
.iter()
.map(AffixReplacement::to_human_readable)
.collect(),
- adds_metadata: self.adds_metadata,
+ adds_metadata: self.adds_metadata
}
}
}
@@ -36,7 +35,7 @@ pub struct HumanReadableExpansion {
pub suffix: bool,
pub cross_product: bool,
pub replacements: Vec<HumanReadableAffixReplacement>,
- pub adds_metadata: WordMetadata,
+ pub adds_metadata: WordMetadata
}
impl HumanReadableExpansion {
@@ -51,7 +50,7 @@ impl HumanReadableExpansion {
suffix: self.suffix,
cross_product: self.cross_product,
replacements,
- adds_metadata: self.adds_metadata,
+ adds_metadata: self.adds_metadata
})
}
}
diff --git a/harper-core/src/spell/hunspell/matcher.rs b/harper-core/src/spell/hunspell/matcher.rs
index f0574185f..0776afd5a 100644
--- a/harper-core/src/spell/hunspell/matcher.rs
+++ b/harper-core/src/spell/hunspell/matcher.rs
@@ -8,7 +8,7 @@ use serde::{Deserialize, Serialize};
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct Matcher {
/// Position-based operators.
- operators: Vec<Operator>,
+ operators: Vec<Operator>
}
impl Matcher {
@@ -42,7 +42,7 @@ impl Matcher {
}
}
'.' => operators.push(Operator::Any),
- _ => operators.push(Operator::Literal(c)),
+ _ => operators.push(Operator::Literal(c))
}
char_idx += 1;
@@ -93,7 +93,7 @@ impl Display for Matcher {
write!(f, "]")?;
}
- Operator::Any => write!(f, ".")?,
+ Operator::Any => write!(f, ".")?
}
}
@@ -106,7 +106,7 @@ enum Operator {
Literal(char),
MatchOne(Vec<char>),
MatchNone(Vec<char>),
- Any,
+ Any
}
impl Operator {
@@ -115,7 +115,7 @@ impl Operator {
Operator::Literal(b) => a == *b,
Operator::MatchOne(b) => b.contains(&a),
Operator::MatchNone(b) => !b.contains(&a),
- Operator::Any => true,
+ Operator::Any => true
}
}
}
@@ -123,7 +123,7 @@ impl Operator {
#[derive(Debug, Clone, Copy, thiserror::Error)]
pub enum Error {
#[error("Unmatched bracket at index: {index}")]
- UnmatchedBracket { index: usize },
+ UnmatchedBracket { index: usize }
}
#[cfg(test)]
diff --git a/harper-core/src/spell/hunspell/mod.rs b/harper-core/src/spell/hunspell/mod.rs
index 5a7d2c559..ed7e4989d 100644
--- a/harper-core/src/spell/hunspell/mod.rs
+++ b/harper-core/src/spell/hunspell/mod.rs
@@ -28,12 +28,13 @@ pub fn parse_default_attribute_list() -> AttributeList {
#[cfg(test)]
mod tests {
+ use hashbrown::HashMap;
use serde_json::json;
use super::word_list::parse_word_list;
use super::{parse_default_attribute_list, parse_default_word_list};
use crate::spell::hunspell::attribute_list::HumanReadableAttributeList;
- use crate::CharString;
+ use crate::{CharString, WordMetadata};
pub const TEST_WORD_LIST: &str = "3\nhello\ntry/B\nwork/AB";
@@ -82,17 +83,17 @@ mod tests {
.unwrap();
let attributes = attributes.to_normal().unwrap();
- let mut expanded = Vec::new();
+ let mut expanded = HashMap::new();
attributes.expand_marked_words(words, &mut expanded);
let expanded: Vec<String> = expanded
.into_iter()
- .map(|v| v.into_iter().collect())
+ .map(|v| v.0.into_iter().collect())
.collect();
assert_eq!(
expanded,
- vec!["hello", "tried", "try", "rework", "reworked", "work", "worked",]
+ vec!["hello", "tried", "reworked", "rework", "worked", "work", "try"]
)
}
@@ -147,18 +148,18 @@ mod tests {
.unwrap();
let attributes = attributes.to_normal().unwrap();
- let mut expanded = Vec::new();
+ let mut expanded: HashMap<CharString, WordMetadata> = HashMap::new();
attributes.expand_marked_words(words, &mut expanded);
- assert!(expanded.contains(&split("giants")))
+ assert!(expanded.contains_key(&split("giants")))
}
- fn build_expanded() -> Vec<CharString> {
+ fn build_expanded() -> HashMap<CharString, WordMetadata> {
let words = parse_default_word_list().unwrap();
let attributes = parse_default_attribute_list();
- let mut expanded = Vec::new();
+ let mut expanded = HashMap::new();
attributes.expand_marked_words(words, &mut expanded);
@@ -172,12 +173,12 @@ mod tests {
#[test]
fn expanded_contains_giants() {
- assert!(build_expanded().contains(&split("giants")));
+ assert!(build_expanded().contains_key(&split("giants")));
}
#[test]
fn expanded_contains_deallocate() {
- assert!(build_expanded().contains(&split("deallocate")));
+ assert!(build_expanded().contains_key(&split("deallocate")));
}
fn split(text: &str) -> CharString {
diff --git a/harper-core/src/word_metadata.rs b/harper-core/src/word_metadata.rs
index 2f45af7c7..40c1c8e08 100644
--- a/harper-core/src/word_metadata.rs
+++ b/harper-core/src/word_metadata.rs
@@ -1,22 +1,23 @@
use serde::{Deserialize, Serialize};
-#[derive(Debug, Default, Copy, Clone, PartialEq, Eq, Serialize, Deserialize)]
+#[derive(Debug, Default, Copy, Clone, PartialEq, Eq, Serialize, Deserialize, PartialOrd)]
pub struct WordMetadata {
kind: Option<WordKind>,
tense: Option<Tense>,
+ plural: Option<bool>
}
-#[derive(Debug, Copy, Clone, PartialEq, Eq, Serialize, Deserialize)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Serialize, Deserialize, PartialOrd)]
pub enum Tense {
Past,
Present,
- Future,
+ Future
}
-#[derive(Debug, Copy, Clone, PartialEq, Eq, Serialize, Deserialize)]
+#[derive(Debug, Copy, Clone, PartialEq, Eq, Serialize, Deserialize, PartialOrd)]
pub enum WordKind {
Verb,
Noun { is_proper: bool },
Adjective,
- Adverb,
+ Adverb
}
diff --git a/harper-ls/src/backend.rs b/harper-ls/src/backend.rs
index c613ff7cb..c678d8410 100644
--- a/harper-ls/src/backend.rs
+++ b/harper-ls/src/backend.rs
@@ -13,7 +13,8 @@ use harper_core::{
Linter,
MergedDictionary,
Token,
- TokenKind
+ TokenKind,
+ WordMetadata
};
use harper_html::HtmlParser;
use serde_json::Value;
@@ -494,7 +495,7 @@ impl LanguageServer for Backend {
let file_url = second.parse().unwrap();
let mut dict = self.load_user_dictionary().await;
- dict.append_word(word);
+ dict.append_word(word, WordMetadata::default());
if let Err(err) = self.save_user_dictionary(dict).await {
error!("Unable to save user dictionary: {}", err);
}
@@ -515,7 +516,7 @@ impl LanguageServer for Backend {
error!("Unable resolve dictionary path: {file_url}");
return Ok(None);
};
- dict.append_word(word);
+ dict.append_word(word, WordMetadata::default());
if let Err(err) = self.save_file_dictionary(&file_url, dict).await {
error!("Unable to save file dictionary: {}", err);
diff --git a/harper-ls/src/dictionary_io.rs b/harper-ls/src/dictionary_io.rs
index 4af97a905..ff3a4bfd1 100644
--- a/harper-ls/src/dictionary_io.rs
+++ b/harper-ls/src/dictionary_io.rs
@@ -1,6 +1,6 @@
use std::path::Path;
-use harper_core::{Dictionary, FullDictionary};
+use harper_core::{Dictionary, FullDictionary, WordMetadata};
use tokio::fs::{self, File};
use tokio::io::{self, AsyncRead, AsyncReadExt, AsyncWrite, AsyncWriteExt, BufReader, BufWriter};
@@ -50,7 +50,10 @@ async fn dict_from_word_list(mut r: impl AsyncRead + Unpin) -> io::Result<FullDi
r.read_to_string(&mut str).await?;
let mut dict = FullDictionary::new();
- dict.extend_words(str.lines().map(|l| l.chars().collect::<Vec<char>>()));
+ dict.extend_words(
+ str.lines()
+ .map(|l| (l.chars().collect::<Vec<char>>(), WordMetadata::default()))
+ );
Ok(dict)
}
diff --git a/harper-tree-sitter/src/lib.rs b/harper-tree-sitter/src/lib.rs
index 36d347eb7..a94809b68 100644
--- a/harper-tree-sitter/src/lib.rs
+++ b/harper-tree-sitter/src/lib.rs
@@ -1,6 +1,6 @@
use std::collections::HashSet;
-use harper_core::{FullDictionary, Mask, Masker, Span};
+use harper_core::{FullDictionary, Mask, Masker, Span, WordMetadata};
use tree_sitter::{Language, Node, Tree, TreeCursor};
/// A Harper [`Masker`] that wraps a given tree-sitter language and a condition,
@@ -47,7 +47,10 @@ impl TreeSitterMasker {
idents.insert(span.get_content(source));
}
- let idents: Vec<_> = idents.into_iter().collect();
+ let idents: Vec<_> = idents
+ .into_iter()
+ .map(|ident| (ident, WordMetadata::default()))
+ .collect();
let mut dictionary = FullDictionary::new();
dictionary.extend_words(idents);