feat: add harper-cli forms and just getforms

harper-cli/Cargo.toml

@@ -14,6 +14,7 @@ harper-literate-haskell = { path = "../harper-literate-haskell", version = "0.19
 harper-core = { path = "../harper-core", version = "0.19.0" }
 harper-comments = { path = "../harper-comments", version = "0.19.0" }
 harper-typst = { path = "../harper-typst", version = "0.19.0" }
+hashbrown = "0.15.2"
 serde = { version = "1.0.214", features = ["derive"] }
 serde_json = "1.0.138"
 

harper-cli/src/main.rs

@@ -1,6 +1,6 @@
 #![doc = include_str!("../README.md")]
 
-use std::collections::{BTreeMap, HashMap};
+use std::collections::BTreeMap;
 use std::path::{Path, PathBuf};
 
 use anyhow::format_err;
@@ -9,8 +9,13 @@ use clap::Parser;
 use harper_comments::CommentParser;
 use harper_core::linting::{LintGroup, LintGroupConfig, Linter};
 use harper_core::parsers::{Markdown, MarkdownOptions};
-use harper_core::{remove_overlaps, Dictionary, Document, FstDictionary, TokenKind};
+use harper_core::spell::hunspell::parse_default_attribute_list;
+use harper_core::spell::hunspell::word_list::parse_word_list;
+use harper_core::{
+    remove_overlaps, CharString, Dictionary, Document, FstDictionary, TokenKind, WordMetadata,
+};
 use harper_literate_haskell::LiterateHaskellParser;
+use hashbrown::HashMap;
 use serde::Serialize;
 
 /// A debugging tool for the Harper grammar checker.
@@ -41,6 +46,8 @@ enum Args {
     },
     /// Get the metadata associated with a particular word.
     Metadata { word: String },
+    /// Get all the forms of a word using the affixes.
+    Forms { word: String },
     /// Emit a decompressed, line-separated list of the words in Harper's dictionary.
     Words,
     /// Print the default config with descriptions.
@@ -156,7 +163,7 @@ fn main() -> anyhow::Result<()> {
                 word_str.clear();
                 word_str.extend(word);
 
-                println!("{}", word_str);
+                println!("{:?}", word_str);
             }
 
             Ok(())
@@ -169,6 +176,22 @@ fn main() -> anyhow::Result<()> {
 
             Ok(())
         }
+        Args::Forms { word } => {
+            let hunspell_word_list = format!("1\n{word}");
+            let words = parse_word_list(&hunspell_word_list.to_string()).unwrap();
+
+            let attributes = parse_default_attribute_list();
+
+            let mut expanded: HashMap<CharString, WordMetadata> = HashMap::new();
+
+            attributes.expand_marked_words(words, &mut expanded);
+
+            expanded.keys().for_each(|form| {
+                let string_form: String = form.iter().collect();
+                println!("{}", string_form);
+            });
+            Ok(())
+        }
         Args::Config => {
             #[derive(Serialize)]
             struct Config {

harper-core/src/lib.rs

@@ -16,7 +16,7 @@ pub mod parsers;
 pub mod patterns;
 mod punctuation;
 mod span;
-mod spell;
+pub mod spell;
 mod sync;
 mod title_case;
 mod token;

harper-core/src/spell/hunspell/mod.rs

@@ -3,7 +3,7 @@ mod attribute_list;
 mod error;
 mod expansion;
 mod matcher;
-mod word_list;
+pub mod word_list;
 
 pub use attribute_list::AttributeList;
 use attribute_list::HumanReadableAttributeList;

harper-core/src/spell/mod.rs

@@ -12,7 +12,7 @@ pub use self::merged_dictionary::MergedDictionary;
 mod dictionary;
 mod fst_dictionary;
 mod full_dictionary;
-mod hunspell;
+pub mod hunspell;
 mod merged_dictionary;
 
 #[derive(PartialEq, Debug)]

justfile

@@ -256,6 +256,9 @@ userdictoverlap:
 # Get the metadata associated with a particular word in Harper's dictionary as JSON.
 getmetadata word:
   cargo run --bin harper-cli -- metadata {{word}}
+# Get all the forms of a word using the affixes.
+getforms word:
+  cargo run --bin harper-cli -- forms {{word}}
 
 bump-versions: update-vscode-linters
   #! /bin/bash