From ab232ceb32b51ac8553692cf8a2b1f86fa975f7d Mon Sep 17 00:00:00 2001
From: Alex Auvolat <alex@adnab.me>
Date: Tue, 28 Nov 2023 16:31:28 +0100
Subject: add furigana override file to fix edge cases

---
 src/datafiles.rs | 20 ++++++++++++++++++++
 src/example.rs   |  9 +++++++--
 src/main.rs      |  5 ++++-
 3 files changed, 31 insertions(+), 3 deletions(-)

(limited to 'src')

diff --git a/src/datafiles.rs b/src/datafiles.rs
index 0e526ef..3065fbf 100644
--- a/src/datafiles.rs
+++ b/src/datafiles.rs
@@ -176,6 +176,26 @@ pub fn read_examples(all_kanji: &Charset) -> Result<Vec<Example>> {
     Ok(ret)
 }
 
+pub fn read_furigana_overrides() -> Result<HashMap<String, String>> {
+    let file = fs::File::open("data/furigana_overrides")?;
+
+    let mut ret = HashMap::new();
+    let re = regex::Regex::new(r#"\|\|\w+\]\]"#)?;
+
+    for line in io::BufReader::new(file).lines() {
+        let line = line?;
+        let line = line.trim();
+        if !line.is_empty() {
+            let clean = re.replace_all(line, "").replace("[[", "");
+            if clean != line {
+                ret.insert(clean, line.to_string());
+            }
+        }
+    }
+
+    Ok(ret)
+}
+
 #[derive(Clone, Debug, Serialize, Deserialize, PartialEq)]
 pub struct JlptVocab {
     pub level: String,
diff --git a/src/example.rs b/src/example.rs
index 71f3f13..770ac8a 100644
--- a/src/example.rs
+++ b/src/example.rs
@@ -1,12 +1,17 @@
-use std::collections::HashSet;
+use std::collections::{HashSet, HashMap};
 
 use crate::charset::Charset;
 use crate::*;
 
 impl Example {
-    pub fn gen_furigana<'a>(&mut self, dict_idx: &DictIndex<'a>) {
+    pub fn gen_furigana<'a>(&mut self, dict_idx: &DictIndex<'a>, overrides: &HashMap<String, String>) {
         use std::fmt::Write;
 
+        if let Some(v) = overrides.get(&self.ja) {
+            self.furigana = Some(v.to_string());
+            return;
+        }
+
         let mut remainder = self.ja.as_str();
         let mut ret = String::new();
 
diff --git a/src/main.rs b/src/main.rs
index 85b278a..b0c46c0 100644
--- a/src/main.rs
+++ b/src/main.rs
@@ -146,16 +146,19 @@ fn main() {
             .expect("parse_jmdict");
             let jmdict_idx = index_jmdict(&jmdict);
 
+            let overrides = read_furigana_overrides().expect("read_furigana_overrides");
+
             for batch in batches.iter_mut() {
                 for ex in batch
                     .examples
                     .iter_mut()
                     .chain(batch.extra_examples.iter_mut())
                 {
-                    ex.gen_furigana(&jmdict_idx);
+                    ex.gen_furigana(&jmdict_idx, &overrides);
                 }
             }
 
+
             save_batches(batches).expect("save_batches");
         }
         Cmd::Format => {
-- 
cgit v1.2.3