enhancement: improve english text normalization in english.py

apps/mofa-asr/src/screen/mod.rs

@@ -262,15 +262,27 @@ impl Widget for MoFaASRScreen {
         if self.paraformer_chat_controller.is_none() {
             let controller = ChatController::new_arc();
             {
-                let mut guard = controller.lock().expect("ChatController mutex poisoned");
+                let mut guard = match controller.lock() {
+                    Ok(g) => g,
+                    Err(poisoned) => {
+                        ::log::warn!("ChatController mutex poisoned; recovering inner state");
+                        poisoned.into_inner()
+                    }
+                };
                 guard.dangerous_state_mut().bots.push(Bot {
                     id: BotId::new("asr"),
                     name: "Paraformer".to_string(),
                     avatar: EntityAvatar::Text("P".to_string()),
                     capabilities: BotCapabilities::new(),
                 });
             }
-            self.paraformer_chat_controller = Some(controller.clone());
+            self.paraformer_chatmatch controller.lock() {
+                    Ok(g) => g,
+                    Err(poisoned) => {
+                        ::log::warn!("ChatController mutex poisoned; recovering inner state");
+                        poisoned.into_inner()
+                    }
+                }
             self.view.messages(ids!(paraformer_messages)).write().chat_controller = Some(controller);
         }
         if self.qwen3_chat_controller.is_none() {
@@ -423,7 +435,13 @@ impl MoFaASRScreen {
         };
 
         let count = {
-            let mut guard = controller.lock().expect("ChatController mutex poisoned");
+            let mut guard = match controller.lock() {
+                Ok(g) => g,
+                Err(poisoned) => {
+                    ::log::warn!("ChatController mutex poisoned; recovering inner state");
+                    poisoned.into_inner()
+                }
+            };
             let state = guard.dangerous_state_mut();
             state.messages.clear();
             for msg in messages {
@@ -465,7 +483,14 @@ impl MoFaASRScreen {
     fn handle_start(&mut self, cx: &mut Cx) {
         // Clear per-engine chat controllers
         if let Some(ref controller) = self.paraformer_chat_controller {
-            controller.lock().expect("ChatController mutex poisoned").dangerous_state_mut().messages.clear();
+            let mut guard = match controller.lock() {
+                Ok(g) => g,
+                Err(poisoned) => {
+                    ::log::warn!("ChatController mutex poisoned; recovering inner state");
+                    poisoned.into_inner()
+                }
+            };
+            guard.dangerous_state_mut().messages.clear();
         }
         if let Some(ref controller) = self.qwen3_chat_controller {
             controller.lock().expect("ChatController mutex poisoned").dangerous_state_mut().messages.clear();

doc/CHECKLIST.md

@@ -543,7 +543,7 @@ pub enum TabId {
 - [x] Simplified code: `contains(&TabId::Profile)` instead of `iter().any(|t| t == "profile")`
 
 **Benefits**:
-- Compile-time checking prevents typos like `"profiel"` or `"setings"`
+- Compile-time checking prevents typos like `"profiel"` or `"settgs"`
 - IDE autocomplete works with enum variants
 - Exhaustive match ensures all cases handled
 - `Copy` trait allows efficient passing without `.clone()` or `.to_string()`

mofa-dora-bridge/src/parser.rs

@@ -430,7 +430,8 @@ nodes:
       tts_log: tts/log
 "#;
 
-        let parsed = DataflowParser::parse_string(yaml, PathBuf::from("test.yml")).unwrap();
+        let parsed = DataflowParser::parse_string(yaml, PathBuf::from("test.yml"))
+            .expect("DataflowParser failed to parse a valid test YAML; check parser changes");
 
         assert_eq!(parsed.mofa_nodes.len(), 2);
         assert_eq!(parsed.mofa_nodes[0].id, "mofa-audio-player");

node-hub/dora-funasr-nano-mlx/src/main.rs

@@ -109,7 +109,14 @@ fn main() -> Result<()> {
                             }
                         }
 
-                        let engine = engine.as_mut().unwrap();
+                        let engine = match engine.as_mut() {
+                            Some(e) => e,
+                            None => {
+                                log::error!("Engine unexpectedly missing after attempted initialization");
+                                send_log(&mut node, "ERROR", "Engine not available")?;
+                                continue;
+                            }
+                        };
 
                         // Extract metadata
                         let question_id = metadata

node-hub/dora-gpt-sovits-mlx/src/ssml.rs

@@ -354,7 +354,8 @@ mod tests {
 
     #[test]
     fn test_plain_text() {
-        let result = parse_ssml("<speak>hello world</speak>").unwrap();
+        let result = parse_ssml("<speak>hello world</speak>")
+            .expect("parse_ssml should parse simple <speak> content");
         assert_eq!(result, vec![SsmlSegment::Text {
             text: "hello world".to_string(),
             speed: 1.0,

node-hub/dora-primespeech/dora_primespeech/moyoyo_tts/text/english.py

@@ -219,11 +219,20 @@ def get_namedict():
 
 
 def text_normalize(text):
-    # todo: eng text normalize
     # 适配中文及 g2p_en 标点
+    if not text:
+        return ""
+
+    # ensure string
+    text = str(text)
+
+    # punctuation compatibility (smart quotes, dashes, ellipsis, CJK punctuation)
     rep_map = {
         "[;:：，；]": ",",
-        '["’]': "'",
+        "[\\u2018\\u2019`‘’]": "'",
+        "[\\u201c\\u201d\\u00ab\\u00bb\\u201e\\\"]": '"',
+        "\\u2026": "...",
+        "[\\u2013\\u2014]": "-",
         "。": ".",
         "！": "!",
         "？": "?",
@@ -233,16 +242,36 @@ def text_normalize(text):
 
     # 来自 g2p_en 文本格式化处理
     # 增加大写兼容
-    text = unicode(text)
-    text = normalize_numbers(text)
-    text = ''.join(char for char in unicodedata.normalize('NFD', text)
-                    if unicodedata.category(char) != 'Mn')  # Strip accents
-    text = re.sub("[^ A-Za-z'.,?!\-]", "", text)
-    text = re.sub(r"(?i)i\.e\.", "that is", text)
-    text = re.sub(r"(?i)e\.g\.", "for example", text)
+    try:
+        text = normalize_numbers(text)
+    except Exception:
+        pass
+
+    # strip accents
+    text = "".join(
+        ch for ch in unicodedata.normalize('NFD', text)
+        if unicodedata.category(ch) != 'Mn'
+    )
+
+    # keep letters, digits, quotes/apostrophes, basic punctuation and hyphen
+    text = re.sub(r"[^ 0-9a-z'\".,?!\-]", "", text)
+
+    # expand abbreviations (word boundaries, case-insensitive)
+    text = re.sub(r"\bi\.e\.\b", "that is", text, flags=re.IGNORECASE)
+    text = re.sub(r"\be\.g\.\b", "for example", text, flags=re.IGNORECASE)
 
     # 避免重复标点引起的参考泄露
-    text = replace_consecutive_punctuation(text)
+    try:
+        text = replace_consecutive_punctuation(text)
+    except Exception:
+        text = re.sub(r"([.,?!\-])\1+", r"\1", text)
+
+    # normalize whitespace
+    text = re.sub(r"\s+", " ", text).strip()
+
+    # ensure terminal punctuation for TTS stability
+    if text and text[-1] not in ".?!":
+        text += "."
 
     return text