From a6a4423a68c1018451a490d378f7f5ad62b3f5fc Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?=E5=B1=85=E6=88=8E=E6=B0=8F?= <chen.sst@gmail.com>
Date: Thu, 16 Jan 2025 10:38:38 +0800
Subject: [PATCH] feat(punctuator): auto-convert digit separator

fixes #972

use ascii punctuation ,.:' after numbers.
they are auto-committed if followed by digit.
or commit manualy with space key.
double strike the key to access the original binding.
---
 src/rime/gear/punctuator.cc | 86 +++++++++++++++++++++++++++++--------
 src/rime/gear/punctuator.h  |  3 +-
 2 files changed, 70 insertions(+), 19 deletions(-)

diff --git a/src/rime/gear/punctuator.cc b/src/rime/gear/punctuator.cc
index a391b85159..ed439c3d7a 100644
--- a/src/rime/gear/punctuator.cc
+++ b/src/rime/gear/punctuator.cc
@@ -58,6 +58,30 @@ static bool punctuation_is_translated(Context* ctx) {
   return cand && cand->type() == "punct";
 }
 
+inline static bool is_digit_separator(char ch) {
+  return ch == '.' || ch == ':' || ch == ',' || ch == '\'';
+}
+
+inline static bool ends_with_digit(const string& text) {
+  auto len = text.length();
+  return len > 0 && isdigit(text[len - 1]);
+}
+
+// recognizes patterns like 3.14 12:30 1,000 1'000
+static bool is_after_number(Context* ctx) {
+  const CommitHistory& history = ctx->commit_history();
+  if (history.empty()) {
+    return false;
+  }
+  const CommitRecord& cr = history.back();
+  return ends_with_digit(cr.text) & (cr.type == "thru" || cr.type == "raw");
+}
+
+static bool is_after_digit_separator(Context* ctx) {
+  const auto& comp = ctx->composition();
+  return !comp.empty() && comp.back().HasTag("punct_number");
+}
+
 ProcessResult Punctuator::ProcessKeyEvent(const KeyEvent& key_event) {
   if (key_event.release() || key_event.ctrl() || key_event.alt() ||
       key_event.super())
@@ -72,29 +96,48 @@ ProcessResult Punctuator::ProcessKeyEvent(const KeyEvent& key_event) {
   if (!use_space_ && ch == XK_space && ctx->IsComposing()) {
     return kNoop;
   }
-  if (ch == '.' || ch == ':') {  // 3.14, 12:30
-    const CommitHistory& history(ctx->commit_history());
-    if (!history.empty()) {
-      const CommitRecord& cr(history.back());
-      if (cr.type == "thru" && cr.text.length() == 1 && isdigit(cr.text[0])) {
-        return kRejected;
-      }
-    }
+  if (isdigit(ch) && is_after_digit_separator(ctx)) {
+    ctx->PushInput(ch) && ctx->Commit();
+    return kAccepted;
   }
+  // sync with full_shape option
   config_.LoadConfig(engine_);
-  string punct_key(1, ch);
-  auto punct_definition = config_.GetPunctDefinition(punct_key);
+  string key(1, ch);
+  auto punct_definition = config_.GetPunctDefinition(key);
   if (!punct_definition)
     return kNoop;
-  DLOG(INFO) << "punct key: '" << punct_key << "'";
-  if (!AlternatePunct(punct_key, punct_definition)) {
-    ctx->PushInput(ch) && punctuation_is_translated(ctx) &&
-        (ConfirmUniquePunct(punct_definition) ||
-         AutoCommitPunct(punct_definition) || PairPunct(punct_definition));
+  DLOG(INFO) << "punct key: '" << key << "'";
+  if (AlternatePunct(key, punct_definition)) {
+    return kAccepted;
+  }
+  if (ToggleNumberMode(key) || ctx->PushInput(ch)) {
+    if (punctuation_is_translated(ctx)) {
+      ConfirmUniquePunct(punct_definition) ||
+          AutoCommitPunct(punct_definition) || PairPunct(punct_definition);
+    }
   }
   return kAccepted;
 }
 
+bool Punctuator::ToggleNumberMode(const string& key) {
+  Context* ctx = engine_->context();
+  if (ctx->input() == key) {
+    Composition& comp = ctx->composition();
+    if (!comp.empty()) {
+      Segment& segment(comp.back());
+      if (segment.HasTag("punct_number")) {
+        segment.tags.erase("punct_number");
+        segment.tags.insert("punct");
+        segment.status = Segment::kVoid;
+        DLOG(INFO) << "exit number mode, key = " << key;
+        ctx->update_notifier()(ctx);
+        return true;
+      }
+    }
+  }
+  return false;
+}
+
 bool Punctuator::AlternatePunct(const string& key,
                                 const an<ConfigItem>& definition) {
   if (!As<ConfigList>(definition))
@@ -170,16 +213,22 @@ bool PunctSegmentor::Proceed(Segmentation* segmentation) {
   char ch = input[k];
   if (ch < 0x20 || ch >= 0x7f)
     return true;
+  // sync with full_shape option
   config_.LoadConfig(engine_);
-  string punct_key(1, ch);
-  auto punct_definition = config_.GetPunctDefinition(punct_key);
+  string key(1, ch);
+  auto punct_definition = config_.GetPunctDefinition(key);
   if (!punct_definition)
     return true;
   {
     Segment segment(k, k + 1);
     DLOG(INFO) << "add a punctuation segment [" << segment.start << ", "
                << segment.end << ")";
-    segment.tags.insert("punct");
+    if (k == 0 && is_digit_separator(ch) &&
+        is_after_number(engine_->context())) {
+      segment.tags.insert("punct_number");
+    } else {
+      segment.tags.insert("punct");
+    }
     segmentation->AddSegment(segment);
   }
   return false;  // exclusive
@@ -235,6 +284,7 @@ an<Translation> PunctTranslator::Query(const string& input,
                                        const Segment& segment) {
   if (!segment.HasTag("punct"))
     return nullptr;
+  // sync with full_shape option
   config_.LoadConfig(engine_);
   auto definition = config_.GetPunctDefinition(input);
   if (!definition)
diff --git a/src/rime/gear/punctuator.h b/src/rime/gear/punctuator.h
index 836ec3aabb..beb4d8dfac 100644
--- a/src/rime/gear/punctuator.h
+++ b/src/rime/gear/punctuator.h
@@ -35,8 +35,9 @@ class Punctuator : public Processor {
   virtual ProcessResult ProcessKeyEvent(const KeyEvent& key_event);
 
  protected:
-  bool ConfirmUniquePunct(const an<ConfigItem>& definition);
+  bool ToggleNumberMode(const string& key);
   bool AlternatePunct(const string& key, const an<ConfigItem>& definition);
+  bool ConfirmUniquePunct(const an<ConfigItem>& definition);
   bool AutoCommitPunct(const an<ConfigItem>& definition);
   bool PairPunct(const an<ConfigItem>& definition);