Refactor quoted dialogue detection (#2138)

novelwriter/text/patterns.py

@@ -28,11 +28,13 @@ import re
 
 from novelwriter import CONFIG
 from novelwriter.common import compact, uniqueCompact
-from novelwriter.constants import nwRegEx
+from novelwriter.constants import nwRegEx, nwUnicode
 
 
 class RegExPatterns:
 
+    AMBIGUOUS = (nwUnicode.U_APOS, nwUnicode.U_RSQUO)
+
     # Static RegExes
     _rxUrl     = re.compile(nwRegEx.URL, re.ASCII)
     _rxWords   = re.compile(nwRegEx.WORDS, re.UNICODE)
@@ -87,16 +89,25 @@ class RegExPatterns:
     def dialogStyle(self) -> re.Pattern | None:
         """Dialogue detection rule based on user settings."""
         if CONFIG.dialogStyle > 0:
-            end = "|$" if CONFIG.allowOpenDial else ""
             rx = []
             if CONFIG.dialogStyle in (1, 3):
                 qO = CONFIG.fmtSQuoteOpen.strip()[:1]
                 qC = CONFIG.fmtSQuoteClose.strip()[:1]
-                rx.append(f"(?:\\B{qO}.*?(?:{qC}\\B{end}))")
+                if qO == qC:
+                    rx.append(f"(?:\\B{qO}.+?{qC}\\B)")
+                else:
+                    rx.append(f"(?:{qO}[^{qO}]+{qC})")
+                if CONFIG.allowOpenDial:
+                    rx.append(f"(?:{qO}.+?$)")
             if CONFIG.dialogStyle in (2, 3):
                 qO = CONFIG.fmtDQuoteOpen.strip()[:1]
                 qC = CONFIG.fmtDQuoteClose.strip()[:1]
-                rx.append(f"(?:\\B{qO}.*?(?:{qC}\\B{end}))")
+                if qO == qC:
+                    rx.append(f"(?:\\B{qO}.+?{qC}\\B)")
+                else:
+                    rx.append(f"(?:{qO}[^{qO}]+{qC})")
+                if CONFIG.allowOpenDial:
+                    rx.append(f"(?:{qO}.+?$)")
             return re.compile("|".join(rx), re.UNICODE)
         return None
 
@@ -106,7 +117,8 @@ class RegExPatterns:
         if CONFIG.altDialogOpen and CONFIG.altDialogClose:
             qO = re.escape(compact(CONFIG.altDialogOpen))
             qC = re.escape(compact(CONFIG.altDialogClose))
-            return re.compile(f"\\B{qO}.*?{qC}\\B", re.UNICODE)
+            qB = r"\B" if (qO == qC or qC in self.AMBIGUOUS) else ""
+            return re.compile(f"{qO}.*?{qC}{qB}", re.UNICODE)
         return None
 
 

tests/test_text/test_text_patterns.py

@@ -312,8 +312,27 @@ def testTextPatterns_DialogueStyle():
     # Straight double quotes are ignored
     assert allMatches(regEx, "one \"two\" three") == []
 
-    # Skipping whitespace is not allowed
+    # Check with no whitespace, single quote
-    assert allMatches(regEx, "one\u2018two\u2019three") == []
+    assert allMatches(regEx, "one\u2018two\u2019three") == [
+        [("\u2018two\u2019", 3, 8)]
+    ]
+    assert allMatches(regEx, "one\u2018two\u2019 three") == [
+        [("\u2018two\u2019", 3, 8)]
+    ]
+
+    # Check with no whitespace, double quote
+    assert allMatches(regEx, "one\u201ctwo\u201dthree") == [
+        [("\u201ctwo\u201d", 3, 8)]
+    ]
+    assert allMatches(regEx, "one\u201ctwo\u201d three") == [
+        [("\u201ctwo\u201d", 3, 8)]
+    ]
+
+    # Check with apostrophe
+    assert allMatches(regEx, "one \u2018two\u2019s three\u2019, \u2018four\u2019 five") == [
+        [("\u2018two\u2019s three\u2019", 4, 17)],
+        [("\u2018four\u2019", 19, 25)],
+    ]
 
     # Open
     # ====
@@ -333,6 +352,63 @@ def testTextPatterns_DialogueStyle():
     ]
 
 
+@pytest.mark.core
+def testTextPatterns_DialoguePlain():
+    """Test the dialogue style pattern regexes for plain quotes."""
+    # Set the config
+    CONFIG.fmtSQuoteOpen  = "'"
+    CONFIG.fmtSQuoteClose = "'"
+    CONFIG.fmtDQuoteOpen  = '"'
+    CONFIG.fmtDQuoteClose = '"'
+
+    CONFIG.dialogStyle = 3
+    CONFIG.allowOpenDial = False
+    regEx = REGEX_PATTERNS.dialogStyle
+    assert regEx is not None
+
+    # Double
+    # ======
+
+    # One double quoted string
+    assert allMatches(regEx, "one \"two\" three") == [
+        [("\"two\"", 4, 9)]
+    ]
+
+    # Two double quoted strings
+    assert allMatches(regEx, "one \"two\" three \"four\" five") == [
+        [("\"two\"", 4, 9)], [("\"four\"", 16, 22)],
+    ]
+
+    # No space
+    assert allMatches(regEx, "one\"two\" three") == []
+    assert allMatches(regEx, "one \"two\"three") == []
+    assert allMatches(regEx, "one\"two\"three") == []
+
+    # Single
+    # ======
+
+    # One single quoted string
+    assert allMatches(regEx, "one 'two' three") == [
+        [("'two'", 4, 9)]
+    ]
+
+    # Two single quoted strings
+    assert allMatches(regEx, "one 'two' three 'four' five") == [
+        [("'two'", 4, 9)], [("'four'", 16, 22)],
+    ]
+
+    # No space
+    assert allMatches(regEx, "one'two' three") == []
+    assert allMatches(regEx, "one 'two'three") == []
+    assert allMatches(regEx, "one'two'three") == []
+
+    # Check with apostrophe
+    assert allMatches(regEx, "one 'two's three', 'four' five") == [
+        [("'two's three'", 4, 17)],
+        [("'four'", 19, 25)],
+    ]
+
+
 @pytest.mark.core
 def testTextPatterns_DialogueSpecial():
     """Test the special dialogue style pattern regexes."""