jupyter-ydoc 3.3.2__py3-none-any.whl → 3.3.3__py3-none-any.whl

jupyter_ydoc/_version.py

@@ -1,4 +1,4 @@
 # This file is auto-generated by Hatchling. As such, do not:
 #   - modify
 #   - track in version control e.g. be sure to add to .gitignore
-__version__ = VERSION = '3.3.2'
+__version__ = VERSION = '3.3.3'

jupyter_ydoc/yunicode.py

@@ -74,30 +74,66 @@ class YUnicode(YBaseDoc):
             # to avoid side-effects such as cursor jumping to the top
             return
 
+        before_bytes = old_value.encode("utf-8")
+        after_bytes = value.encode("utf-8")
+
         with self._ydoc.transaction():
-            matcher = SequenceMatcher(a=old_value, b=value)
+            matcher = SequenceMatcher(a=before_bytes, b=after_bytes)
 
             if (
                 matcher.real_quick_ratio() >= SIMILARITY_THREESHOLD
                 and matcher.ratio() >= SIMILARITY_THREESHOLD
             ):
                 operations = matcher.get_opcodes()
-                offset = 0
+
+                # Fix byte ranges and check for problematic overlaps
+                fixed_operations = []
+                prev_end = 0
+                prev_tag = None
+                has_overlap = False
+
                 for tag, i1, i2, j1, j2 in operations:
-                    match tag:
-                        case "replace":
-                            self._ysource[i1 + offset : i2 + offset] = value[j1:j2]
-                            offset += (j2 - j1) - (i2 - i1)
-                        case "delete":
-                            del self._ysource[i1 + offset : i2 + offset]
-                            offset -= i2 - i1
-                        case "insert":
-                            self._ysource.insert(i1 + offset, value[j1:j2])
-                            offset += j2 - j1
-                        case "equal":
-                            pass
-                        case _:
-                            raise ValueError(f"Unknown tag '{tag}' in sequence matcher")
+                    # Fix byte ranges to proper UTF-8 character boundaries
+                    i1_fixed, i2_fixed = _fix_byte_range_to_char_boundary(before_bytes, i1, i2)
+                    j1_fixed, j2_fixed = _fix_byte_range_to_char_boundary(after_bytes, j1, j2)
+
+                    # Check if this operation overlaps with the previous one
+                    # which can happen with grapheme clusters (emoji + modifiers, etc.)
+                    if i1_fixed < prev_end and prev_tag != "equal":
+                        has_overlap = True
+                        break
+
+                    prev_end = i2_fixed
+                    prev_tag = tag
+                    fixed_operations.append((tag, i1_fixed, i2_fixed, j1_fixed, j2_fixed))
+
+                # If we detected overlapping operations, fall back to hard reload
+                if has_overlap:
+                    self._ysource.clear()
+                    if value:
+                        self._ysource += value
+                else:
+                    # Apply granular operations
+                    offset = 0
+                    for tag, i1, i2, j1, j2 in fixed_operations:
+                        match tag:
+                            case "replace":
+                                self._ysource[i1 + offset : i2 + offset] = after_bytes[
+                                    j1:j2
+                                ].decode("utf-8")
+                                offset += (j2 - j1) - (i2 - i1)
+                            case "delete":
+                                del self._ysource[i1 + offset : i2 + offset]
+                                offset -= i2 - i1
+                            case "insert":
+                                self._ysource.insert(
+                                    i1 + offset, after_bytes[j1:j2].decode("utf-8")
+                                )
+                                offset += j2 - j1
+                            case "equal":
+                                pass
+                            case _:
+                                raise ValueError(f"Unknown tag '{tag}' in sequence matcher")
             else:
                 # for very different strings, just replace the whole content;
                 # this avoids generating a huge number of operations
@@ -118,3 +154,28 @@ class YUnicode(YBaseDoc):
         self.unobserve()
         self._subscriptions[self._ystate] = self._ystate.observe(partial(callback, "state"))
         self._subscriptions[self._ysource] = self._ysource.observe(partial(callback, "source"))
+
+
+def _is_utf8_continuation_byte(byte: int) -> bool:
+    """Check if a byte is a UTF-8 continuation byte (10xxxxxx)."""
+    return (byte & 0xC0) == 0x80
+
+
+def _fix_byte_range_to_char_boundary(data: bytes, start: int, end: int) -> tuple[int, int]:
+    """
+    Adjust byte indices to proper UTF-8 character boundaries.
+
+    :param data: The byte data.
+    :param start: The start byte index.
+    :param end: The end byte index.
+    :return: A tuple of (adjusted_start, adjusted_end).
+    """
+    # Move start backward to the beginning of a UTF-8 character
+    while start > 0 and start < len(data) and _is_utf8_continuation_byte(data[start]):
+        start -= 1
+
+    # Move end forward to the end of a UTF-8 character
+    while end < len(data) and _is_utf8_continuation_byte(data[end]):
+        end += 1
+
+    return start, end

{jupyter_ydoc-3.3.2.dist-info → jupyter_ydoc-3.3.3.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: jupyter-ydoc
-Version: 3.3.2
+Version: 3.3.3
 Summary: Document structures for collaborative editing using Ypy
 Project-URL: Homepage, https://jupyter.org
 Project-URL: Source, https://github.com/jupyter-server/jupyter_ydoc

{jupyter_ydoc-3.3.2.dist-info → jupyter_ydoc-3.3.3.dist-info}/RECORD

@@ -1,14 +1,14 @@
 jupyter_ydoc/__init__.py,sha256=itUidK7o0_wS6YcbKKIyt1su7hM3-YppQshFheTQQdw,428
-jupyter_ydoc/_version.py,sha256=v4KIZUrlasa2fZGLAsEoZ179ShRb4xGFjH2KEtqJAUY,171
+jupyter_ydoc/_version.py,sha256=N8PaAgfH_KWhDzwQorENf6K8gehfU0qesXEFAAPh4ZY,171
 jupyter_ydoc/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 jupyter_ydoc/utils.py,sha256=yKvcuqhpylMinmjuscuZ_kY8KPEseFbwcg5K9VzYOfs,810
 jupyter_ydoc/ybasedoc.py,sha256=c0jwhULtTNCjOYHbXhDhKaD6OJYn7hpL4hcLZWyGJsU,5115
 jupyter_ydoc/yblob.py,sha256=JZiXQhONqFS8Cqdglx__AVeS18gyRq0yHq-AQKFPVfw,2316
 jupyter_ydoc/yfile.py,sha256=XTMtAXDWgIOLU2KUQxkLJz2cGvSPlOxpvJc4daXCV6I,198
 jupyter_ydoc/ynotebook.py,sha256=Xplr1x57Y6gWAU84u-Krh9_Xkn_-AN-Vte4DJY5VOfw,13452
-jupyter_ydoc/yunicode.py,sha256=nPNhXLT06TwxvbZL8IhFiMvRYkFoS6Hnpm1MNVlW4b4,4004
-jupyter_ydoc-3.3.2.dist-info/METADATA,sha256=-2RhQiCsCQ_aryHI6YjqCnYIeL06LzjENOxtRduz8_8,2282
-jupyter_ydoc-3.3.2.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
-jupyter_ydoc-3.3.2.dist-info/entry_points.txt,sha256=lgvRG-rpsjRKf8cy7LpO7fqwwXy0sBVMCwhGOHgn4mc,164
-jupyter_ydoc-3.3.2.dist-info/licenses/LICENSE,sha256=dqphsFbhnlzPK7Vlkc66Zc7g7PS-e1dln07GXIVpFCQ,1567
-jupyter_ydoc-3.3.2.dist-info/RECORD,,
+jupyter_ydoc/yunicode.py,sha256=gmWNvkItUlbHAtc0JeVdq2ffAUMi9gr9lE83vDuNotE,6491
+jupyter_ydoc-3.3.3.dist-info/METADATA,sha256=gE7ys9Pdw3JCEmVT0s6QrVrd7fejgGvHnGVeHzp6pk4,2282
+jupyter_ydoc-3.3.3.dist-info/WHEEL,sha256=WLgqFyCfm_KASv4WHyYy0P3pM_m7J5L9k2skdKLirC8,87
+jupyter_ydoc-3.3.3.dist-info/entry_points.txt,sha256=lgvRG-rpsjRKf8cy7LpO7fqwwXy0sBVMCwhGOHgn4mc,164
+jupyter_ydoc-3.3.3.dist-info/licenses/LICENSE,sha256=dqphsFbhnlzPK7Vlkc66Zc7g7PS-e1dln07GXIVpFCQ,1567
+jupyter_ydoc-3.3.3.dist-info/RECORD,,