Merge pull request #2357 from OsaAjani/issue-2260

Fix text without spaces overflowing TextClip boundaries
Zulko · Jan 31, 2025 · ff4dca0 · ff4dca0
2 parents 71eb429 + 9d7d668
commit ff4dca0
Show file tree

Hide file tree

Showing 3 changed files with 32 additions and 14 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -33,6 +33,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Fix GPU h264_nvenc encoding not working.
 - Improve perfs of decorator by pre-computing arguments
 - Fix textclip being cut or of impredictable height (see issues #2325, #2260 and #2268)
+- TextClip now properly breaklines on text without spaces, such as chinese (see #2260)
 - Fix TimeMirror and TimeSymmetrize cutting last second of clip
 - ImageSequenceClip was wrong when calculating fps with duration and no fps (see issue #2351)
 - More consistent frame seek (see issue #2115 and PR #2117)

diff --git a/moviepy/video/VideoClip.py b/moviepy/video/VideoClip.py
@@ -1577,7 +1577,9 @@ def __init__(
                 _ = ImageFont.truetype(font)
             except Exception as e:
                 raise ValueError(
-                    "Invalid font {}, pillow failed to use it with error {}".format(font, e)
+                    "Invalid font {}, pillow failed to use it with error {}".format(
+                        font, e
+                    )
                 )
 
         if filename:
@@ -1777,9 +1779,15 @@ def __break_text(
 
         lines = []
         current_line = ""
-        words = text.split(" ")
-        for word in words:
-            temp_line = current_line + " " + word if current_line else word
+
+        # We try to break on spaces as much as possible
+        # if a text dont contain spaces (ex chinese), we will break when possible
+        last_space = 0
+        for index, char in enumerate(text):
+            if char == " ":
+                last_space = index
+
+            temp_line = current_line + char
             temp_left, temp_top, temp_right, temp_bottom = draw.multiline_textbbox(
                 (0, 0),
                 temp_line,
@@ -1790,11 +1798,20 @@ def __break_text(
             )
             temp_width = temp_right - temp_left
 
-            if temp_width <= width:
-                current_line = temp_line
+            if temp_width >= width:
+                # If we had a space previously, add everything up to the space
+                # and reset last_space and current_line else add everything up
+                # to previous char
+                if last_space:
+                    lines.append(temp_line[0:last_space])
+                    current_line = temp_line[last_space + 1 : index + 1]
+                    last_space = 0
+                else:
+                    lines.append(current_line[0:index])
+                    current_line = char
+                    last_space = 0
             else:
-                lines.append(current_line)
-                current_line = word
+                current_line = temp_line
 
         if current_line:
             lines.append(current_line)

diff --git a/moviepy/video/io/ffmpeg_reader.py b/moviepy/video/io/ffmpeg_reader.py
@@ -492,12 +492,12 @@ def parse(self):
                 # for default streams, set their numbers globally, so it's
                 # easy to get without iterating all
                 if self._current_stream["default"]:
-                    self.result[f"default_{stream_type_lower}_input_number"] = (
-                        input_number
-                    )
-                    self.result[f"default_{stream_type_lower}_stream_number"] = (
-                        stream_number
-                    )
+                    self.result[
+                        f"default_{stream_type_lower}_input_number"
+                    ] = input_number
+                    self.result[
+                        f"default_{stream_type_lower}_stream_number"
+                    ] = stream_number
 
                 # exit chapter
                 if self._current_chapter: