Inconsistency between direct source execution and command line execution for Chinese subtitle generation

[sumanit]

When executing the subtitle generation logic directly from the source code, it correctly segments sentences based on punctuation, and there are no spaces between the characters, which is the expected behavior for Chinese text.
However, when running the same logic via the command line interface, the sentence segmentation appears to be inaccurate, and there are unexpected spaces between Chinese characters.

input：
在忙碌和挑战中，我们的内心有时会感到疲惫。尤其是当我们发现自己脱发时，

sourceCode：

00:00:00,100 --> 00:00:01,538
在忙碌和挑战中

00:00:01,625 --> 00:00:04,237
我们的内心有时会感到疲惫

00:00:04,787 --> 00:00:07,225
尤其是当我们发现自己脱发时

command line:
00:00:00.100 --> 00:00:03.400
在 忙碌 和 挑战 中 我们 的 内心 有时 会

00:00:03.413 --> 00:00:07.875
感到 疲惫 尤其是 当 我们 发现 自己 脱发 时 心里

[rany2]

Does this still happen?

[rany2]

Nevermind I see what you mean now. It's the same issue as #156