I have a srt file named subtitles.srt that is non-English, and I followed the instructions of the documentation and source code of the moviepy package (https://moviepy.readthedocs.io/en/latest/_modules/moviepy/video/tools/subtitles.html):
from moviepy.video.tools.subtitles import SubtitlesClip
from moviepy.video.io.VideoFileClip import VideoFileClip
generator = lambda txt: TextClip(txt, font='Georgia-Regular', fontsize=24, color='white')
sub = SubtitlesClip("subtitles.srt", generator, encoding='utf-8')
And this gives the error TypeError: __init__() got an unexpected keyword argument 'encoding'.
In the source code, the class SubtitlesClip does have a keyword argument encoding. Does that mean the version of the source code is outdated or something? And what can I do about this? I even attempted to copy the source code for moviepy.video.tools.subtitles with the encoding keyword argument directly to my code, yet it led to more errors like at the line:
from moviepy.decorators import convert_path_to_string
it failed to import the decorator convert_path_to_string.
The source code does not seem to agree with what I have installed. Anyway to fix it? If not, are there any good alternatives of Python libraries for inserting subtitles or video editing in general?
Edit: My current solution is to create a child class of SubtitlesClip and override the constructor of the parent class:
from moviepy.video.tools.subtitles import SubtitlesClip
from moviepy.video.VideoClip import TextClip, VideoClip
from moviepy.video.compositing.CompositeVideoClip import CompositeVideoClip
import re
from moviepy.tools import cvsecs
def file_to_subtitles_with_encoding(filename):
""" Converts a srt file into subtitles.
The returned list is of the form ``[((ta,tb),'some text'),...]``
and can be fed to SubtitlesClip.
Only works for '.srt' format for the moment.
"""
times_texts = []
current_times = None
current_text = ""
with open(filename,'r',encoding='utf-8') as f:
for line in f:
times = re.findall("([0-9]*:[0-9]*:[0-9]*,[0-9]*)", line)
if times:
current_times = [cvsecs(t) for t in times]
elif line.strip() == '':
times_texts.append((current_times, current_text.strip('\n')))
current_times, current_text = None, ""
elif current_times:
current_text += line
return times_texts
class SubtitlesClipUTF8(SubtitlesClip):
def __init__(self, subtitles, make_textclip=None):
VideoClip.__init__(self, has_constant_size=False)
if isinstance(subtitles, str):
subtitles = file_to_subtitles_with_encoding(subtitles)
#subtitles = [(map(cvsecs, tt),txt) for tt, txt in subtitles]
self.subtitles = subtitles
self.textclips = dict()
if make_textclip is None:
make_textclip = lambda txt: TextClip(txt, font='Georgia-Bold',
fontsize=24, color='white',
stroke_color='black', stroke_width=0.5)
self.make_textclip = make_textclip
self.start=0
self.duration = max([tb for ((ta,tb), txt) in self.subtitles])
self.end=self.duration
def add_textclip_if_none(t):
""" Will generate a textclip if it hasn't been generated asked
to generate it yet. If there is no subtitle to show at t, return
false. """
sub =[((ta,tb),txt) for ((ta,tb),txt) in self.textclips.keys()
if (ta<=t<tb)]
if not sub:
sub = [((ta,tb),txt) for ((ta,tb),txt) in self.subtitles if
(ta<=t<tb)]
if not sub:
return False
sub = sub[0]
if sub not in self.textclips.keys():
self.textclips[sub] = self.make_textclip(sub[1])
return sub
def make_frame(t):
sub = add_textclip_if_none(t)
return (self.textclips[sub].get_frame(t) if sub
else np.array([[[0,0,0]]]))
def make_mask_frame(t):
sub = add_textclip_if_none(t)
return (self.textclips[sub].mask.get_frame(t) if sub
else np.array([[0]]))
self.make_frame = make_frame
hasmask = bool(self.make_textclip('T').mask)
self.mask = VideoClip(make_mask_frame, ismask=True) if hasmask else None
I actually only changed two lines, but I have to create a new class and redefine the whole thing, so I doubt whether it's really necessary. Any better solution than this?
The latest version in the documentation (the one you are looking at) corresponds to a dev version 2.x, which is not released to PyPI yet. The version you have installed through pip is most likely 1.0.3, which is the latest on PyPI, and it doesn't allow an encoding parameter.
From the PR where the feature was introduced, you can see that it's only been tagged for release in 2.x versions.
Copying only that file to your source code will most likely not work, because it will depend on changes that happened in between the two versions. However, if you feel adventurous, you can install the dev version of the package, by following the Method by hand section in moviepy's docs.