Generate Subtitles With Highlighting And Stroke (Outline)

"""
Subtitles in moviepy are still annoying, but doable

I struggled with this, so maybe somebody else did too.  You can't use pango with stroke (stroke_width) so if you want to change one word, you're screwed. I run this on linux cuz it's easier than figuring out how to get ImageMagick working on Windows, lol

I don't do the timings manually (I'd die).  I use AWS Transcribe, and as long as it's less than 60 minutes a month, it's free for the first year.  It's cheap if you go over anyway

The video is generated with Deforum notebook (Stable Diffusion)

If you navigated here from somewhere other than my video, here's a link to see the end result

https://youtu.be/YN_XWoMTGdY
"""

import math
from moviepy.editor import *
from moviepy.video.tools import *

screensize = (1024, 768)
audioclip = AudioFileClip("untitled.mp3")
v4 = VideoFileClip('20230205043321.mp4')
clip_list = [v4]


def do_text(text, font_size, main_color, highlight_color, highlight_index, start_time, end_time, y_pos= 573, wobble = 1.0, backcolor = "white"):
    def lambda_with_var(i, yi, w):
        return lambda t: (i, yi + w * t * 10 * math.sin(t * 3 + i))
    items = []
    widths = [0]
    te = text.split(' ')
    co = [main_color] * len(te)
    if highlight_index >= 0 and highlight_index < len(co):
        co[highlight_index] = highlight_color

    for i in range(len(te)):

        txt_clip = TextClip(te[i], method='label', fontsize=font_size, color=co[i], stroke_color=backcolor,
                            stroke_width=1.5, font='URW-Chancery-L-Medium-Italic').set_duration(end_time - start_time)
        txt_clip = txt_clip.set_start(start_time)
        widths.append(widths[-1] + txt_clip.size[0] + 10)
        items.append(txt_clip)

    assert widths[-1] < screensize[0], f"Too damn wide! {widths[-1]}"
    indent = (screensize[0] - widths[-1]) // 2
    for i in range(len(items)):
        items[i] = items[i].set_pos(lambda_with_var(widths[i] + indent, y_pos, wobble))

    clip_list.extend(items)


do_text("I want to own a sword", 120, "purple", "red", 0, 0.39, 0.62)
do_text("I want to own a sword", 120, "purple", "red", 1, 0.62, 1.06)
do_text("I want to own a sword", 120, "purple", "red", 2, 1.06, 1.19)
do_text("I want to own a sword", 120, "purple", "red", 3, 1.2, 1.44)
do_text("I want to own a sword", 120, "purple", "red", 4, 1.44, 1.5)
do_text("I want to own a sword", 120, "purple", "red", 5, 1.5, 2.17)
do_text("But not because Andrew Tate", 90, "purple", "red", 0, 2.17, 2.36, 500)
do_text("But not because Andrew Tate", 90, "purple", "red", 1, 2.36, 2.63, 500)
do_text("But not because Andrew Tate", 90, "purple", "red", 2, 2.63, 2.97, 500)
do_text("But not because Andrew Tate", 90, "purple", "red", 3, 2.97, 3.38, 500)
do_text("But not because Andrew Tate", 90, "purple", "red", 4, 3.38, 3.7, 500)
do_text("But not because Andrew Tate", 90, "purple", "red", -1, 3.7, 4.94, 500)
do_text("told me that I should!", 90, "purple", "red", -1, 2.17, 3.7, 620)
do_text("told me that I should!", 90, "purple", "red", 0, 3.7, 3.99, 620)
do_text("told me that I should!", 90, "purple", "red", 1, 3.99, 4.14, 620)
do_text("told me that I should!", 90, "purple", "red", 2, 4.14, 4.29, 620)
do_text("told me that I should!", 90, "purple", "red", 3, 4.29, 4.4, 620)
do_text("told me that I should!", 90, "purple", "red", 4, 4.4, 4.94, 620)
do_text("Also, I think I figured", 110, "pink", "red", -1, 5.79, 8.2, 500, backcolor = "black", wobble=0)
do_text("out the subtitle thing.", 110, "yellow", "red", -1, 5.79, 8.2, 620, backcolor = "black", wobble=0)
final_clip = CompositeVideoClip(clip_list).set_audio(audioclip)
final_clip.write_videofile("my_concatenation.mp4", fps=24, codec='libx264', threads=7)