Add option to cache fetched emoji to disk for big speedup

[rambling-ai]

This PR adds an option to cache fetched emojis to disk. For my use case this provides a massive boost to performance.

To demonstrate the speed up, I called Pilmoji three times, the first time with disk_cache = False, the next two times with disk_cache=True. Here are the results:

Block 1 duration: 1.4033 seconds
Block 2 duration: 1.3994 seconds
Block 3 duration: 0.0041 seconds

The first two calls are roughly equivalent, but the third call benefits from the emojis already having been fetched and stored to disk locally.

And here's the speed test code:

import time
from pilmoji import Pilmoji
from PIL import Image

test_image = Image.new("RGBA", (1080, 1920), (255, 255, 255, 0))
x = 100
y = 100
text_color = (255, 255, 255)
line = "⏱️⏳🏎️🐇🐢💨🚀⏩"

# Block 1
start_time = time.time()
with Pilmoji(test_image) as pilmoji:
    pilmoji.text((x, y), line, text_color)
end_time = time.time()
print(f"Block 1 duration: {end_time - start_time:.4f} seconds")

# Block 2
start_time = time.time()
with Pilmoji(test_image, disk_cache=True) as pilmoji:
    pilmoji.text((x, y), line, text_color)
end_time = time.time()
print(f"Block 2 duration: {end_time - start_time:.4f} seconds")

# Block 3
start_time = time.time()
with Pilmoji(test_image, disk_cache=True) as pilmoji:
    pilmoji.text((x, y), line, text_color)
end_time = time.time()
print(f"Block 3 duration: {end_time - start_time:.4f} seconds")

This speed-up made it possible to use Pilmoji for my use case, so I hope it can help others too!