Dmitriy Kovalev dmitriykovalev

Decibel

bits	2 ^ bits	dB (power)	dB (amplitude)
1	2	3.01	6.02
2	4	6.02	12.04
3	8	9.03	18.06
4	16	12.04	24.08
5	32	15.05	30.10
6	64	18.06	36.12

Quick Action: Play in mpv (macOS)

You need to have [mpv][mpv] installed. For example, you can run brew install mpv for that.

Run the built-in [Automator][automator] app and create a new "Quick Action". Name it "Play in mpv" and set it up according to the screenshot below:

Workflow receives current files or folders in any application
Input is entire selection
Image ▶ Play

Python: CheckRange for argparse

The Python argparse module lacks a built-in mechanism for validating argument ranges. For instance, while you can specify an argument type as int, there is no straightforward way to restrict it to only positive values. Similarly, when dealing with float arguments, there is no native method to ensure that the provided values fall within the range of [0.0, 1.0).

CheckRange is a possible solution to this problem. It is a subclass of argparse.Action. It allows you to validate open, closed, and half-open intervals. Each endpoint can be either a number or positive or negative infinity:

[a, b] → min=a, max=b

	import math

	def print_audio_frames(num_samples, freq_hz, sample_rate_hz, num_channels):
	for i in range(num_samples):
	a = math.sin(2 * math.pi * freq_hz * i / sample_rate_hz)
	print(f'{i // num_channels:4d} => ', end='')
	for j in range(num_channels):
	value = ((1 << (15 - j)) - 1) * a;
	print(f'{int(value):6d}', end='')
	print()

	import argparse
	import hashlib
	import math
	import struct
	import wave

	def rms(samples):
	return math.sqrt(sum(s*s for s in samples) / len(samples))

	def md5(data):

	#!/bin/bash

	osascript -l JavaScript > $(date '+%Y.%m.%d-%H.%M.%S')-chrome-tabs.json <<EOF
	var app = Application('Google Chrome');

	let result = app.windows().map((w) => {
	return {
	'name': w.name(),
	'tabs': w.tabs().map((t) => {
	return {'title': t.title(), 'url': t.url()};

	from math import degrees, radians, tan, atan, sqrt

	# Diagonal FOV
	dfov = 90
	# Sensor size
	width, height = 4032, 3024
	r = width / height
	# Horizontal FOV (corresponds to width)
	hfov = degrees(2 * atan(tan(radians(dfov / 2)) / sqrt(1 + 1 / (r*r))))
	# Vertial FOV (corresponds to height)

	import argparse
	from PIL import Image

	def nv12_to_rgb(nv12: bytes \| bytearray, size: tuple[int, int]) -> Image:
	w, h = size
	n = w * h
	y, u, v = nv12[:n], nv12[n + 0::2], nv12[n + 1::2]
	yuv = bytearray(3 * n)
	yuv[0::3] = y
	yuv[1::3] = Image.frombytes('L', (w // 2, h // 2), u).resize(size).tobytes()

	set -e

	if [[ -f /.dockerenv ]]; then
	exit 1
	fi

	readonly SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"

	readonly DOCKER_IMAGE_TAG=$(docker build --quiet - <<EOF
	FROM debian:bullseye

	import argparse
	import wave

	def crop(src_filename, dst_filename, num_frames):
	with wave.open(src_filename, 'rb') as src:
	with wave.open(dst_filename, 'wb') as dst:
	dst.setnchannels(src.getnchannels())
	dst.setsampwidth(src.getsampwidth())
	dst.setframerate(src.getframerate())