#!/bin/bash
# This script converts Base64url-encoded Twitter filenames into human readable dates and appends them as prefix
# Written by yours truly
#
# Hot singles in your area at https://endchan.net/agatha2/
#
# Inconsistent reference materials:
# https://en.wikipedia.org/wiki/Snowflake_ID
# https://github.com/twitter-archive/snowflake/blob/snowflake-2010/src/main/scala/com/twitter/service/snowflake/IdWorker.scala
#
# Requirements: GNU/bash I guess
#
# Usage
# Run script by sourcing, asterisk picks up every file in current working directory:
# $ source /home/user/path/twitter-filename-to-timestamp.sh *
#
# Or select files, drag-and-drop into terminal:
# $ source /home/user/path/twitter-filename-to-timestamp.sh '/home/user/path1/E2lkq7dXwAARMsY.jpg' '/home/user/path2/E6B2fedXMAIm80Z.jpg'
# 
# Test file:
# https://twitter.com/Wikipedia/status/1707751173812269539
# F7MmpC0XcAEggXE.jpg
# Post timestamp: [2023-09-29 13:36 UTC]
#
# Create separate subdir to put final files in
mkdir "timestamped"
#
# Consume multiple file paths as array
for FILE_PATH in "${@}" ;do
	#
	# Get raw filename
	FILENAME="$(basename "$FILE_PATH")";
	# echo "Filename: $FILENAME"
	# > F7MmpC0XcAEggXE.jpg
	#
	# Strip file extension and dot
	SNOWFLAKE="${FILENAME%.*}";
	# echo "Snowflake ID: $SNOWFLAKE"
	# > F7MmpC0XcAEggXE
	#
	# Decode from Base64 to Base2/binary number string, '=' padding to avoid errors
	BIN_STRING="$(echo "$SNOWFLAKE=" | basenc -d --base64url | basenc -w0 --base2msbf)";
	# echo "Binary string: $BIN_STRING"
	# > 0001011110110011001001101010010000101101000101110111000000000001001000001000000101110001
	#
	# Extract first 42 characters/bits (7x6) 
	BIN_TSTAMP="${BIN_STRING:0:42}";
	# echo "Binary timestamp: $BIN_TSTAMP"
	# > 000101111011001100100110101001000010110100
	#
	# Convert binary (Base2 or Radix 2) number to decimal (Base10 or Radix 10) using bash arithmetic expansion
	DEC_TSTAMP="$(echo "obase=10; ibase=2; $BIN_TSTAMP" | bc)";
	# echo "Decimal timestamp: $DEC_TSTAMP"
	# > 407159607476
	#
	# Get post date by adding Twitter epoch [1288834974657] and removing milliseconds
	POST_DATE="$(date --utc -d @$((("$DEC_TSTAMP"+1288834974657)/1000)) +'[%Y-%m-%d %H:%M:%S UTC]')";
	# echo "$POST_DATE";
	# > [2023-09-29 13:36:22 UTC]
	#
	NEW_FILENAME="$POST_DATE $FILENAME";
	echo "$NEW_FILENAME";
	# > [2023-09-29 13:36:22 UTC] F7MmpC0XcAEggXE.jpg
	#
	# Copy new files into subdir
	cp "$FILE_PATH" "timestamped/$NEW_FILENAME";
done;