rethread-tweets.sh

#!/bin/bash

curlicue_path="${HOME}/src/curlicue"
  # edit this to point wherever you `git pull https://github.com/decklin/curlicue.git` to...

function print_usage {
	cat <<-_END_OF_HELP_
	rethread-tweets
	
	USAGE
	  rethread-tweets [tweet_id] [other_tweet_id(s) ...]
	
	Given the tweet ID of the last tweet in a thread, recursively retrieve up to the head
	of the thread, then print out the text of the tweets in sequence (lowest-to-highest
	tweet ID, roughly equivalent to chronological order, in theory)
	
	If you pass in more than one tweet ID, the threads will be reconstructed one by one.
	
	DEPENDENCIES
	  - openssl
	  - curl (with SSL support for HTTPS retrieval)
	  - jq
	    https://stedolan.github.io/jq/
	  - curlicue:
	    https://github.com/decklin/curlicue
	  - a Twitter API access token to pass to curlicue
	    https://dev.twitter.com/oauth/overview/application-owner-access-tokens
	
	_END_OF_HELP_
	return 0
}

if ((0 == $#)) ; then
	print_usage
	exit 0
fi

while (($# > 0)) ; do
#	twid='816631120196501504'
	twid=$1
	ids=( "$twid" )
	while [[ $twid == +([0-9]) ]] ; do
		twid_next=$("${curlicue_path}"/contrib/twitpull -n -j '.in_reply_to_status_id_str' 'statuses/show' "id=${twid}")
		  # I use `in_reply_to_status_id_str` here rather than `in_reply_to_status` because
		  # my version of `jq` apparently lacks bignum arithmetic and truncates the numeric
		  # values to 53 bits of precision which IS NOT THE SAME THING AT ALL kthxbyeeeeee
		ids+=("$twid_next")
		twid=$twid_next
		printf . >> /dev/stderr
	done
	unset ids[$((${#ids[@]} -1))]
	ids_str=${ids[*]}
	ids_str=${ids_str// /,}
	"${curlicue_path}"/contrib/twitpull -n -j 'sort_by(.id_str) | .[].full_text' 'statuses/lookup' "id=${ids_str}" 'tweet_mode=extended'
	printf '\n'

	shift
done

This pairs well with a little dingus I whipped up to run twitter images through tesseract; I've got it as a little scriptlent in ~/bin/ named tw_ocr:

#!/bin/bash

COLUMNS=$(tput cols)

while (($# > 0)) ; do
	url_tw=$1
	url_img=($( \
		curl -L -s "$url_tw" | \
		  awk '/property="og:image"/ {
			i=$0;
			sub(/.*content="/, "", i);
			sub(/".*/,         "", i);
			print i;
		}' \
	))
	for ((i=0; i<${#url_img[@]}; i++)) ; do
		fn="$(mktemp -p /dev/shm/ --suffix=.jpeg)" && \
		  curl -Ls -o "$fn" "${url_img[i]}" && \
		  tesseract "$fn" stdout -l eng && \
		  rm -- "$fn"
		(((${#url_img[@]}-i)>1)) && printf '\n\n----\n\n'
	done | fmt -w "$COLUMNS" | less
	shift
done

Inspired by https://twitter.com/gnomon/status/817226973340725248 and upthread.