Removes attachments and emoji from Mattermost bulk export file

remove-attachments.rb

#!/usr/bin/env ruby
require 'json'

## How to use
#
#  1. Generate a Mattermost bulk export file with attachments
#  2. Unzip the file and find the .jsonl file with the post contents and put its filename on the next line
filename = 'import.jsonl'
#  3. Run it and direct the output to a new file, e.g.
#
#	ruby remove-attachments.rb > import_no_attachments.jsonl

posts = []
emoji = []
other = []


File.readlines(filename).each do |line|
	data = JSON.parse(line)

	if data['type'] == 'post' or data['type'] == 'direct_post'
		$stderr.puts "Found #{data['type']}"
		type = data['type']
		if data[type].key?('attachments')
			$stderr.puts "#{type} has an attachments key "
			
			begin
				if !data[type]['attachments'].nil? and data[type]['attachments'].length > 0
					$stderr.puts "#{type} has #{data[type]['attachments'].length} attachment(s). Resetting"
					data[type]['attachments'] = []
				end	
			rescue Exception => e
				$stderr.puts e.to_s
				$stderr.puts data[type].to_json
			end
			
		end

		posts << data.to_json
	else
		other << line
	end
end

other.each do |line|
	puts line
end

posts.each do |post|
	puts post
end