Cleaning up disqus comments

[hamoid]

I noticed the comments imported from disqus into isso contain a lot of html, so I plan to clean them up.

I created this small script to edit one comment:

#!/bin/bash

id=$1
db="my-isso-file.db"
table="comments"
tempfile=".comment-edit"

if (( $# != 1 )); then
    echo "Use edit-comment ID"
else
    echo "edit $id"
    sqlite3 $db "SELECT text FROM $table WHERE id=$id" > $tempfile
    vim $tempfile
    sqlite3 $db "UPDATE $table SET text=readfile('$tempfile') WHERE id=$id"
fi

It needs a recent version of sqlite3, otherwise readfile() is not available.

It can be used in combination with searches, for instance sqlite3 my-isso-file.db "SELECT id,text FROM comments WHERE text LIKE '%problem%'", to find some kind of problematic text, then you type ./edit-comment ID using the returned ID to edit the comment in vim.

Probably before going into editing single comments I'll try using pandoc to convert all HTML in the comments to Markdown. Usage: cat .comment-edit | pandoc --from=html --to=markdown --no-wrap.