tsql

A stupid stupid twitter client that uses sql commands to get tweets. Did I mention that this is totally stupid?

const config = {
  "twitter" : {
    "consumer_key": "",
    "consumer_secret": "",
    "access_token": "",
    "access_token_secret": ""
  }
}

const t = require('./twsql.js');
const tsql = new t(config);

tsql.query("select id, text from home").then(function(data){
  console.log(data);
});

or from the command line

chmod 755 console.js
./console.js select id, text, created_at from home where id between 1 and 900000000000000000 limit 10

What columns can I select

Any property of the tweet. Can link into sub-objects as if they were joined tables. e.g.

select user.name from home

Child objects of the Tweet object are implicitly 'joined'.

Can use the as keyword to alias the columns.

What 'tables'

home is a protected keyword, just like on twitter.com, and uses the authenticated user's home timeline. Anything else is treated as a screen name and searches that user's tweets.

What about joins?

Can join media and hashtags. Fields on those tables are accessed as if they are joined tables. e.g. select media.media_url from home join media or select hashtags.text from home join hashtags

The timeline being queried will always be assumed to the first table in the from list.

What filters can I use

id can be filtered using <, >, and between. include_retweets is a boolean to filter retweets from the results. exclude_replies is a boolean to filter replies from the results.

LIMIT

Uses the MySql syntax for limiting

LIMIT 2 Returns 2 rows.

LIMIT 1,2 Returns 2 rows, offset by 1

TODO

• Further parsing of WHERE clauses into timeline filters
• Better Table joining. Should be able to handle LEFT vs INNER joins
• ORDER BY
• The rest of the entities object, co-ordinates objects, and other sub-objects that are arrays.