This tutorial is part of a series on ‘How To Build Bot’s Using Steem-JS Open Source Library’ it aims to demonstrate practical applications for bots and demystify the code used to create them.

Previous Tutorials

• #1 Auto-liking bot
• #2 Curation bot
• #3 Auto follow bot
• #4 Paid upvote bot
• #5 @ Mention Notifier bot
• #6 Auto Reward Claimer bot
• #7 Bots In The Background

The completed bot code is here - I know some people find it easier to see everything in one block and follow along that way. ⬅️

Outline 📝

This tutorial is looking at creating a responder bot. One of the most popular responderbots is origionalworks another well known one is treeplanter there are likely many others.

When mentioned a responder bot will perform an action and respond with the results in a comment. In this tutorial we’ll create a responder bot that will calculate statistics about the account who requested it.

I have used my steemversary account for testing this bot.

Requirements

• A plain text editor (I use atom.io )
• Your private posting key - This is found in your wallet on the permissions tab on steemit.com, click ‘show private key’ (remember to keep this safe). This is what allows the bot to reply when mentioned.

Difficulty

Intermediate - Please run through all other tutorials in the series to better understand what we’re working on.

(I appreciate feedback on the teaching level).

Learning Goals

• Replying to @ mention with a broadcast.comment() using Steem-js
• Look at the getState() function of Steem-js
• Use Javascript Array Reduce function to create averages

Step 1 - Setup & Recap 💻

The starting point for this bot will be where tutorial 5 left off. We’ll strip out the email code and replace it with out statistics calculation and comment reply. This bot will once again run in the browser but there is no reason why it couldnt also run with node.js as in the previous tutorial.

starting point using code created in bot #5

Fill in your details above in the NAME & KEY variables. When you run the code above notice the console output notifying of checking and receiving mentions. Send yourself a mention to make sure it’s running correctly.

Step 2 - Get stats about Author 💻

We’re going to calculate three stats about an author, stats that do not exist by just querying the blockchain. Looking at up to the last 20 author posts we’ll calculate -

• Average Votes Per Post
• Average Replies Per Post
• Average Post Word count

this could be the start of an interesting responder bot, I could imagine it including many more stats.

If people find this idea interesting let me know and I’ll make it a real robot

These stats don’t exist on the blockchain. The getState() function allows us to retrive a number of posts from the blockchain these can be related to an author if we used getState(‘sambillingham’) or a category getState(‘trending’), it differers from getDiscussionsBy() in that it give data about the replies to post also and can be used to look into the comment section. First try that and look at the results

steem.api.getState(`/@sambillingham/`, (err, result) => {
    console.log(err, result)
})

If you click the dropdown you’ll notice we can view all of an accounts latest content. Each post also contains individual data, comments, replies, rewards etc. The first thing I noticed was that the data is not in the friendliest format. The content is an object, where each property of the object is another object of post data.

When working with data my first aim is to get it’s format close to what I want to work with. I this case An Array where each item in the array relates to stats about each post will be easier to work with than objects in objects. Loop over the object and reformat the data selecting only the items we’re interested in.

let resultsArray = [];
for ( post in result.content ){
    resultsArray.push({
        votes: result.content[post].net_votes,
        replies: result.content[post].children,
        length: result.content[post].body.split(' ').length
    })
}

1. Initialise an empty Array
2. Loop over result content object
3. for each post in content create a new object containing votes, replies and length
4. split the body content at spaces to get word count
5. add each item to the new array

Output of the new Array.

The data we’ve now constructed contains only what we’re interested in and will be far easier to manage.

Next we’ll create averages. The Array reduce function works perfectly for our needs here. Reduce takes a list of numbers in an array and gives a single number, this might be used for finding minimum, maximum, total etc. Total we’ll use it to find an average. When looking online you’ll often see .reduce((a,b) => a + b, 0) which looks confusing until you realise a represents the total and b represents the current item, there is no reason why you couldn’t write it out like below.

resultsArray.reduce((totalVotes,currentItem) => total + currentItem.votes, startingNumber) / resultsArray.length

We only use a/b or x/y variables names to save keystrokes. Reduce will take the the total so far add on the current item and move along the array. Divide by the length of array aka number of posts to get an average.

For all of our stats we would use reduce for each,

let statsData = {
    averageVotes: resultsArray.reduce((a,b) => a + b.votes, 0) / resultsArray.length,
    averageReplies: resultsArray.reduce((a,b) => a + b.replies, 0) / resultsArray.length,
    averageLength:  resultsArray.reduce((a,b) => a + b.length, 0) / resultsArray.length
}

Leaving us with a single object contains averages for votes, replies and word count.

Step 3 - Reply With A Comment 💻

steem.broadcast.comment() does what you’d expect sending comments out over the blockchain. Simply fill in the required fields in the correct order

steem.broadcast.comment(
            ACCOUNT_KEY, 
            AUTHOR,
            AUTHOR_PERMLINK,
            ACCOUNT_NAME,
            NEW_PERMLINK,
            TITLE,
            COMMENT_CONTENT,
            CUSTOM_JSON,
            function(err, result) {
              console.log(err, result);
            }
          );

ACCOUNT_KEY & ACCOUNT_NAME are the posting account, we added at the top of our code.

AUTHOR & AUTHOR_PERMLINK are who we’re replying to and will be part of the transaction data we retrieve when listening for the mention.

NEW_PERMLINK is a random alphanumeric string that we’ll need to generate to make sure our comment has a unique url

TITLE & COMMENT_CONTENT are fairly self explainitary. Note the Title is only viewable when accessing the comment directly not when looking at comment in a post.

CUSTOM_JSON is used for specifying tags and the app we’re posting from.

Create a sending function that has access to the transaction data and the statistics data we’ve just calculated. We’ll use this as part of those functions in just a moment.

 function reply(txData, statsData){}

Create a comment for our bot to send back to the user.

function reply(txData, statsData){
    let comment = `I'm the statistics responder bot, reporting as requested.<br><br>Average Votes Per Post: ${statsData.averageVotes} <br>Average Replies Per Post: ${statsData.averageReplies} <br>Average Post Word count: ${statsData.averageLength} <br>`
}

Fill in the sending data

  steem.broadcast.comment(
            ACCOUNT_KEY,
            txData.author,
            txData.permlink,
            ACCOUNT_NAME,
            randomString(),
            '',
            comment,
            { tags: ['stats'], app: 'statsbot' },
            function(err, result) {
              console.log(err, result);
            }
          );

We can create a randomString() function that generates a unique string used for our url.

function randomString() {
    let string = ''
    let allowedChars = "abcdefghijklmnopqrstuvwxyz0123456789";
    for (var i = 0; i < 32; i++){
        string += allowedChars.charAt(Math.floor(Math.random() * allowedChars.length));
    }
    return string + '-post';
}

1. initialise an empty string
2. create a set of characters that are permlink valid within the steem blockchain
3. create a loop that will be 32 character long.
4. for each step in the loop add choose a random character for the available list and add it to our new string
5. after we have 32 characters in the string return it as the result

Step 4 - bring it all together

With the statistics generated and the reply function setup we need to combine this with the transaction stream listening for mentions.

Start the Stream. When a mention is found pass the transaction data through to the statistics function.

steem.api.streamTransactions('head', function(err, result) {
    let txType = result.operations[0][0]
    let txData = result.operations[0][1]
    let includesMention = checkContentForMention(txType,txData)
    if (includesMention) {
        console.log('MENTION FOUND: ', txData)
        getStats(txData)
    }
});

Generate statistics and pass transaction data and generated statics through to the reply function.

function getStats(txData){
  steem.api.getState(`/@${txData.author}/`, (err, result) => {
    let resultsArray = [];
    for ( post in result.content ){
      resultsArray.push({
        votes: result.content[post].net_votes,
        replies: result.content[post].children,
        length: result.content[post].body.split(' ').length
      })
    }
    let statsData = {
      averageVotes: resultsArray.reduce((a,b) => a + b.votes, 0) / resultsArray.length,
      averageReplies: resultsArray.reduce((a,b) => a + b.replies, 0) / resultsArray.length,
      averageLength:  resultsArray.reduce((a,b) => a + b.length, 0) / resultsArray.length
    }
    reply(txData, statsData)
  })
}

Veiew the full code below if putting it all together is a little confusing in written form

Here’s the full code 🤖

The Bot is setup. When the bot sees a mention with the account name it will calculate the stats and send a reply. There are a lot of opportunities with these kind of bots. Be careful to make sure your bot adds value, nobody likes spammy bots in the comments!

While outside of the scope of this tutorial it might be interesting to have this sort of bot also upvote if you statistics increase since the previous day or week.

Let me know what you think of this robot tutorial.

Other Posts in This series

• Part 1 - Beginner friendly - Build your first Steem bot in Javascript - 30minutes
• Part 2 - Beginner friendly - Building bots With steem-js #2
• PArt 3 - Copy Follow Bot - Building Bots With steem-js #3
• Part 4 - Paid Voting Bot - Building Bots With steem-js #4
• Part 5 - Mention Notifier Bot - Building Bots With steem-js #5
• Part 6 - 🤖 TUTORIAL - Auto Reward Claimer - Building bots With steem-js #6
• 🤖 TUTORIAL - Bots In The Background - Building bots With steem-js #7

---

Posted on Utopian.io - Rewarding Open Source Contributors

---