How to Post a Random Tweet using Twitter API
The Twitter API enables developers programmatically posting tweets.
We will use Spring Boot to build a Client App that posts tweets periodically.
@Scheduled is a Spring annotation that marks a method to be scheduled.
To execute @Scheduled annotations, we add an @EnableScheduling annotation on the main class.
A cron job is a Linux job scheduler that sets up tasks periodically that run at a fixed date or interval.
A cron attribute - as in Unix-based systems - enables the @Scheduled method to be executed at a specified date/time.
1
@Scheduled(cron = "0 0 0/3 1/1 * ?")
The requirement
We want to delegate our manual tweets to an automated process, even when we are sleeping, to support our social media automation.
We have created a Tweet scheduling client to publish tweets under a Scheduler’s control.
We have a text file listing all the articles we have written.
1
2
3
4
5
6
SOLID principles;article_URL;#solid #objectoriented
Binary Search Tree;article_URL;#algorithm #datastructure
Clean code;article_URL;#programming #codingpractices
.
.
.
We have the following metadata for every article.
1
2
3
4
5
public class Article {
String title;
String link;
String hashTags;
}
Load all the articles sequentially into a hashmap. We define the hashmap key as an integer to uniquely identify every article.
1
2
3
4
5
6
7
8
9
10
private Map<Integer, Article> articlesMap;
.
.
.
int n = 1;
for (String line: allLines) {
Article article = new Article();
//code omitted for brevity
articlesMap.put(n++, article);
}
We choose a random number with the following method.
1
2
3
private int getRandomNumber(int min, int max) {
return ThreadLocalRandom.current().nextInt(min, max + 1);
}
The following code snippet shows a method executed every three hours to retrieve a random article and send it to the Twitter API as a scheduled tweet.
1
2
3
4
5
6
7
8
9
@Scheduled(cron = "0 0 0/3 1/1 * ?")
public void postInTwitter() throws Exception {
int n = getRandomNumber(1, articlesMap.size());
Article article = articlesMap.get(n);
twitterService.sendPost(article);
logger.info("postInTwitter random article: " + article.getTitle());
}
But once an article is retrieved, the following article will probably be the same if we use the previous getRandomNumber method.
Don’t post duplicate articles in a time frame.
What we want is to post no duplicate articles in a time frame.
For example, if we post every three hours, eight posts are in one day. Well, we want to show eight no-duplicate tweets for our users throughout the day.
As developers, we need to understand the inner workings of data structures to support the behavior - method - we want to implement. Could an array, linked list, queue, or stack satisfy our requirement?
If we can publish only eight articles daily, then our data structure should allow us to store only eight elements. Once we have filled our data structure, there will be no space for the following article. Then, we must be able to remove the oldest article (in this case, the first one) and insert the newest article (in the last position). The Queue data structure supports our requirement.
A Queue is an abstract data type, which includes a collection of objects that follow the first-in, first-out (FIFO) principle, i.e., the element inserted at first is the first element to come out of the list.
A MAX_ARTICLES variable defines your frame time. You can increase or decrease it according to the number of your articles.
We define a queue of integers that is implemented by a linked list.
1
2
private int MAX_ARTICLES = 8;
private Queue<Integer> queue = new LinkedList<>();
Every three hours, we check if the new random number is not included in the queue. If the validation is successful, we retrieve a new article and publish it.
The remove method deletes the oldest element from the queue (remember the element inserted at first is the first element to come out of the list).
The new method executes (do) the following instructions while a new random number is not found.
The new random number represents an article that is included in our articlesMap variable.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
private Article getRandomArticle() {
Article article = null;
boolean newNumberFound = false;
do {
int newNumber = getRandomNumber(1, articlesMap.size());
if (!queue.contains(newNumber)) {
newNumberFound = true;
article = articlesMap.get(newNumber);
if (queue.size() < MAX_ARTICLES) {
queue.add(newNumber);
} else {
queue.remove();
queue.add(newNumber);
}
}
} while (newNumberFound == false);
return article;
}
The Big O notation to add and remove elements is O(1).
This is how it looks at the new postInTwitter method. To evaluate the random behavior, we iterate the queue - only for test purposes.
1
2
3
4
5
6
7
8
9
10
@Scheduled(cron = "0 0 0/3 1/1 * ?")
public void postInTwitter() throws Exception {
Article article = getRandomArticle();
twitterService.sendPost(article);
logger.info("postInTwitter random article: " + article.getTitle());
for (Integer item: queue) {
System.out.print(item + " ");
}
System.out.println();
}
Once you deploy the Tweet scheduling to any cloud provider, its log files show the queue’s random numbers, for example.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
42
42 70
42 70 44
42 70 44 1
42 70 44 1 61
42 70 44 1 61 26
42 70 44 1 61 26 65
42 70 44 1 61 26 65 28
70 44 1 61 26 65 28 52
44 1 61 26 65 28 52 31
1 61 26 65 28 52 31 33
61 26 65 28 52 31 33 6
26 65 28 52 31 33 6 44
65 28 52 31 33 6 44 24
With the help of this Twitter bot, you can now focus on writing more articles. Look at it in action on my Twitter account!.
When finding a random number is exponential
When you have written more than 100 articles, finding a new random number that is not still included in the queue variable could be difficult.
When the queue has space for only three more elements, the following validation could be exponential when trying to find a new random number between 100 possible articles.
1
2
int newNumber = getRandomNumber(1, articlesMap.size());
if (!queue.contains(newNumber)) {
Next few days we will how to redesign our algorithm to avoid this exponential issue.
Please donate to maintain and improve this website if you find this content valuable.
Similar problems you can find in my book about algorithms and the inner workings of Data Structures. Learn how to apply common algorithms to practical problems.
