The platform manager can check statistics about the number of users and posts. They can also review all news and articles.
A news aggregator platform that delivers content from 100 news sources
Our client wanted to create a news aggregator platform called Khabar on which everyone could find something interesting. Our task was to develop iOS and Android applications along with a landing page.
To ensure that Khabar always has enough new articles and blog posts, we integrated around 100 news sources and let users subscribe to the ones they nally find useful and entertaining. The purpose of Khabar is to be an all-in-one news platform that lets users quickly learn what’s happening in the spheres they care about without having to go to dozens and dozens of separate sources.
It’s been estimated that by 2019 there will be 789.2 million mobile phone users in the Africa and Middle East region. And it’s only natural that once people have mobile devices, they want to have everything the internet offers at the tip of their finger.
Apps like Khabar have a lot to offer. Not only do they let people read news and blog posts whenever they want, but they also let people choose their favorite categories so they only see content they really care about. Push notifications alert users when anything notable has happened, making sure that not a single piece of news gets past them.
We've implemented a fully automated scraping service that delivers content from more than 100 sources in real time. Also, we built a system that automatically defines the scapers that don't work and sends a notification to the support team about this issue.
Readers can create profiles, subscribe to interesting news categories, and get notifications with the latest news. Readers can add favorite news items and save them as well as share news on social networks.
We started developing Khabar from mood boards and wireframes.
Our client put a lot of emphasis on minimalism and wanted to avoid any unnecessary details. To this end, we selected light hues that don’t distract readers from articles. The CTA buttons (the only highlighted parts of the page) make sure that users can easily share their favorite pieces. Because the app needed to support both English and Arabic we couldn’t really offer much font-wise, as the readability of Arabic is extremely font-sensitive. To prevent the design from being bland, we added some accents. We agreed that the app would be a white canvas with little touches of color.
As for the logo, we decided to tie it to Middle East-styled ornaments and made it in the form of a decorated Arabic letter “K.”
The backend consists of two major modules: a parsing module (which collects news from various sources) and a REST API module (which transfers those news to the applications).
The parser module is based on Scrapy. Every five minutes, it runs a command to parse each separate news resource. All of the data from these sources is listened to and collected by one of our background processes, after which it’s stored in a Redis-based database. From here, another background process takes every new piece of news and registers it in the PostgreSQL database.
In the other words, the parser parses and registers raw data to Redis. Then, while the news is being imported, that data is converted to the necessary format and saved to PostgreSQL. We used two databases to separate parsing and importing, which increases the speed of both of these processes.
Khabar’s full-text Amazon ElasticSearch system operates with the help of Django Haystack.
The second module is an API implemented with Django-REST-Auth. When a new article appears in one of the categories to which a user is subscribed, the user receives a push notification telling them about it. This API makes that happen.
We chose the MVP architectural pattern for the Android application. The MVP architecture makes the project’s structure very clean and flexible, helping a great deal with expandability. It’s very simple to make corrections and fixes and to add new features, which is absolutely vital for a product that’s just entering the market. Also, a clean architecture means a higher quality product with fewer bugs.
For connectivity with the REST API, we decided to go with the popular and stable combination of RxAndroid and Retrofit. Сalligraphy helps out with font management. We integrated Crashlytics to track crash analytics, which helps us prevent future bugs and crashes. Google Support libraries make sure that the app works fine on older versions of Android. We used Picasso for the beautiful images added to every article, and integrated GCM for push notifications.
The Khabar iOS application is built on the MVC architectural pattern, which offers several benefits: it’s a very simple system with a unifying vision and a simplified mechanism for debugging. By simple, we mean that it’s easy and comfortable to manage, which is perfect for any mobile app from the perspective of both clients and developers.
For efficient REST API interactions, we used Alamofire and ObjectMapper. SDWebImage was implemented to load and cache images, while TYBlurImage helps with blur effects. For social media logins, we used the Google/SignIn and TwitterKit libraries, and we supported push notifications with the help of APNs.
Khabar around 100 news sources, pulling articles from a bunch of different fields and making sure that users always have something interesting to read.
Khabar can notify users whenever something happens in a category they follow, such as culture, business, and so on. Users don’t even need to sign up in Khabar to get notifications; they just need to allow notifications in device settings.
Khabar allows users to sign up via Facebook, Twitter, and Google+. This also ensures that users can share articles with their friends through at least one of these channels.