This project is used to teach the basics of core Akka.
We have discussed
Create a database actor called DbActor that stores values of type Any in a Map, indexed by a key of type String.
The actor should be created in the actor.db package.
There is a simply REST service for the application now. The goal of this chapter is to add some endpoints to interact
with the DbActor.
Create a REST endpoint that will return the content stored in the DbActor for a given key.
GET http://localhost:9000/db?key=[key]
If the guid (key) is not in the AkkaDB, return a 404 error, Not Found.
Create a REST endpoint that will store a value in the DbActor. Specify the key and value to be stored in key and
value query parameters
POST http://localhost:9000/db?key=foo&value=cheese
Store a piece of content
curl -i -X POST "http://localhost:9000/db?key=foo&value=cheese"
Retrieve a piece of content
curl -i http://localhost:9000/db?key=foo
The goal of this chapter is to have your application retrieve an RSS feed and store the contents of each article in the feed in the DbActor.
Create a REST endpoint that accepts a URL of content that should be downloaded via an RSS feed. The application will
retrieve the RSS feed from the URL provided, parse out the articles (items) and store the content of those articles in
the DbActor.
Add only ONE actor to the system for this implementation. Call the actor RssActor.
When storing content in the AkkaDb, you should use the <guid> element as the key for each <item> in the RSS feed. The value
stored in the AkkaDb could be the <description> element of the item or you may opt to download the content from the
item's URL found in the <link> of the item. The description element is often wrapped in a CDATA which you’ll need to discard.
If you grab the content from the URL in the <link> element, you can use the BoilerPipe library included in the Learning
Akka project to retrieve the content stripped of extraneous HTML elements/tags from it’s specific URL. For example:
de.l3s.boilerpipe.extractors.ArticleExtractor.INSTANCE.getText(new java.net.URL( \ "http://www.cbc.ca/news/canada/ottawa/ottawa-weather-forecast-october-5-1.3791908?cmp=rss"))
The endpoint should be
POST http://localhost:9000/contents/url
The request body should be JSON indicating the URL to fetch the RSS feed from.
{
"url", "[RSS feed url]"
}
If content cannot be downloaded from the URL ,return a 404 error, Not Found.
If the URL has content but it’s not in the format you expect, return a 400 error, Bad Request. If the supplied URL is invalid, return a 400 error, Bad Request.
Create a REST endpoint that will return the content for a given guid.
GET http://localhost:9000/content/guid/[guid]
If the guid (key) is not in the AkkaDB, return a 404 error, Not Found.
Create a REST endpoint that will return the guids for all content in the system. You do not need to paginate this data.
GET http://localhost:9000/contents/guids
Parse an RSS feed
curl -i -H "Content-Type: application/json" -X POST -d '{"url":"http://www.cbc.ca/cmlink/rss-world"}' http://localhost:9000/contents/url
List the guid keys in the DB
curl -i http://localhost:9000/contents/guids
Retrieve a piece of content
curl -i http://localhost:9000/contents/guid/[a guid from the content in the DB]
If you are grabbing articles and cleaning them up using boilerpipe, you likely needed to increase your ask Timeout
in Main to 10 seconds or so. This is because the RssActor is downloading articles from a URL and stripping out tags
in a serial fashion: one article at a time.
Start by refactoring the code to download an article, strip it of tags and store it in the DbActor to a separtate
actor. Then, change that actor into a Router (with a pool of actors created by it). What happens to the RSS fetching
time with 2 actors in the pool? How about with 5 actors in the pool?