A web application to be used as a benchmarking tool for IaaS systems.
It exposes RESTfull endpoints to perform both CPU-intensive and memory-intensive tasks.

Pwitter stores pweets, pieces of text that are not limited in length, and indexes them by the polarity of their sentiment.
Thus, pweets are stored together with their sentiment analysis.

Pwitter exposes a single endpoint: /pweets, that can be requested using method GET or POST.
[GET] /pweets is meant to be the memory-intensive task, given that at most 20 pweets (each of them could contain a big amount of text) are loaded in memory; while [POST] / pweets is meant to be the CPU-intensive task, because the new pweet has to be processed to obtain its sentiment polarity.

For CPU and Memory behavior while stressing those endpoints, go to http://affo.github.io/pwitter/.

Receives a new pweet and stores it together with its sentiment analysis.

Sample request:

{
    "user": "Alan Touring", # not compulsory
    "body": "Having a great time with 0s and 1s!"
}

Sample response:

201

{
    "body": "Having a great time with 0s and 1s!",
    "polarity": 1.0,
    "subjectivity": 0.75,
    "user": "Alan Touring"
}

Returns at most 20 pweets with polarity between lower_bound and upper_bound.
If lower_bound and/or upper_bound are/is not specified, they/it default/s, respectively, to -1.0 and 1.0.

Sample requests:

[GET] /pweets?gte=-0.5&lte=0.23`
[GET] /pweets?gte=0.42`
[GET] /pweets`

Sample response:

200

{
    "pweets": [
        {
          "body": "Having a great time with 0s and 1s!",
          "polarity": 1.0,
          "subjectivity": 0.75,
          "user": "Alan Touring"
        }
    ]
}

The installation of both server and client is performed using Docker Engine and Docker Compose.
Use the Makefile to run tasks:

• make clean: destroys web and database containers;
• make fullclean: as above, plus it destroys also the web image, in order to force Docker re-build it on next run;
• make WEB_OPTS="<opts>", or make, or make run: runs clean task and then launches the application (docker-compose up) in daemon mode. You can specify opts for the Gunicorn web server (except from -b, which is fixed). At the end of the processing the IP address of the web container is returned (use it with the client).
• make client: builds the client Docker image.

The client is dockerized. To build it for the first time, run make client, in this way, you will obtain a Docker image called pwitter.
From now on you can launch commands as docker run --rm pwitter -H <web_ip> <command> <opt>. web_ip is the one obtained after make call.

For a documentation on available commands, run docker run --rm pwitter --help.