Hackage :: [Package]

eros-http: JSON HTTP interface to Eros.

[ bsd3, program, text ] [ Propose Tags ]

This is a small program that runs an HTTP server.

Changelog

0.6.0.1
Remove unneeded dependency on wai-responsible.
0.6
Update to be compatible with Eros v.0.6
[Skip to Readme]
Versions [faq] 0.5.0.0, 0.6.0.0, 0.6.0.1
Dependencies aeson (==0.8.*), base (==4.7.*), blaze-html (==0.7.*), bytestring (==0.10.*), eros (==0.6.*), http-types (==0.8.*), markdown (==0.1.*), text (==1.1.*), wai (==3.0.*), warp (==3.0.*) [details]
License BSD-3-Clause
Copyright 2014, Peter Harpending
Author Peter Harpending
Maintainer Peter Harpending <pharpend2@gmail.com>
Category Text
Uploaded by pharpend at Sun Jul 20 10:00:10 UTC 2014
Distributions NixOS:0.6.0.1
Executables eros-http
Downloads 1245 total (44 in the last 30 days)
Rating (no votes yet) [estimated by rule of succession]
Your Rating
  • λ
  • λ
  • λ
Status Hackage Matrix CI
Docs not available [build log]
Last success reported on 2015-05-19 [all 6 reports]

Downloads

Maintainer's Corner

For package maintainers and hackage trustees

Readme for eros-http-0.6.0.1

[back to package description]

eros-http

This is an HTTP front-end to the Eros library. Eros is a text censorship library, that I wrote.

Usage

If the server receives a GET request, it returns an HTML representation of this file.

It takes an input string via POST, and returns some data in JSON mapping each phraselist to the score for the input string. The JSON is compressed.

I fed the server the GPL, here are the results, which have been prettified, and alphabetized by key.

{
  "chat": 0,
  "conspiracy": 0,
  "drug-advocacy": 0,
  "forums": 0,
  "gambling": 0,
  "games": 0,
  "gore": 0,
  "id-theft": 0,
  "illegal-drugs": 0,
  "intolerance": 0,
  "legal-drugs": 0,
  "malware": 0,
  "music": 0,
  "news": 0,
  "nudism": 0,
  "peer2peer": 0,
  "personals": 0,
  "pornography": 20,
  "proxies": 0,
  "secret-societies": 0,
  "self-labeling": 0,
  "sport": 30,
  "translation": 0,
  "upstream-filter": 0,
  "violence": 0,
  "warez-hacking": 0,
  "weapons": 0,
  "webmail": 0
}

The home page of PornHub yields this result:

{
  "chat": 0,
  "conspiracy": 0,
  "drug-advocacy": 0,
  "forums": 0,
  "gambling": 0,
  "games": 0,
  "gore": 0,
  "id-theft": 0,
  "illegal-drugs": 0,
  "intolerance": 0,
  "legal-drugs": 0,
  "malware": 0,
  "music": 0,
  "news": 0,
  "nudism": 0,
  "peer2peer": 0,
  "personals": 0,
  "pornography": 8790,
  "proxies": 0,
  "secret-societies": 0,
  "self-labeling": 0,
  "sport": 0,
  "translation": 0,
  "upstream-filter": 0,
  "violence": 0,
  "warez-hacking": 0,
  "weapons": 0,
  "webmail": 0
}

This is the command I used

curl www.pornhub.com | curl -d @- localhost:8000 | prettify.rb

This uses my JSON prettifier script.

Overflow bug

Edit - This bug was fixed in version 0.6.

~~There's a fundamental flaw in the algorithm where if it gets a lot of data with~~ ~~a lot of flagged phrases, it takes a long time to calculate the result. The~~ ~~server times out after 30 seconds. I haven't figured out how to get~~ ~~multiprocessing, so this will remain a bug until I do.~~

~~This won't be a problem for typical usage, only if you try to send paragraphs of~~ ~~dirty data in one string.~~

~~Typical usage is for SMS messages, which are at most 160 characters long.~~