| Category | Third party tool integration |
|---|---|
Purpose | Integrates the third party tool MT-ComparEval. |
| Third party URL | https://github.com/choko/MT-ComparEval |
Bootstrap Class | editor_Plugins_MtComparEval_Bootstrap |
| Type | Core plug-in (delivered with translate5 core) |
Description
Integrates the third party tool MT-ComparEval into translate5.
MT-ComparEval is a tool for comparison and evaluation of machine translations. It allows users to compare translations according to several criteria, such as:
- automatic metrics of machine translation quality computed either on whole documents or single sentences
- quality comparison of single sentence translation by highlighting confirmed, improving and worsening n-grams
- summaries of the most improving and worsening n-grams for the whole document.
MT-ComparEval also plots a chart with absolute differences of metrics computed on single sentences and a chart with values obtained from paired bootstrap resampling.
This description is from https://github.com/choko/MT-ComparEval.
Installation
On the MT-ComparEval homepage you find also the installation instructions for MT-ComparEval, currently only available for Linux.
In addition to the installation instructions there, you can change the default server configuration of MT-ComparEval from localhost:8080 to whatever you want by changing this in the bin/server.sh directly.
The server address http://localhost:8080 is the default configured URL for MT-ComparEval in translate5, this can be changed in translate5 config, see below.
Configuration
| Config name | Values | Default | Description |
|---|---|---|---|
| runtimeOptions.plugins.MtComparEval.url | URL | http://localhost:8080 | Base URL to the MT-ComparEval Tool, for format see default value |
Debugging
Adding the following line to your installation.ini enables debugging output for MtComparEval Plugin.
runtimeOptions.debug.plugin.MtComparEval = 1
Enabled debugging with the value 1 enables more verbose output of the communication between translate5 and MT-ComparEval.