danielnaber/oxygen-languagetool-plugin

LanguageTool plugin for the Oxygen XML editor

oxygen-languagetool-plugin's Issues

Timeout with large files

A test file contains approximately 4000 lines, each line of which is 'This is a example input to to show you how LanguageTool works.'

LT gives a 'timed out' message:

The number of lines at which the error occurred was not constant. But with approximately 3000 lines I did not get the 'timed out' message.

Plugin doesn't resolve keyref/conref in DITA to properly analyze sentence.

The plugin fails to resolve keyref/conkeyref/conref properly, which results in false positives. For example:

<p>All decisions in connection with the implementation of this application are at the sole
        decision of the Employer utilizing this application. Employers agree that use of the
        application is not a representation by <ph keyref="company-no-reg-tm"/> regarding the
        application’s compliance with any law or regulation and any suggested language provided out
        of the box with the application does not constitute legal advice by <ph
          keyref="company-no-reg-tm"/>. </p>

and:

<p><ph keyref="company-no-reg-tm"/> shall not be responsible for any
        implementation or configuration costs associated with use of the application unless
        separately purchased.</p>

Errors/suggestions tooltip is empty

After installing this plugin in Oxygen Author 25, i had to download and start the language tool server by following the instructions on this page https://dev.languagetool.org/http-server

Now, the tool can detect the grammar issues but on hover the message about what exactly is the error/suggestion is not displayed.

Error handling in plugin

We should look over what possible errors the server might give and give some feedback to the user

Authentication support for LanguageTool Premium

More of a feature request, but is there any chance of getting support for authenticating LanguageTool Premium users so that we can take advantage of other features and larger API queries?

Possibly better segmenation for checked text

In an XML document some elements may be considered inline (like bold, italic, underline). Ideally when checking a paragraph's text (paragraph containing inline elements) the server should receive the entire text from the paragraph as if those inline elements would not be present.

We could use this API method:

ro.sync.ecss.extensions.api.AuthorDocumentController.getFilteredContent(int, int, AuthorNodesFilter)

which should do this (it was developed for another external spell checker integration) instead of creating those ranges. But this needs to be tested though.

Mac: plugin works only when oXygen started from command-line

User report: the plugin installs but doesn't do anything when oXygen is started the usual way. When started from the command line with oxygenMac.sh, it works. Using LT local server or internet server doesn't make a difference.

OS X Yosemite 10.10.2
Java 7, update 75

Plugin Preferences -> Choose spell check language

In the plugin's preferences page we should have a combo box in which users would select the language used for spell checking.
We could query the server using:

https://languagetool.org:8081/Languages

but also have a list of default values if the server cannot be contacted.

'Use above settings for server' is ignored for an external file

In LanguageTool Options, when 'Use above settings for server' is selected, the LanguageTool Check in oXygen does not use the settings that are specified for rules that are in an external file (http://wiki.languagetool.org/tips-and-tricks#toc10).

I noticed this problem using LT 3.5 and LT 36 snapshot, and oXygen 18.1 (I haven't used oXygen and LT for many months, so I don't know when the problem was introduced).

Plugin Preferences -> Choose URL for connecting to the server

In the plugin preferences page we could also allow the user to configure the server URL to connect to, by default it would be: http://localhost:8081