Question on plugins management

[亚历山大]

  Dear community members,

  I have looked at the Savant manuals and was under impression that
all the jar files stored in the plugins directory in the Savant
project (I use the latest release release-2.0.1) are the only ones
used for the plugins. It appears that this is not the case and the
JARS corresponding to the plugins are stored somewhere. Moreover,
plugin manager makes it difficult to refresh the plugins that I
recompile. The JARs corresponding to the plugins that appear in the
plugins directory apparently not the ones used by the system. So I
need to remove plugins first with the plugin manager, restart the
system (since I cant re-install the plugins under the same name
immediately). Then install updated plugins and only then it appears as
updated recompiled program.
  Please suggest what is the current mechanism of storing these
plugin jars and how do I recompile plugins quickly.

  Thanks,

  Alexander

[Eric J. M. Smith]

2012/4/9 亚历山大 <alex...@big.ac.cn> wrote:
>   Please suggest what is the current mechanism of storing these
> plugin jars and how do I recompile plugins quickly.

Plugins are stored in the .savant/plugins sub-directory
("savant\plugins" on Windows) of your home directory. If you're
developing and want to make sure Savant is always using the latest
build of your plugin, just set your project's JAR output to go into
that directory.

Just so you know, all the Plugin Manager dialog does is copy the JAR
file to that directory. If you set your project up to send the output
there, you can ignore the Plugin Manager altogether.

Hope that helps,

Eric Smith
Tarkvara Design Inc.

[亚历山大]

Dear Eric,

Thanks for the prompt response. I ended up editing the file called project.properties that is part of the plugin project. It appears that these plugins jars are deposited in C:\Users\Owner\savant\plugins on my computer, so that I set dist.dir=C:/Users/Owner/savant/plugins and also modified the build.dir from build.dir=../out to build.dir=../out.plugin.splicing to avoid interference with the main Savant compiled code (otherwise I would need to recompile Savant every time I recompile plugin). Now it works!

Sincerely,

Alexander

> -----Original Messages-----
> From: "Eric J. M. Smith" <er...@tarkvara.org>
> Sent Time: Tuesday, April 10, 2012
> To: savant-c...@googlegroups.com
> Cc:
> Subject: Re: [Savant Community Post] Question on plugins management

[亚历山大]

Dear Savant community,

I have question about naming colunms in the Track class. When I get the columns the following way columnsNames[0] is always null.

String[] columnsNames = t.getDataSource().getColumnNames();

But when I get values from TabixRichIntervalRecord, the values[0] is not null, i.e. column names are shifted by 1 compared to column index.

for (Record r : records) {
TabixRichIntervalRecord rr = (TabixRichIntervalRecord) r;
String[] values = rr.getValues();
}

Is it supposed to be this way, or should it be considered for repair in the following versions?

Thanks,

Alexander



> -----Original Messages-----
> From: "Eric J. M. Smith" <er...@tarkvara.org>
> Sent Time: Tuesday, April 10, 2012
> To: savant-c...@googlegroups.com
> Cc:
> Subject: Re: [Savant Community Post] Question on plugins management
>

[Eric J. M. Smith]

On 3 May 2012 02:41, 亚历山大 <alex...@big.ac.cn> wrote:
>   Dear Savant community,
>
>   I have question about naming colunms in the Track class. When I get the columns the following way columnsNames[0] is always null.
>
>   String[] columnsNames = t.getDataSource().getColumnNames();
>
>   But when I get values from TabixRichIntervalRecord, the values[0] is not null, i.e. column names are shifted by 1 compared to column index.
>
>   for (Record r : records) {
>      TabixRichIntervalRecord rr = (TabixRichIntervalRecord) r;
>      String[] values = rr.getValues();
>   }
>
>   Is it supposed to be this way, or should it be considered for repair in the following versions?

Depending on the file in question, this may be correct behaviour.

For instance, RefSeq genes downloaded from UCSC store a "bin" in
column 0. Savant ignores the UCSC "bin" column, so
getColumnNames()[0] will be null. However, there is still data in
that column, so getValues()[0] will contain some integer value. The
first real data will be in column 1, so getColumnNames()[1] ==
"Transcript name" and getValues()[1] will contain the transcript name.