reactome / curatortool Goto Github PK

Clicking Report Bugs in the tool opens the browser with the address http://brie8.cshl.edu/bugzilla which is unreachable. Please update.

The two classes have to do with Stable identifier history

To do this we will need to create an ant file to compile the command line runner and then add the command line runner commands to the Jenkinsfile for sliceQA

During Reactome release 66 we needed to rebuild the WebELVTool jar to account for the introduction of the new renderable classes. When we ran it initially, we got a 'NoClassDefFoundError' for /org/gk/gkCurator/authortool/ directories.

The fix we made to the WebELVDiagram.xml was to add the following two lines:

<include name="org/gk/gkCurator/**" />
<include name="org/gk/gkCurator/authorTool/**" />

The Help-->Curator Guide and Report Bugs items open a file choice dialog, asking for a browser to use. This is unusual. As I understand, pages can be opened from any application using a default browser.

Typing a browser name in the mentioned dialog gives an error message. Trying a different input is not possible because choosing the menu item now immediately gives the error. So, the first input changed a state somewhere which needs to be reset before trying the item again. Looking in Tools-->Options has nothing about browser configuration. Finally I find it in $(HOME)/.reactome/browser.prop .

So I think the tool should either use the default browser or, at least, make it possible to change a wrong browser setting in Options.

As someone who has to spend hours staring at the Reactome Curator Tool, it would be really nice if there was a "dark mode" that isn't so hard on my eyes.

I have no idea how hard this would be to implement, but it would be very nice.

Clicking Help-->Curator Guide brings me to https://reactome.org/documentation which does not show any Curator Guide. I think there would be no harm in directly opening https://devwiki.reactome.org/index.php/New_Curators_Guide from Help-->Curator Guide.

Hello Guanming,
supposedly the curator tool uses pubmed or crossref to import PMID entities. Doing the same for an URL that starts with http://doi.org/ could use api.crossref.org to import title and authors of a DOI URL. At the moment I'm adding authors by hand, and I need an hour for a paper that has no PMID. Can you please make my work a bit easier?

JNLP is a dependency of reactome-base: https://github.com/reactome/CuratorTool/blob/master/reactome_base_sonatype/pom.xml#L93-L97

Unfortunately, this breaks projects that depend on reactome-base, as jnlp is not on maven central. It is available here: https://github.com/reactome/CuratorTool/blob/master/lib/jnlp-1.2.jar, meaning that anyone who wants to use a project that depends on reactome-base must install this jar into their local m2 repo before using reactome-base.

The current recommendation for using JNLP in a project is to include it as a system dependency:

    <dependency>
      <groupId>javax.jnlp</groupId>
      <artifactId>jnlp-api</artifactId>
      <version>7.0</version>
      <scope>system</scope>
      <systemPath>${java.home}/lib/javaws.jar</systemPath>
    </dependency>

I don't know if this would work for the CuratorTool or not, but I'm not really sure that JNLP should even be included in reactome-base as it is a very niche dependency, and I suspect that most downstream projects that rely on reactome-base will not be making use of Java web-launch.

If dropping jnlp is not an option, I'd suggest maybe marking it as "optional" (https://maven.apache.org/guides/introduction/introduction-to-optional-and-excludes-dependencies.html) and then test that release-common-lib (or some other project that depends on reactome-base) can be built cleanly (clean out local .m2 and then run mvn clean compile).

As it currently is, it is difficult to build anything the relies on reactome-base without first manually installing jnlp-1.2.

This is an error in production where the Update Tracker is throwing an error when trying to put data in the gk_central