[jira] [Commented] (AIRAVATA-3007) [GSoC] General plotting capabilities of experiment outputs

["Aravind Parappil (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/AIRAVATA-3007?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16810520#comment-16810520 ] 

Aravind Parappil commented on AIRAVATA-3007:
--------------------------------------------

[~marcuschristie] – Very interesting. You mentioned a declarative configuration approach for the admin users to specify graphing options per output file. Are you suggesting that the admin users specify them in, say, a JSON with pre-defined fields and then pass it along to the Django code? Something like:
{code:java}
{
   "xAxisTitle": "Time (ms)",
   "yAxisTitle": "Magnitude",
    "columnsReq": ["col2","col1","col8"],
   "chartType": "scatter",
    ......
}
{code}
Is it viable for us to create a UI form that takes this data while an admin user registers a new application? (Instead of asking the admin to provide a JSON, perhaps keep drop-downs, check-boxes etc. that will gather the same data?) Or is the application registration interface not within our control? 

There is a high possibility that I am completely off target here, so please feel free to ignore!

> [GSoC] General plotting capabilities of experiment outputs
> ----------------------------------------------------------
>
>                 Key: AIRAVATA-3007
>                 URL: https://issues.apache.org/jira/browse/AIRAVATA-3007
>             Project: Airavata
>          Issue Type: New Feature
>          Components: Django Portal
>            Reporter: Marcus Christie
>            Priority: Major
>              Labels: gsoc2019
>
> Add general plotting capabilities of computational experiment outputs to the Airavata Django Portal [1]. Portal admins that register an application should be able to specify what kind of visualizations should be provided to users for each file using a declarative configuration.  Code should not be required for simple charts, but for more complex use cases it should be possible to provide Python code needed to support it.
> As an example, let's say there is an application that generates a CSV file with time series data. The admin user who is registering the application knows what sort of plotting is desired, which columns of data are needed from the CSV file, etc.  The admin user should be able to provide some JSON configuration that describe how to generate the plot. The Django Portal can then use this JSON configuration to process the CSV file and provide the data and configuration to a frontend UI component (implemented in Vue.js) that will then render the desired chart.  The charting technology should allow some interactive features so that users can explore the data easily.
> Some charting libraries that look particularly interesting:
> * https://plot.ly/javascript/
> * http://echarts.apache.org/
> [1] https://github.com/apache/airavata-django-portal



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)