Json Api help!

About a third of the utility/value of Roku is the ability to read content data from the internet.  It can read rss / xml / json / txt, and any other format that you can retrieve and program the app to make sense of.  When rebuilding the parsing routines, or modifying them to fit your content stream, speed is a key requirement.  I haven't looked at the "hero grid" code, but it is extremely popular.  If you put in a stop after the data is parsed, but before values are assigned, you can use the debugger to see what is there by printing the variable, and trying to get data out of the parsed array.  Ideally, for a content node, you would want to use setfields() as it is the fastest way to push data into a content node but to do that your source feed tags/names need to match what a content node uses.
Note that "Artist" is not currently included in the scenegraph content node, so you would need to use addfield to add it during runtime if you want to use that.
https://sdkdocs.roku.com/display/sdkdoc/ifSGNodeField

https://sdkdocs.roku.com/display/sdkdoc ... +Meta-Data

You want to read the data in whatever format it comes in, in your case JSON. Don't waste time (and resources) converting JSON to XML just so you can cut and paste the code from a demo.
To use JSON in brightscript, first you convert it to an BS associative array using parseJson()
For example, 

url = CreateObject("roUrlTransfer")
url.SetUrl("http://www.mycontentprovider.com/some_feed?format=json")
rsp = url.GetToString()
response_as_associative_array = parseJSON(rsp)

Like destruk said, you then loop through the nodes of your data and assign the values to the SG content node fields. Do you have a link to the hero grid example? (duh I can read.)

looking at the Hero Grid example, you would rewrite the parseResponse() in Parser.brs function to read your JSON data instead of XML.
https://github.com/rokudev/hero-grid-ch ... Parser.brs

"destruk" wrote:
Note that "Artist" is not currently included in the scenegraph content node, so you would need to use addfield to add it during runtime if you want to use that.

You should be able to extend ContentNode to add whatever fields you want, and use your extended ContentNode as if it were an actual ContentNode:

<?xml version="1.0" encoding="UTF-8"?>
<component name="ContentItem" extends="ContentNode">
  <interface>
    <field id="HLS_URL" type="string" />
    <field id="MP3_URL" type="string" />
    <field id="artist" type="string" />
  </interface>
</component>

Good point Belltown -- if the firmware is altered to support Artist for a contentnode, would using the extended class cause a conflict or would adding the extra field this way simply be silently removed so the base class takes precedence?

"destruk" wrote:
Good point Belltown -- if the firmware is altered to support Artist for a contentnode, would using the extended class cause a conflict or would adding the extra field this way simply be silently removed so the base class takes precedence?

According to https://sdkdocs.roku.com/display/sdkdoc/Creating+Custom+Components, under Characteristics of Components Extended from Custom Components:

<interface> fields accumulate: the extended component includes all of the interface fields of the component from which it is extended in addition to its own. In the case where an extended component field name is the same as the component from which it is extended, the definition of the extended component field is used, similar to functions.

It would seem to me that there would be no conflict; however, it's the extended component field name rather than the base class field name that takes precedence.
Note - I haven't tried this out yet. It would be an easy experiment to try overriding another existing field and see what happens.

Thanks Belltown.

Regarding JSON, check out this thread, I provided some example code: https://forums.roku.com/viewtopic.php?f=34&t=102328&sid=bb80464627580f0a1878995549bf7e31

"destruk" wrote:
About a third of the utility/value of Roku is the ability to read content data from the internet.  It can read rss / xml / json / txt, and any other format that you can retrieve and program the app to make sense of.  When rebuilding the parsing routines, or modifying them to fit your content stream, speed is a key requirement.  I haven't looked at the "hero grid" code, but it is extremely popular.  If you put in a stop after the data is parsed, but before values are assigned, you can use the debugger to see what is there by printing the variable, and trying to get data out of the parsed array.  Ideally, for a content node, you would want to use setfields() as it is the fastest way to push data into a content node but to do that your source feed tags/names need to match what a content node uses.
Note that "Artist" is not currently included in the scenegraph content node, so you would need to use addfield to add it during runtime if you want to use that.
https://sdkdocs.roku.com/display/sdkdoc/ifSGNodeField

https://sdkdocs.roku.com/display/sdkdoc ... +Meta-Data

Artist appears to be in the table of attributes on that second link yet I can't seem to use it.

Correct, it's a bug.  You can use a different field if available or add one.