How to build JSON object with python script?

[slashnburn]

I am making a custom module and cannot, for the life of me, figure out how the python script is working. I am using the Example 6 script as a guide and my main trouble is with this piece:

dataset = getattr(job, entity_name)[offset: offset+count]
        outputJSON = {}
        for i, result in enumerate(dataset):
            tdict = {}
            tdict[str(result.get('processor', None))] = str(result.get('value', None))
            name = str(result.get('name', None))
            if name not in outputJSON:
                outputJSON[name] = dict()
            outputJSON[name].update(tdict)

I can see what it is doing, but my results are much different than a simple table. I need a way of pulling out the first value in a table cell, then the second, etc. This script outputs a JSON array of
{name:foo, {processor1 : value}, {processor2:value}, {etc}} .

I need a JSON array of something more complex, like:

{
 "name": "flare",
 "children": [
  {
   "name": "analytics",
   "children": [
    {
     "name": "cluster",
     "children": [
      {"name": "AgglomerativeCluster", "value": 3938},
      {"name": "CommunityStructure", "value": 3812},
      {"name": "HierarchicalCluster", "value": 6714},
      {"name": "MergeEdge", "value": 743}
     ]
    },
    {
     "name": "graph",
     "children": [
      {"name": "BetweennessCentrality", "value": 3534},
      {"name": "LinkDistance", "value": 5731},
      {"name": "MaxFlowMinCut", "value": 7840},
      {"name": "ShortestPaths", "value": 5914},
      {"name": "SpanningTree", "value": 3416}
     ]
    },
    {
     "name": "optimization",
     "children": [
      {"name": "AspectRatioBanker", "value": 7074}
     ]
    }
   ]
  },
  {
   "name": "animate",
   "children": [
    {"name": "Easing", "value": 17010},
    {"name": "FunctionSequence", "value": 5842},
    {
     "name": "interpolate",
     "children": [
      {"name": "ArrayInterpolator", "value": 1983},
      {"name": "ColorInterpolator", "value": 2047},
      {"name": "DateInterpolator", "value": 1375},
      {"name": "Interpolator", "value": 8746},
      {"name": "MatrixInterpolator", "value": 2202},
      {"name": "NumberInterpolator", "value": 1382},
      {"name": "ObjectInterpolator", "value": 1629},
      {"name": "PointInterpolator", "value": 1675},
      {"name": "RectangleInterpolator", "value": 2042}
     ]

In other words, I need to build a JSON object that is much more complex than a 1:1 mapping, like in this example. Does anyone have any ideas on how to complete this? I still haven't been able to figure out if things like tdict[str(result.get( is just a python thing, or a splunk thing.

[CynthiaMhav]

Not sure if I understood your question corrently, but (having already made a long reply, getting deleted by "You cannot include external links with this low karma" <insert cursing>, I'll make it short)

All the function and methods used in your code snippet are from Python's built-in functions and types' methods. None of them are from Splunk's Python SDK.

For reference, see:

https://docs.python.org/2/library/functions.html (getattr(), enumerate(), str())

https://docs.python.org/2/library/stdtypes.html#dict.get

As for how to form a more complex JSON object; you do exactly similar dictionary & list processing.

JSON.dumps() converts python dictionaries { } to JSON objects, and python lists [ ] to JSON arrays.

See https://docs.python.org/2/library/json.html#encoders-and-decoders for full details on how default JSON encoder/decoder works.