[bug] plotly doesn't work w/ python

[Almenon]

I get an error when trying to generate a plotly graph w/ python:

from json import dumps
import random

while(True):
    data = []
    value = 0
    for i in range(10000):
        delta = 1 if random.random() > .5 else -1
        data.append(value)
        value += delta

    json_data = dumps({
	    "kind": { "plotly": True },
	    "data": data
    })
    print('done')

I evaluated json_data in the debug visualizer console and there was just a blank screen. I opened up the webview console logs and I found this:

bootstrap:105 Uncaught (in promise) ChunkLoadError: Loading chunk 2 failed.
(error: )
    at Function.i.e (http://localhost:52480/main.js:1:724)
    at Object.loader (http://localhost:52480/main.js:35:1189248)
    at e.load (http://localhost:52480/main.js:6:12838)
    at http://localhost:52480/main.js:6:13195
    at http://localhost:52480/main.js:236:538979
    at e.fire (http://localhost:52480/main.js:1:63465)
    at e._getOrCreateMode (http://localhost:52480/main.js:236:451829)
    at e._createModeAndGetLanguageIdentifier (http://localhost:52480/main.js:236:451509)
    at t._selector (http://localhost:52480/main.js:236:451123)
    at new t (http://localhost:52480/main.js:236:449819)
bootstrap:105 Uncaught (in promise) ChunkLoadError: Loading chunk 64 failed.
(error: )
    at Function.i.e (http://localhost:52480/main.js:1:724)
    at p (http://localhost:52480/main.js:35:2001026)
    at http://localhost:52480/main.js:35:2001448
    at http://localhost:52480/main.js:236:538979
    at e.fire (http://localhost:52480/main.js:1:63465)
    at e._getOrCreateMode (http://localhost:52480/main.js:236:451829)
    at e._createModeAndGetLanguageIdentifier (http://localhost:52480/main.js:236:451509)
    at t._selector (http://localhost:52480/main.js:236:451123)
    at new t (http://localhost:52480/main.js:236:449819)
    at e.create (http://localhost:52480/main.js:236:451045)
ExpressionInput.tsx:105 json_data
factory.js:104 Error while plotting: TypeError: Cannot use 'in' operator to search for 'bardir' in 0
    at Object.n.cleanData (plotly.js:110967)
    at Object.n.plot (plotly.js:111744)
    at Object.n.newPlot (plotly.js:112273)
    at Object.n.react (plotly.js:114298)
    at factory.js:89
(anonymous) @ factory.js:104
Promise.catch (async)
value @ factory.js:99
value @ factory.js:115
yl @ react-dom.production.min.js:251
t.unstable_runWithPriority @ scheduler.production.min.js:18
Gi @ react-dom.production.min.js:120
bl @ react-dom.production.min.js:244
ol @ react-dom.production.min.js:223
(anonymous) @ react-dom.production.min.js:121
t.unstable_runWithPriority @ scheduler.production.min.js:18
Gi @ react-dom.production.min.js:120
Qi @ react-dom.production.min.js:121
Yi @ react-dom.production.min.js:120
al @ react-dom.production.min.js:224
Je @ mobx.module.js:1905
Xe @ mobx.module.js:1881
Ge @ mobx.module.js:1581
e.reportChanged @ mobx.module.js:258
t.setNewValue @ mobx.module.js:1066
e.write @ mobx.module.js:4032
set @ mobx.module.js:4239
updateState @ Model.ts:186
(anonymous) @ TypedChannelContracts.js:127
handleNotification @ TypedChannel.js:210
handleNotification @ TypedChannel.js:53
processRequestOrNotification @ StreamBasedChannel.js:54
(anonymous) @ StreamBasedChannel.js:16
onMessage @ MessageStream.js:17
e.onmessage @ index.js:34

[hediet]

Your data field of the json object does not match the plotly data type.
This json does:

{ kind: { plotly: true }, data: [
          {
            "x": [1, 2, 3],
            "y": [2, 6, 3],
            "type": "scatter",
            "mode": "lines+markers",
            "marker": { "color": "red" },
          }
] }

[Almenon]

Ah, that makes sense. I thought it would work because the script was a direct conversion of https://github.com/hediet/vscode-debug-visualizer/blob/master/demo/js/src/demo_random-walks.ts to python. But like you said in the other issue python doesn't have the extractors necessary to automatically make sure the schema matches. Or at least not yet.

[hediet]

Yes. Closing this as it works as intended. Data Extractors for Python can be tracked in a separate issue.