conflict with jupyterlab-variableInspector

[zhehao]

I found that beakerx-jupyterlab is conflicted with jupyterlab-variableInspector, variableInspector widget becomes static(no reaction to variable redefinition or new variables) when I import beakerx into the kernel, here is my jupyterlab environment:

JupyterLab v0.35.4
Known labextensions:
   app dir: ~\AppData\Local\Continuum\anaconda3\share\jupyter\lab
        @jupyter-widgets/jupyterlab-manager v0.38.1 enabled  ok
        @jupyterlab/celltags v0.1.4 enabled  ok
        @jupyterlab/toc v0.6.0 enabled  ok
        @pyviz/jupyterlab_pyviz v0.6.3 enabled  ok
        beakerx-jupyterlab v1.2.0 enabled  ok
        jupyterlab-drawio v0.5.0 enabled  ok
        jupyterlab_variableinspector v0.1.0 enabled  ok*

Is the conflict resolvable? As variableInspector is quite useful for data science, I think many people would like to use it with beakerx.

[piorek]

@LeeTZ
I debugged this and it looks the problem lays in https://github.com/lckr/jupyterlab-variableInspector/blob/8f975ae5cd4624660a1400a7121a148d063cf9b9/src/inspectorscripts.ts#L116

After doing from beaker import * certain objects (more than one) are throwing an exception when doing obj = eval(v) and that leads to vardic being [] hence the error.

Im not an python expert but changing that function to

def _jupyterlab_variableinspector_dict_list():
    def keep_cond(v):
        try:
            obj = eval(v)
            if isinstance(obj, str):
                return True
            if tf and isinstance(obj, tf.Variable):
                return True
            if pd and pd is not None and (
                isinstance(obj, pd.core.frame.DataFrame)
                or isinstance(obj, pd.core.series.Series)):
                return True
            if str(obj)[0] == "<":
                return False
            if  v in ['np', 'pd', 'pyspark', 'tf']:
                return obj is not None
            if str(obj).startswith("_Feature"):
                # removes tf/keras objects
                return False
            return True
        except:
            return False
    values = _jupyterlab_variableinspector_nms.who_ls()
    vardic = [{'varName': _v,
    'varType': type(eval(_v)).__name__,
    'varSize': str(_jupyterlab_variableinspector_getsizeof(eval(_v))),
    'varShape': str(_jupyterlab_variableinspector_getshapeof(eval(_v))) if _jupyterlab_variableinspe$
    'varContent': str(_jupyterlab_variableinspector_getcontentof(eval(_v))),
    'isMatrix': _jupyterlab_variableinspector_is_matrix(eval(_v))}
            for _v in values if keep_cond(_v)]
    return json.dumps(vardic, ensure_ascii=False)

fixed issue for me.

Should I make issue and/or PR in extension repo?

[piorek]

@LeeTZ I will create an issue and PR tomorrow

[piorek]

@zhehao my fix in jupyterlab-variableInspector was merged into master so it should be OK now