dj.ERD(schema) runs into TypeError

[horsto]

Bug Report

schema = dj.schema('MY_DATABASE')
dj.ERD(schema) 

runs into

File ~\miniconda3\envs\detectron\lib\site-packages\datajoint\diagram.py:428, in Diagram.make_svg(self)
    425 def make_svg(self):
    426     from IPython.display import SVG
--> 428     return SVG(self.make_dot().create_svg())

File ~\miniconda3\envs\detectron\lib\site-packages\datajoint\diagram.py:306, in Diagram.make_dot(self)
    304 def make_dot(self):
--> 306     graph = self._make_graph()
    307     graph.nodes()
    309     scale = 1.2  # scaling factor for fonts and boxes

File ~\miniconda3\envs\detectron\lib\site-packages\datajoint\diagram.py:292, in Diagram._make_graph(self)
    288 nx.set_node_attributes(
    289     graph, name="node_type", values={n: _get_tier(n) for n in graph}
    290 )
    291 # relabel nodes to class names
--> 292 mapping = {
    293     node: lookup_class_name(node, self.context) or node
    294     for node in graph.nodes()
    295 }
    296 new_names = [mapping.values()]
    297 if len(new_names) > len(set(new_names)):

File ~\miniconda3\envs\detectron\lib\site-packages\datajoint\diagram.py:293, in <dictcomp>(.0)
    288 nx.set_node_attributes(
    289     graph, name="node_type", values={n: _get_tier(n) for n in graph}
    290 )
    291 # relabel nodes to class names
    292 mapping = {
--> 293     node: lookup_class_name(node, self.context) or node
    294     for node in graph.nodes()
    295 }
    296 new_names = [mapping.values()]
    297 if len(new_names) > len(set(new_names)):

File ~\miniconda3\envs\detectron\lib\site-packages\datajoint\table.py:980, in lookup_class_name(name, context, depth)
    972 elif (
    973     node["depth"] > 0
    974     and inspect.ismodule(member)
    975     and member.__name__ != "datajoint"
    976 ):
    977     try:
    978         nodes.append(
    979             dict(
--> 980                 context=dict(inspect.getmembers(member)),
    981                 context_name=node["context_name"] + "." + member_name,
    982                 depth=node["depth"] - 1,
    983             )
    984         )
    985     except ImportError:
    986         pass  # could not import, so do not attempt

File ~\miniconda3\envs\detectron\lib\inspect.py:458, in getmembers(object, predicate)
    454 # :dd any DynamicClassAttributes to the list of names if object is a class;
    455 # this may result in duplicate entries if, for example, a virtual
    456 # attribute with the same name as a DynamicClassAttribute exists
    457 try:
--> 458     for base in object.__bases__:
    459         for k, v in base.__dict__.items():
    460             if isinstance(v, types.DynamicClassAttribute):

TypeError: '_ClassNamespace' object is not iterable

Description

Sometimes (and I don't know exactly when), the generation of dj Diagrams does not work. See error above.

Reproducibility

Include:

• OS: Windows 11 Pro
• Python 3.10.6
• DataJoint Version 0.13.8

[kabilar]

Hi @horsto, is this still an issue for you? If so, could you please send me a copy of your module so that I can replicate this error?

[dimitri-yatsenko]

@horsto If you can reproduce this, please let us know. Otherwise, we will close this.

[dimitri-yatsenko]

This may be already fixed and we cannot reproduce it. This may have been fixed by #1177