[BUG] ctx.triggered_id is not recognized as a dict #3968

Activity

1. ProfessorOwl commented on Sep 1, 2026

   

   ProfessorOwl

   on Sep 1, 2026

   Author

   More actions

   The same problem also seems to exist with ctx.outputs_grouping.
   I am defining the outputs like this:

          output={
               "drawer": Output("drawer", "opened"),
               "entries": {
                   "name": Output("entry_name", "value"),
                   "amount": Output("entry_amount", "value"),
                   "paid": Output("entry_paid", "checked"),
                   "category": Output("entry_category", "value"),
                   "date": Output("entry_date", "value"),
                   "interval": Output("entry_interval", "value"),
                   "description": Output("entry_description", "value"),
               },
           },

   This gives the following error when calling ctx.outputs_grouping["entries"]

   no overloads for "__getitem__" match the provided arguments
   

   The function .get() is also not recognized as a valid function.

   Both ["entries"] and .get() work, nonetheless
2. AnnMarieW commented on Sep 1, 2026

   

   AnnMarieW

   on Sep 1, 2026

   Collaborator

   More actions

   @ProfessorOwl

   Could you please include a minimal example to replicate the issue? Here is one that show that ctx.triggered_id can be None when the callback first runs:

   from dash import Dash, html, Input, Output, ALL, ctx
   
   app = Dash()
   
   app.layout = html.Div([
       html.Button("Entry 1", id={"type": "entry", "index": 1}),
       html.Button("Entry 1", id={"type": "entry", "index": 1}),
       html.Div(id="out")
   ])
   
   
   @app.callback(
       Output("out", "children"),
       Input({"type": "entry", "index": ALL}, "n_clicks"),
       #prevent_initial_call=True,
   )
   def update_entries(n_clicks):
       return f"n: {n_clicks},    ctx: {ctx.triggered_id}"
   
   
   if __name__ == "__main__":
       app.run(debug=True)
   
3. ProfessorOwl commented on Sep 2, 2026

   

   ProfessorOwl

   on Sep 2, 2026

   Author

   More actions

   HI @AnnMarieW, sorry, I should have done that from the beginning.
   If I modify your example and put the .get function and ["index"] subscript behind ctx.triggered_id, then Pylance displays an error (reportOptionalMemberAccess for .get() and reportOptionalSubscript for ["index"]. I now see that if I test beforehand that ctx.triggered_idis not None, then the errors disappear. But since prevent_initial_call=True is set, only inputs with a dict as an id can trigger the callback, so the error is obsolete. But I can understand now that it can be argumented that – for type safety – there should be an additional check in place, independent from prevent_initial_call.

   from dash import Dash, html, Input, Output, ALL, ctx
   from dash.exceptions import PreventUpdate
   
   
   app = Dash()
   
   app.layout = html.Div(
       [
           html.Button("Entry 1", id={"type": "entry", "index": 1}),
           html.Button("Entry 1", id={"type": "entry", "index": 2}),
           html.Div(id="out"),
       ]
   )
   
   
   @app.callback(
       Output("out", "children"),
       Input({"type": "entry", "index": ALL}, "n_clicks"),
       prevent_initial_call=True,
   )
   def update_entries(n_clicks):
       # if ctx.triggered_id is None:
       #     raise PreventUpdate
       return f"n: {n_clicks},    ctx.get: {ctx.triggered_id.get("index")},    ctx.subscript: {ctx.triggered_id["index"]}"
   
   if __name__ == "__main__":
       app.run(debug=True)

   The error regarding ctx.outputs_grouping can also be circumvented by checking if isinstance(ctx.outputs_grouping, list), but since the outputs are statically written out as a dict, it would be nice if that could be somehow given over to Pylance. The issue Pylance reports is reportAttributeAccessIssue

   Here is a MWE:

   from dash import Dash, html, Input, Output, ALL, ctx
   from dash.exceptions import PreventUpdate
   
   app = Dash()
   
   app.layout = html.Div(
       [
           html.Button("Entry 1", id={"type": "entry", "index": 1}),
           html.Button("Entry 1", id={"type": "entry", "index": 2}),
           html.Div(id="out"),
       ]
   )
   
   
   @app.callback(
       output={"div1": Output("out", "children")},
       inputs=Input({"type": "entry", "index": ALL}, "n_clicks"),
       prevent_initial_call=True,
   )
   def update_entries(n_clicks):
       # if isinstance(ctx.outputs_grouping, list):
       #     raise PreventUpdate
       return {key: f"n: {n_clicks}" for key in ctx.outputs_grouping.keys()}
   
   
   if __name__ == "__main__":
       app.run(debug=True)

   I understand now that the type problems can be mitigated by extensive type checking. If this is too extensive to implement, then I fully understand, though it is cumbersome to have all these lines of typechecking in place, where the types are pretty much hardcoded.
4. camdecoster commented on Sep 2, 2026

   

   camdecoster

   on Sep 2, 2026

   Contributor

   More actions

   I'm inclined to leave the types as is, but I'll leave this open for now to see if anyone else would like to see a change.

   👍React with 👍1Reacted by ProfessorOwl
5. camdecoster

   added

   P3backlogbacklog

   taskone-off taskone-off task

   size: 1

   on Sep 2, 2026