Imagery does not load and render in the labelling window

[dbuscombe-usgs]

I have made doodler python conda environments following the posted instructions (py38) as well as envs based on python versions 3.7, 3.9, and 3.10. I also trialled firefox, chrome and edge browsers. I made environments using WSL on windows 11 and windows 10 (2 different machines). Each time, I ran into the same issue. When I select an image, and navigate to the labeling screen, the image does not appear.

I suspect something is going wrong with the timing

dash_doodler/app.py

Line 506 in 41ce96b

dcc.Interval(id='interval-component', interval=500, n_intervals=0)]),

@2320sharon investigated this some (thank you) and her summary is below:

In update_output which is called periodically to update the figure I configured it to display the default image if the select_image_value (which is the filename of the image selected by the user to label) is None. When I did this the main page kept updating to the default image even after I selected an image which displayed for a second then updated to the default image.

Something is causing select_image_value to be set to None and update_output is being called every 500ms causing the main figure to clear the image. The question is why is select_image_value being cleared

I can't seem to figure out where select_image_value is from. I can only find it defined in the update_output function where it is passed in as a variable

[dbuscombe-usgs]

select_image_value is inherited from the UI components and is present in the callback. Plotly works by funneling all inputs, outputs, and states, into a single callback function

It is set by Input("select-image", "value"), below

@app.callback(
    [
    Output("select-image","options"),
    Output("graph", "figure"),
    Output("image-list-store", "data"),
    Output("masks", "data"),
    Output('my-div', 'children'),
    Output("segmentation", "data"),
    Output('thisimage_output', 'value'),
    Output("pen-width-display", "children"),
    Output("theta-display", "children"),
    Output("mu-display", "children"),
    Output("crf-downsample-display", "children"),
    # Output("crf-gtprob-display", "children"),
    Output("rf-downsample-display", "children"),
    Output("numscales-display", "children"),
    Output("classified-image-store", "data"),
    ],
    [
    Input("upload-data", "filename"),
    Input("upload-data", "contents"),
    Input("graph", "relayoutData"),
    Input(
        {"type": "label-class-button", "index": dash.dependencies.ALL},
        "n_clicks_timestamp",
    ),
    Input("crf-theta-slider", "value"),
    Input('crf-mu-slider', "value"),
    Input("pen-width", "value"),
    Input("crf-show-segmentation", "value"),
    Input("crf-downsample-slider", "value"),
    Input("rf-downsample-slider", "value"),
    Input("numscales-slider", "value"),
    Input("select-image", "value"),
    Input('interval-component', 'n_intervals'),
    ],
    [
    State("image-list-store", "data"),
    State('my-id', 'value'),
    State("masks", "data"),
    State("segmentation", "data"),
    State("classified-image-store", "data"),
    ],
)

[dbuscombe-usgs]

I tried switching backend server, from flask

server = Flask(__name__)
app = dash.Dash(server=server)

to

app = dash.Dash()
server = app.server

but that did not fix