simdec-python
simdec-python copied to clipboard
Sensitivity analysis using simulation decomposition
Doc for JOSS: https://joss.readthedocs.io/en/latest/submitting.html#example-paper-and-bibliography
Seems like when the dashboard is starting in GCP, it needs to be reloaded once to work. This is not the case with the sampling page. Which tells me that...
The default procedure does not return states with an equal amount of observations. The screenshot (tested in the dashboard) and the data are attached.  [case1_data.csv](https://github.com/Simulation-Decomposition/simdec-python/files/14308522/case1_data.csv)
With 1.3.1, there is now a Guest user and we can update our OAuth custom code to take advantage of the new feature. https://holoviz-dev.github.io/panel/how_to/authentication/guest_users.html
Should we separate the dashboard from SimDec-Python itself? As we discuss adding more capabilities such as a sampler dashboard or a more general SA app, maybe it would make more...
Stacked histogram: https://discourse.holoviz.org/t/stacked-histogram/6205 Boxplot is fine: ```python df_melted = res.bins.melt() df_melted.hvplot(kind="box", y='value', by='variable', c='variable', invert=True, xlabel="variable", ylabel="value", cmap=colors_hex[::-1], legend=False, width=500, height=500, padding=0.4) ```
For now it's still ok/easy enough to QA. Might be a good idea to add a proper testing of the UI as the app grows. There are some tips there...
Initial `value` are sorted. It comes from the underlying JS library. This is supposed to be fixed in Bokeh 3.0, but there are still issues.
If the output is changed, the input selection field updates to default - all inputs. For a situation when a dataset has multiple outputs and a fixed set of inputs...
Things we need to realize for the two-output analysis: 1. **switch** to two outputs 2. **selection box** for the second output 3. **scatterplot** 4. reduction of points **slicer** (e.g. every...