Nixtla's Open Source Time Series Ecosystem.

Nixtla Community

Hi guys, in the <https://github.com/Nixtla/hierarchicalforecast/blob/162e42f4143201d44fdd0480306dd4cc19776038/hierarchicalforecast/utils.py#L805|current implementation> of the `plot_hierarchical_predictions_gap` function, we cannot get the matplotlib Figure (fig) object. Is that correct? I don't

I want to get access to the fig object so that I can save the plot as a svg file:

```fig.savefig("images/output_plot.svg", format="svg", bbox_inches="tight")```
Should we return the fig object at the end of the function?

```plt.show()
return fig, ax```

If you only want to save the figure, just returning `fig` is enough. But I don't see a problem with returning the `ax` either...

yeah returning a fig is enough. Should I submit a PR for this?

<@U06H2MDHK7H> Can you review <https://github.com/Nixtla/hierarchicalforecast/pull/371|this PR>?

You need to modify the notebook and use nbdev to make that kind of changes. You can look at CONTRIBUTING.md for details

Is it normal for conda to take forever to solve the environment using this command? We only use a few packages

```$ conda env update -f environment.yml
Collecting package metadata (repodata.json): done
Solving environment: ```

I wonder if running `nbdev_clean` and `nbdev_export` in the <https://github.com/Nixtla/hierarchicalforecast/blob/main/CONTRIBUTING.md|CONTRIBUTING.md> is necessary if we have <https://github.com/Nixtla/hierarchicalforecast/blob/162e42f4143201d44fdd0480306dd4cc19776038/.pre-commit-config.yaml#L10|nbdev pre-commit hooks>