You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Metadata for ArviZ's example datasets (which are loadable by {func}`~arviz_base.load_arviz_data`) is stored in the [`arviz_example_data`](https://github.com/arviz-devs/arviz_example_data) repository.
4
+
This repo has been embedded into the `arviz` repo at `/src/arviz_base/example_data` using [git subtree](https://www.atlassian.com/git/tutorials/git-subtree) with the following command:
5
+
6
+
```bash
7
+
$ git subtree add --prefix src/arviz_base/example_data https://github.com/arviz-devs/arviz_example_data.git main --squash
8
+
```
9
+
10
+
When `arviz_example_data` is updated, the subtree within the `arviz` repo also needs to be updated with the following command:
11
+
12
+
```bash
13
+
$ git subtree pull --prefix src/arviz_base/example_data https://github.com/arviz-devs/arviz_example_data.git main --squash
14
+
```
15
+
16
+
Moreover, to keep the size of the wheel and sdist for `arviz-base` as small as possible, the `code` folder of `arviz_example_data` is also removed as it contains some jupyter notebooks and csv files.
0 commit comments