Hello, can anyone point me to some example kedro s...
# questions
b
Hello, can anyone point me to some example kedro starter template repos? Im working on creating one at work, but theres some confusing parts of the documentation that I think would be cleared up with an example repo
h
Someone will reply to you shortly. In the meantime, this might help:
m
Hi Ben, we've got the official `kedro-starters` repo that contains the templates used when going through the
kedro new
flow. There's also https://github.com/kedro-org/awesome-kedro which contains a list of example projects created by users.
y
And @Ben Shaughnessy What part is confusing in the documentation? Or rather, where are you getting stuck?
b
Hello, thank you for the links! Currently we can't figure out how to create a new project with our own starter package. Whatever we try gives us an error message that boils down to:
Copy code
Kedro project template not found at my_kedro_starter. Specified tag 0.19.10. The following tags are available: . The aliases for the official Kedro starters are:
- astro-airflow-iris
- databricks-iris
- spaceflights-pandas
- spaceflights-pandas-viz
- spaceflights-pyspark
- spaceflights-pyspark-viz
Im able to install one of these fine... but not my own. We have tried using:
kedro new --starter git+<http://github.com/username/my-repo.git>
and also tried creating a plugin.py in a local folder like:
Copy code
# plugin.py
from kedro.framework.cli.starters import KedroStarterSpec

starters = [
    KedroStarterSpec(
        alias="default_ds_starter",
        template_path="<https://github.com/MY-ORG/my-kedro-starter/>",
        directory="default-starter",
    )
]
then running:
kedro new --starter default_ds_starter
which gives the same error
m
You need to provide the directory as well as the starter github repo path so:
kedro new --starter git+<http://github.com/username/my-repo.git> --directory default-starter
b
Hello, the directory works now, but I think I may have found a bug with the
--checkout
command? Im trying to
--checkout my-branch-name
which seems to successfully work with cookiecutter (the temp dir has my git repo and the correct branch) but later on in the process, I get an error saying
Copy code
File "C:\Users\bshaughnessy\AppData\Local\miniconda3\envs\py311-ryi-margin-model\Lib\site-packages\kedro\framework\cli\starters.py", line 467, in _get_cookiecutter_dir
    raise KedroCliError(...)

Specified tag poc-starter. The following tags are available: . The aliases for the official Kedro starters are:
- astro-airflow-iris
- databricks-iris
- spaceflights-pandas
- spaceflights-pandas-viz
- spaceflights-pyspark
- spaceflights-pyspark-viz
with command:
kedro new --starter git@github.com:COMPANY/kedro-starter.git --directory default-starter --checkout my-branch-name --verbose
If im understanding correctly - its treating the branch name and tag name interchangeably, preventing me from using any branches (unless I name it something like
spaceflights-pandas
) This seems like a bug no?
and some more info related to this potential bug: when I run this command:
kedro new --starter C:\Users\bshaughnessy\code\starter-repo\default-starter
I get:
Copy code
kedro.framework.cli.utils.KedroCliError: Kedro project template not found at C:\Users\bshaughnessy\code\dsg-kedro-starters\default-starter. Specified tag 0.19.8. The following tags are available: . The aliases for the official Kedro starters are:
- astro-airflow-iris
- databricks-iris
- spaceflights-pandas
- spaceflights-pandas-viz
- spaceflights-pyspark
- spaceflights-pyspark-viz
note I am not specifying a tag - and the default value is failing