Hi again, Speaking of `kedro ipython` : Is there ...
# questions
m
Hi again, Speaking of
kedro ipython
: Is there a way to specify /override the
conf_source
? 🙂 Thx
f
👋 , if you use
--env
, you can use the environment you wish.
a
From
CONF_SOURCE
in
settings.py
, I believe
f
My bad, I misunderstood the question.
m
@Flavien: No worries. Thx anyway 🙂
@Ankita Katiyar Thx. That’s how I did it so far. But not really convenient. My problem comes from the fact that we have a single pipeline / repo for all our clients. To better enforce “tenant isolation” I have organized
conf
and
data
as follow:
Copy code
conf
└── client_A
    ├── base
    └── local
└── client_B
    ├── base
    └── local
Copy code
data
├── client_A
│   └── 01_raw
│   └── 02_intermediate
│   └── ...
└── client_B
│   └── 01_raw
│   └── 02_intermediate
│   └── ...conf
I did so because I liked the idea of not being able to run the pipeline without being explicit about which client it should be done for… (the structure above makes it so that
kedro run
without
--conf-source
will raise an error) Another reason I did so (if I recall correctly) was to avoid having “duplicate keys” conflicts in the different yaml files across clients…
n
I think this should be available in the
reload_kedro
magic, similarly you can do
reload_kedro --env=prod
(which we haven’t added)
@Ankita Katiyar Can you help to draft an issue?
a
@Nok Lam Chan Yep!
thankyou 1
m
Thx @Nok Lam Chan & @Ankita Katiyar Another similar issue (which I’ll leave here since related) With the structure above, I can’t use
kedro viz
It doesn’t have a
conf-source
param and
kedro viz --env client_A/base
does not work. Am I just being “a pain” or does this sound like a legit expectation 🙂
n
I think this is valid - though in this case settings.py would still be a workaround for now.
m
Thx Nok. Indeed, for now, I’ll specify the conf_source with setting.py 🙂
👍🏼 1
👍 1
n
feel free to open an issue in viz repo https://github.com/kedro-org/kedro-viz/issues
m
Thx. Will do.