Hi. I had a question about <cards and kubernetes y...
# ask-metaflow
h
Hi. I had a question about cards and kubernetes yesterday. I was wondering before heading down the metadata service route if there is any simple way of viewing cards from flows triggered via Argo? I did a test flow and it appears that the cards are being written to s3.
1
c
the cards are html files, so one way if you see them in your S3 bucket is to download that object and open the file locally.
h
I must be doing something wrong them. I get an error when I download and open them.
Copy code
Card of type default is unable to be rendered with arguments None.
Stack trace :  Traceback (most recent call last):
  File "/metaflow/metaflow/plugins/cards/card_cli.py", line 503, in create
    rendered_info = render_card(mf_card, task, timeout_value=timeout)
  File "/metaflow/metaflow/plugins/cards/card_cli.py", line 384, in render_card
    rendered_info = mf_card.render(task)
  File "/metaflow/metaflow/plugins/cards/card_modules/basic.py", line 601, in render
    ).render()
  File "/metaflow/metaflow/plugins/cards/card_modules/basic.py", line 434, in render
    p.id for p in self._task.parent.parent["_parameters"].task if p.id != "name"
  File "/metaflow/metaflow/client/core.py", line 471, in __getitem__
    raise KeyError(id)
KeyError: '_parameters'
Maybe I need to download the whole dir
c
hmm, how are you trying to open the .html file that you downloaded? should just be able to open up in the browser by clicking (e.g., in finder on mac) if I understand the situation correctly
h
ok so I downloaded the whole
.metaflow
dir and it seems to show up correctly. Previously I was only copying the
.metaflow/mf.cards
dir.
makes sense. It must be reading from the run’s metadata to display anything.
Thanks 🙂
c
of course!