https://kedro.org/ logo
Join Slack
Powered by
Hi, after upgrading to kedro-viz 5.2.0 i get the f...
# questions
s
Hi, after upgrading to kedro-viz 5.2.0 i get the following error:
Copy code
kedro.framework.cli.utils.KedroCliError: not enough values to unpack (expected 3, got 1)
Run with --verbose to see the full exception
Error: not enough values to unpack (expected 3, got 1)
Worked with the previous version, anyone knows why this happens? (full stack trace in comments)
d
can you run with verbose option?
s
How do I run it in verbose mode? Have looked in both 
kedro viz
and 
kedros
help page and it doesn't say anything about the 
--verbose
flag. (I also tried, but said "no such option")
d
kedro viz --verbose
s
Something must be very wrong here:
d
weird!
does a normal 
kedro run
work?
s
Yup
a
Uhoh, I think I know what will have caused this… @Merel https://github.com/kedro-org/kedro-viz/pull/1214 something like this: 
_, dataset_type, dataset_file = "a.b".split(".")
😬 1
@Tynan if you were about to do a release announcement, I’d hold off for a second since I think this might need a quick patch fix.
t
ok
a
@Simen Husøy in your data catalog, do you have any entries that specify 
type
as something a bit weird? Like custom dataset types?
s
Yep! Can send you a screenshot
Copy code
radar_data:
  type: birdslib.common.kedro_datasets.falcon_dataset.Falcon_radar_dataset

falcon_test_data:
  type: birdslib.common.kedro_datasets.falcon_dataset.Falcon_test_dataset
  args:
    amount: 2

cross_section_plot:
  type: matplotlib.MatplotlibWriter
  filepath: data/08_reporting/cross_section_plot.png
  save_args:
    format: png

cross_section_plot_plotly:
  type: plotly.JSONDataSet
  filepath: data/08_reporting/cross_section_plot_plotly.json
  layer: reporting
  versioned: true
This is my data_catalog
So two custom dataset classes
a
Hmmm, just to test: please could you comment out those datasets?
s
When running kedro viz, or kedro run?
a
kedro viz
s
Tried to comment out the custom ones, and then the plotting ones as well. Gave equal results
a
Weird… I’m not immediately sure but we should definitely try and fix that 😬 For the time being if you want to try out the plotly image datasets then I’d just use kedro 0.18.3 + kedro-viz 5.1.1 and I think all should work well.
@Jannic Holzer @Rashida Kanchwala can you see why this isn’t working? At a glance I’m 95% sure it’s that PR I linked to above.
s
Yeah, got that working earlier today, so thats nice!
m
Yeah it’s probably the bit where we parse the module and class. Strange viz still doesn’t work when commenting those datasets out though..
j
Taking a look into this now
👍 2
t
@Jannic Holzer any update on this?
r
i faced this issue while writing the tests. it's to do with how dataset. _ class _. _module _ is handled differently for AbstractDataSet and AbstractVersionedDataSet but I am not sure how to solve this as it's framework side issue
j
Still looking into it, it's either (or both) 
_get_dataset_type
(
flowchart.py
) or 
get_dataset_type
(
experiment_tracking.py
) that's causing the issue
👍 1
t
@Simen Husøy what's your GitHub username?
j
Hey @Simen Husøy, we've worked on a potential fix for this problem, though we haven't been able to replicate your error. Could I give you a 
.whl
file of a modified 
5.2.0
release for you to test and see if it resolves your error?
s
Yes, I can see if it solves the problem. Don't know if I'll be able to do it this week, but should be possible next week
👍 2
j
Great, thank you 🙂 I'll prepare the file
a
One quick check @Simen Husøy: do you use transcoding in your project? i.e. are there entries with 
@
in the dataset name in the catalog?
👍 1
s
No, I don't think so
j
@Simen Husøy we released a patch version (5.2.1) to PyPI which could solve your issue, feel free to try and upgrade again.
😍 2
👍 1
s
Cool, I'll try it out within the week
Open in Slack
PreviousNext
Powered by