Conversation
|
The schemas etc are still severely broken, but if @kyleniemeyer and @bryanwweber would like to look at the |
|
The example JSR yaml needs the CSV file to go alongside it |
|
We made some progress. A bit of refactoring, so that |
|
one quick comment: should the csv file have a |
pyked/chemked.py
Outdated
| if units: break | ||
| #todo: schema should enforce at most 1 units entry | ||
|
|
||
| value_properties = [ f'{value} {units}' ] |
There was a problem hiding this comment.
no f-strings in python 3.5, which travis includes?
|
I updated |
Based on https://github.com/comocheng/jsrcomparison/blob/master/Autoscience/new_zhang_jsr_phi0.25.yaml which was updated after a telecon with Kyle and Bryan and Krishna and Richard and Nate. Still has a few comments that need resolving
The idea is we can have multiple types of DataPoint such as IgnitionDataPoint and JSRDataPoint each will inherit from DataPoint. The process_quantitity method looks like a good candidate for something to be in the superclass.
for testing purposes
|
@sevyharris Should I start looking to review this PR or should it be considered in progress? |
|
Hi @mefuller, it's still in progress, but I would love for you to review it when it's ready, hopefully in a day or so! |
|
@sevyharris @rwest great to hear - I am definitely interested in reviving this project and would be happy to see PRs from your group (especially anything related to the issues with cerberus) |
|
@mefuller @sevyharris although I don't have a particular interest in making any final decisions, I'd suggest replacing Cerberus with another library for the data validation. There's been a ton of innovation in this space in the last six years, especially with type checking becoming a true part of the language. I don't have any specific suggestions, but perhaps something like Pydantic would work here (although that's primarily built for web validation). |
7 participants
Not yet ready for review.
CHANGELOG.mdChanges proposed in this pull request:
@pr-omethe-us/chemked