Uh oh!
There was an error while loading. Please reload this page.
Conversation
| "uninstall-local": "pip uninstall dash-core-components -y", | ||
| "build:js": "webpack --mode production", | ||
| "build:js-dev": "webpack --mode development", | ||
| "build:r": "node ./extract-meta src/components > dash_core_components/metadata.json && cp package.json dash_core_components && npm run generate-python-classes && npm run generate-r-classes", |
There was a problem hiding this comment.
Does generate-r-classes actually require generate-python-classes to be run first? regardless, the extract-meta and cp package.json parts are going to be shared so I wonder if we just want to have one command build:backend or something that does both languages (and any others we add later)? And use that in build:all (which is currently an exact copy of prepublish but perhaps shouldn't be cc plotly/dash-html-components#83)
There was a problem hiding this comment.
Sorry for not replying to these comments late last month; when I last spoke with @nicolaskruchten and @T4rk1n about component generation in Python vs. R, I believe we decided that support for the latter remains experimental.
Currently this requires setting the --r-prefix flag when calling dash-generate-components, but we could eventually remove the flag and generate both at once.
It does (for now) require that the Python classes are generated first, since JS dependency information is sourced from __init__.py and the R package version number from version.py.
There was a problem hiding this comment.
remind me again why the R transpiler gets the version info from a generated Python file? I know we've talked about this before but I don't recall.
There was a problem hiding this comment.
Realized I had already provided the same reply previously, which didn't really answer your question. I seem to remember talking with Philippe about this, and he suggested reading these data from js_dist in __init__.py, though we don't have to do it this way.
| \item{id}{} | ||
| } | ||
No newline at end of file |
There was a problem hiding this comment.
There is inconsistent whitespace at the end of these .Rd files, as well as trailing whitespace on some of the blank lines above. Can we get rid of these?
| "Textarea", | ||
| "Upload" | ||
| "Tabs", | ||
| "Slider" |
There was a problem hiding this comment.
These were sorted before, now they're not. What changed?
There was a problem hiding this comment.
OK, I've looked into this. The reshuffled order isn't present when cloning the repo, but occurs after running dash-generate-components ./src/components/ dash-core-components, with or without the --r-prefix.
It seems to be outside of the R component generation methods, but I might be mistaken.
There was a problem hiding this comment.
if that's the case, we should make a separate PR that commits that batch of changes, so we can review/approve the Python and R changes separately :)
There was a problem hiding this comment.
@nicolaskruchten Good call; I'll see if I can do just that, and will cross-reference here for posterity.
There was a problem hiding this comment.
@alexcjohnson By commenting about this change, you revealed two distantly related issues that should be addressed. Thanks -- talking with @rmarren1, this is probably related to swapping out _get_metadata for json.loads, and inadvertently not using an OrderedDict anymore.
I'm going to verify this, will then open PR for Dash if true, and link back to here.
| self._valid_wildcard_attributes = [] | ||
| self.available_events = [] | ||
| self.available_properties = ['children', 'id', 'contents', 'filename', 'last_modified', 'accept', 'disabled', 'disable_click', 'max_size', 'min_size', 'multiple', 'className', 'className_active', 'className_reject', 'className_disabled', 'style', 'style_active', 'style_reject', 'style_disabled'] | ||
| self.available_properties = ['children', 'disabled', 'style', 'multiple', 'contents', 'disable_click', 'accept', 'id', 'min_size', 'className', 'style_reject', 'style_active', 'max_size', 'last_modified', 'className_active', 'filename', 'style_disabled', 'className_reject', 'className_disabled'] |
There was a problem hiding this comment.
Similarly, these ones were not sorted before and they're a different unsorted order now. I don't really care about the sort order here, but I don't want unnecessary diffs. What changed?
There was a problem hiding this comment.
Please see above comment, same situation.
After discussion with @alexcjohnson, closing this PR and opening a new one since the commit log is somewhat ungainly and hard to review. Will cross-reference here. |
The JSON >> R 'transpiler' has been rewritten in Python, and was used to generate an installable R package for version 0.42.1 of
dash-core-components.@nicolaskruchten@T4rk1n