Introduce decorator: wrap and autowrap #7
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit adds a minimal implementation of @tfplot.wrapper.as_op decorator which can wrap a python function to TensorFlow operation (factory). One additional feature is auto-injection of `fig`, `ax` instance to avoid manual invocation of `tfplot.subplots()` if appropriate. Batch support and agglomeration wrapper will be added later.
plus some style improvement on the unit tests
The existing function tfplot.wrap() can actually be turned into
a python decorator providing the same functionality. To this end,
we use biwrap to support two usage of tfplot.wrap():
- As a function: tfplot.wrap(plotter, option='..')
- As a decorator:
```python
@tfplot.wrap
def plotter(tensor_val, option='..'):
# ...
return fig
```
to create a factory of plotting Tensors.
In the same vein, a new decorator @tfplot.autowrap is added.
It provides with additional features such as automatic injecting
matplotlib Figure and Axes objects if specified and handling of
return values (e.g. return ax directly). This will make most of
use cases easier.
Python2 may raise SyntaxError on collecting test suites that contains a function with keyword-only arguments. To work around this, we use a exec trick to dynamically define a decorator-wrapped function. Argspec is handled as well.
wookayin
added a commit
that referenced
this pull request
Dec 19, 2018
As a result, `tfplot.autowrap()` can totally replace `tfplot.wrap_axesplot()` (which now shall be deprecated). Follow-up of #7.
wookayin
added a commit
that referenced
this pull request
Dec 25, 2018
wookayin
added a commit
that referenced
this pull request
Dec 25, 2018
Additional kwargs of autwrap() should be passed through to the python function when executed inside a TF graph. Follow-up of #7.
ferrine
reviewed
Feb 16, 2019
|
|
||
|
|
||
| @biwrap | ||
| def wrap(plot_func=REQUIRED, _sentinel=None, |
There was a problem hiding this comment.
Why not using keyword only arguments? Compatibility?
There was a problem hiding this comment.
I want to support python 2.7 for the time being (See TF's cross entropy loss function).
| @functools.wraps(plot_func) | ||
| def _wrapped_plot_fn(*args, **kwargs_call): | ||
| # (1) auto-inject fig, ax | ||
| if fig_ax_mode: |
| (as a decorator or with normal function call), but provides with additional | ||
| features such as auto-creating matplotlib figures. | ||
|
|
||
| - (``fig``, ``ax``) matplotlib objects are automatically created and |
|
@ferrine Thanks for your comments! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
A prototype implementation of #3.
tfplot.wrap: It can be used as a decorator as well as with a normal function call.tfplot.autowrap: Similar towrap, but supports additional features (e.g. automatic injection offig,ax, etc.)This PR is for reviewing.