Skip to content

accelerate plotly JSON encoder for numpy arrays without nans #2880

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Nov 17, 2020
Merged

accelerate plotly JSON encoder for numpy arrays without nans #2880

Changes from 2 commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
fcf2a5b
accelerate plotly JSON encoder for numpy arrays without nans
emmanuelle Nov 6, 2020
0fa810f
limit change to numerical data types
emmanuelle Nov 6, 2020
561edf1
solution that works for 99% of cases but lists with composite types w…
emmanuelle Nov 13, 2020
8392ead
other method: brute force string matching of Infinity or NaN
emmanuelle Nov 13, 2020
627c00f
black
emmanuelle Nov 13, 2020
c8e2922
added test
emmanuelle Nov 15, 2020
224082e
black
emmanuelle Nov 15, 2020
5b18b2d
removed ununsed variable
emmanuelle Nov 16, 2020
e528271
changelog entry
emmanuelle Nov 16, 2020
c46bae0
NaN tested before Infinity
emmanuelle Nov 17, 2020
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
13 changes: 11 additions & 2 deletions packages/python/plotly/_plotly_utils/utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -40,10 +40,11 @@ def encode(self, o):
Note that setting invalid separators will cause a failure at this step.

"""

# this will raise errors in a normal-expected way
self.unsafe = False
encoded_o = super(PlotlyJSONEncoder, self).encode(o)

if self.unsafe:
return encoded_o
Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I'm fine with this for now given your description of the performance characteristics.

Copy link
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

maybe check for NaN first, as it feels like that'd appear more frequently?

# now:
# 1. `loads` to switch Infinity, -Infinity, NaN to None
# 2. `dumps` again so you get 'null' instead of extended JSON
Expand Down Expand Up @@ -95,6 +96,14 @@ def default(self, obj):
Therefore, we only anticipate either unknown iterables or values here.

"""
numpy = get_module("numpy", should_load=False)
if numpy:
if isinstance(obj, numpy.ndarray) and numpy.issubdtype(
obj.dtype, numpy.number
):
if numpy.all(numpy.isfinite(obj)):
self.unsafe = True
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not sure if I am missing something. We can test all the numpy arrays that are in the dict being encoded, but we'd not catch nans and infs in any sublists, right?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes this will work only for numpy arrays but for other kinds of objects the existing process (encoding - decoding - reencoding) will still be executed. Or at least this is the intended behaviour :-), maybe I am the one missing something! I need to write tests anyway.

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

but for other kinds of objects the existing process (encoding - decoding - reencoding) will still be executed

I don't think this is the case, since as I understand this class, encode() is not called recursively, or is it?


# TODO: The ordering if these methods is *very* important. Is this OK?
encoding_methods = (
self.encode_as_plotly,
Expand Down