Defer importing ens._normalization to speed up web3 import by 27% #3285
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
As of version 7.0.0-beta.2 the `import web3` command takes 375 ms to
run*, which is quite long for a library import operation. This affects
the usability of command-line tools built around web3.py, where import
times are critical.
Profiling* showed that the largest contributor to this is the
`ens._normalization` module that performs two lengthy operations at
import time:
- Loading `ens/spec/normalization_spec.json` that is 3 Mbytes in size
takes 7% of the total import time
- Executing the `_extract_valid_codepoints()` function that performs
transformations on the loaded data takes 16% of the total import time
These are very costly operations to perform at every single import, with
the sole purpose of initializing the `ens.utils.normalize_name`
function.
By deferring import until the first call of the function, the execution
time of `import web3` is reduced to 275 ms* i.e. by 27%.
*: Measurements were taken with `timeit` and `pyinstrument` on an AMD
Ryzen 7 machine with M.2 SSD using Python 3.10.12:
>>> timeit.timeit("import web3")
and
$ pyinstrument --show-regex '.*(eth|web3|ens).*' -c 'import web3'
kclowes
approved these changes
Mar 15, 2024
|
Just kidding, looks like I/maintainers can't add to your PR. Once you add a newsfragment I can merge. Thanks! |
|
Newsfragment added, thank you! |
|
Thank you! |
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Outdated suggestions cannot be applied.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.
What was wrong?
As of version 7.0.0-beta.2 the
import web3command takes 375 ms to run*, which is quite long for a library import operation. This affects the usability of command-line tools built around web3.py, where import times are critical.How was it fixed?
By deferring the import of
ens._normalizationuntil the first call of theens.utils.normalize_namefunction, the execution time ofimport web3is reduced to 275 ms* i.e. by 27%.Todo:
Details
Profiling* showed that the
ens._normalizationmodule performs two lengthy operations at import time:ens/spec/normalization_spec.jsonthat is 3 Mbytes in size takes 7% of the total web3 import time_extract_valid_codepoints()function that performs transformations on the loaded data takes 16% of the total web3 import timeThese are very costly operations to perform at every single import, with the sole purpose of initializing the
ens.utils.normalize_namefunction.*: Measurements were taken with
timeitandpyinstrumenton an AMD Ryzen 7 machine with M.2 SSD using Python 3.10.12:and