Skip to content
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

Jump to bottom

WIP: Avoid recursion in _extract_serialize #4258

Closed
wants to merge 2 commits into from
Closed

WIP: Avoid recursion in _extract_serialize #4258

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
d9eaa0a
Drop extra type check in `_extract_serialize`
jakirkham Nov 25, 2020
41f4a0c
Avoid recursion in `_extract_serialize`
jakirkham Nov 25, 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
58 changes: 35 additions & 23 deletions distributed/protocol/serialize.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
from array import array
from collections import deque
from functools import partial
from itertools import repeat
import traceback
import importlib
from enum import Enum
Expand Down Expand Up @@ -445,34 +445,46 @@ def extract_serialize(x):
>>> extract_serialize(msg)
({'op': 'update'}, {('data',): <Serialize: 123>}, set())
"""
x2 = type(x)()
ser = {}
bytestrings = set()
_extract_serialize(x, x2, ser, bytestrings)
return x2, ser, bytestrings


def _extract_serialize(x, x2, ser, bytestrings, path=()):
typ_x = type(x)
if typ_x is dict:
x_items = x.items()
x2 = {}
elif typ_x is list:
x_items = enumerate(x)
x2.extend(repeat(None, len(x)))

for k, v in x_items:
path_k = path + (k,)
typ_v = type(v)
if typ_v is dict or typ_v is list:
x2[k] = v2 = typ_v()
_extract_serialize(v, v2, ser, bytestrings, path_k)
elif typ_v is Serialize or typ_v is Serialized:
ser[path_k] = v
elif (typ_v is bytes or typ_v is bytearray) and len(v) > 2 ** 16:
ser[path_k] = to_serialize(v)
bytestrings.add(path_k)
x2 = len(x) * [None]

ser = {}
bytestrings = set()
_extract_serialize(x_items, x2, ser, bytestrings)
return x2, ser, bytestrings


def _extract_serialize(x_items, x2, ser, bytestrings, path=()):
q = deque()
Copy link
Member

Choose a reason for hiding this comment

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

If we're using this as only a stack then maybe this should be a list rather than a deque

Copy link
Member Author

Choose a reason for hiding this comment

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

Yeah I tried both. This performed better based on the benchmarking I did. That said, it seems that recursion itself is already handled quite efficiently.

while True:
for k, v in x_items:
path_k = path + (k,)
typ_v = type(v)
if typ_v is dict:
v_items = v.items()
x2[k] = v2 = {}
q.append((v_items, v2, path_k))
elif typ_v is list:
v_items = enumerate(v)
x2[k] = v2 = len(v) * [None]
q.append((v_items, v2, path_k))
elif typ_v is Serialize or typ_v is Serialized:
ser[path_k] = v
elif (typ_v is bytes or typ_v is bytearray) and len(v) > 2 ** 16:
ser[path_k] = to_serialize(v)
bytestrings.add(path_k)
else:
x2[k] = v

if q:
x_items, x2, path = q.pop()
else:
x2[k] = v
break


def nested_deserialize(x):
Expand Down