PERF: Use bytearray instead of b"" in encode_pdfdocencoding

[zuypt]

Since b"" is not mutable it causes python to allocate and deallocate memory repeatedly in the for loop which cause hang/long runtime when handle very large string. For example when using add_js to to add a very big javascript code.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (40e25ec) 94.37% compared to head (e3ec6cc) 94.37%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2325   +/-   ##
=======================================
  Coverage   94.37%   94.37%           
=======================================
  Files          43       43           
  Lines        7660     7660           
  Branches     1515     1515           
=======================================
  Hits         7229     7229           
  Misses        267      267           
  Partials      164      164           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stefan6419846]

Could you please update the title to use the recommended naming scheme? https://pypdf.readthedocs.io/en/latest/dev/intro.html#commit-messages

[MartinThoma]

@zuypt Do you have an example that shows the difference? (It could be a toy-example - I'm just curious :-) )

[zuypt]

Could you please update the title to use the recommended naming scheme? https://pypdf.readthedocs.io/en/latest/dev/intro.html#commit-messages

I'm too lazy if some one have permission please help

[zuypt]

@zuypt Do you have an example that shows the difference? (It could be a toy-example - I'm just curious :-) )

just create a PdfWriter then call add_js with a super large string you will see. This is a pretty common python programming error.

[MartinThoma]

I've already adjusted the title

[MartinThoma]

import timeit

def benchmark_empty_bytes_literal():
    result = b""
    for _ in range(100000):
        result += b"a"

def benchmark_bytes_object():
    result = bytearray()
    for _ in range(100000):
        result += b"a"

if __name__ == "__main__":
    empty_bytes_literal_time = timeit.timeit(benchmark_empty_bytes_literal, number=100)
    bytes_object_time = timeit.timeit(benchmark_bytes_object, number=100)

    print(f"Empty Bytes Literal Time: {empty_bytes_literal_time:.1f}")
    print(f"bytearray Time: {bytes_object_time:.1f}")

shows:

Empty Bytes Literal Time: 21.4
bytearray Time: 0.5

[MartinThoma]

@zuypt Thanks for your contribution! If you want, I can add you to https://pypdf.readthedocs.io/en/latest/meta/CONTRIBUTORS.html

[MartinThoma]

It will be part of the next release on Sunday.

[zuypt]

@zuypt Thanks for your contribution! If you want, I can add you to https://pypdf.readthedocs.io/en/latest/meta/CONTRIBUTORS.html

sure. Thanks for the recognition