test: add benchmark tests for encode/decode operations

[sumanjeet0012]

Fixes #42

Description

This pull request adds benchmark tests for all encodings to establish performance baselines and help catch regressions. Similar to the benchmarks in go-multibase (BenchmarkEncode, BenchmarkDecode, BenchmarkRoundTrip), these benchmark tests run across all supported implementations.

Changes

• Installed pytest-benchmark as a development dependency in pyproject.toml.
• Added tests/test_benchmarks.py containing test_bench_encode, test_bench_decode, and test_bench_roundtrip parameterized across all ENCODINGS.

[acul71]

Review (maintainer)

Thanks for the clean benchmark addition — it matches #42 and the go-multibase encode/decode/roundtrip shape, and local/CI validation looks good once pytest-benchmark is present.

Blocking

Missing newsfragment — this is a mandatory requirement for approval.

Please add something like:

• newsfragments/42.misc.rst (or 42.internal.rst)

with a short ReST line describing the change for HISTORY, ending with a trailing newline. Example:

Added pytest-benchmark encode/decode/roundtrip benchmarks covering all multibase encodings.

Non-blocking but important

1. Branch is 6 commits behind master — please rebase (or merge) onto current origin/master before landing.
2. CI cost: these benchmarks run in the normal pytest suite (~40s+ locally for 72 timed cases). Consider default-disabling them in regular CI (--benchmark-disable / config) and documenting how to run them explicitly, unless we intentionally want always-on timings.
3. pyproject.toml overlap with open PRs #49 and #57 — informational only; expect a trivial dependency-list conflict when those merge.

Full structured review saved locally as AI-PR-REVIEW-56-0.md.