perf: optimize timestamp parsing in metadata processing

[raj-prince]

Description:
Optimizes GCSFileSystem._parse_timestamp by replacing legacy string slicing and zero-padding logic with @staticmethod and datetime.fromisoformat(timestamp.replace("Z", "+00:00")).

This yields a ~2.2x speedup on timestamp parsing and reduces total _process_object latency by ~33%.

Rationale & Backward Compatibility

• The manual zero-padding and rsplit logic was originally added as a workaround for older Python versions (<3.10) where fromisoformat() was restrictive about fractional-second padding.
• Since gcsfs requires python >= 3.10, standard GCS RFC 3339 timestamps (milliseconds .000Z, microseconds .ffffffZ, and whole seconds Z) are natively supported by fromisoformat() when replacing "Z" with "+00:00".
• Reference: Python 3.10 datetime.fromisoformat() Specification

Benchmarks (timeit, 500k iterations)

Scenario	Legacy	Optimized	Speedup
_parse_timestamp (milliseconds .000Z)	0.298 µs	0.136 µs	2.19x (54.4% less CPU)
_parse_timestamp (microseconds .123456Z)	0.269 µs	0.139 µs	1.93x (48.2% less CPU)
_parse_timestamp (whole seconds Z)	0.250 µs	0.130 µs	1.92x (47.9% less CPU)
Full _process_object (2 timestamps)	0.870 µs	0.584 µs	1.49x (32.9% less CPU)

Testing

• Added test_parse_timestamp_formats and test_process_object_timestamps to gcsfs/tests/test_core.py covering all GCS timestamp formats and precisions.

[gemini-code-assist]

Code Review

This pull request refactors the _parse_timestamp method in gcsfs/core.py to be a static method and simplifies its implementation using datetime.fromisoformat. It also adds comprehensive unit tests to verify various timestamp formats. However, the simplified implementation introduces a compatibility issue on Python 3.10, where datetime.fromisoformat does not support arbitrary fractional second lengths (such as 1, 2, 4, or 5 digits), leading to potential runtime crashes and test failures. A conditional fallback based on the Python version is recommended to maintain compatibility.

[gemini-code-assist]

In Python 3.10, datetime.fromisoformat does not support arbitrary fractional second lengths (it only supports exactly 0, 3, or 6 digits). If a timestamp has a different number of fractional digits (such as 1, 2, 4, or 5 digits, which are explicitly covered in the new test cases like 2024-01-15T10:30:00.12Z), fromisoformat will raise a ValueError on Python 3.10.

Since gcsfs supports Python >= 3.10, this change will cause test failures and runtime crashes for Python 3.10 users.

To resolve this while maintaining maximum performance:

1. On Python 3.11+, we can use the native datetime.fromisoformat directly (which natively supports the Z suffix and arbitrary fractional seconds, making it even faster as it avoids string replacement).
2. On Python 3.10, we can use a fast compatibility fallback that only pads the fractional seconds when their length is not 3 or 6.

We can achieve this with zero runtime overhead by conditionally defining the @staticmethod at class definition time using sys.version_info.

    if sys.version_info >= (3, 11):
        @staticmethod
        def _parse_timestamp(timestamp):
            return datetime.fromisoformat(timestamp)
    else:
        @staticmethod
        def _parse_timestamp(timestamp):
            if timestamp.endswith('Z'):
                timestamp = timestamp[:-1]
                if '.' in timestamp:
                    frac_len = len(timestamp.rsplit('.', 1)[-1])
                    if frac_len not in (3, 6):
                        timestamp = timestamp + '0' * (6 - frac_len)
                return datetime.fromisoformat(timestamp + '+00:00')
            return datetime.fromisoformat(timestamp)

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 90.18%. Comparing base (dd48b07) to head (5956280).
⚠️ Report is 8 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #1045      +/-   ##
==========================================
+ Coverage   90.10%   90.18%   +0.07%     
==========================================
  Files          16       16              
  Lines        3679     3739      +60     
==========================================
+ Hits         3315     3372      +57     
- Misses        364      367       +3     

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

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.