fix: preserve punctuation in job metric names

[rioyu123]

Issue #, if available:

N/A

Description of changes:

log_metric writes names such as train/loss and eval/loss unchanged, but LogMetricsParser reads both as loss. The metrics end up in one column, where MIN/MAX selects between unrelated values. Hyphens and periods also cause names to lose their prefixes.

This change preserves the complete non-whitespace name token before =. Value parsing, delimiters and statistics handling are unchanged. Names containing whitespace, = or ; remain outside this fix.

Punctuated names now appear under their full name instead of the truncated suffix. Hand-formatted logs should separate names from preceding text with whitespace, as log_metric already does.

Testing done:

• Added two regression cases that capture actual log_metric output and check slash, hyphen, period and underscore names under MIN and MAX. Both fail on the original parser.
• Linux, CPython 3.12: 3,498 unit tests passed, 131 xfailed.
• tox -e linters_check,type-check,docs,coverage-check passed; coverage is 100%. The docs build completed with 467 warnings.
• Live AWS integration tests were not run. This change only affects local parsing of fetched log text.

Merge Checklist

General

• I have read the CONTRIBUTING doc
• I used the PR title format described in CONTRIBUTING
• I have updated any necessary documentation, including READMEs and API docs (if appropriate)

No documentation change is needed for this parser correction.

Tests

• I have added tests that prove my fix is effective or that my feature works (if appropriate)
• I have checked that my tests are not configured for a specific region or account (if appropriate)

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.