Robustify SQLA version determination logic

[Dev-iL]

1. What GitHub issue is this PR addressing? Make sure that there is an accompanying issue to your PR.

   Fixes SNOW-3066571: SQLAlchemy compat script doesn't handle versions containing letters  #646

2. Fill out the following pre-review checklist:

   • I am adding a new automated test(s) to verify correctness of my new code
   • I am adding new logging messages
   • I am adding new credentials
   • I am adding a new dependency

3. Please describe how your code solves the related issue.

   The version string SA_VERSION can contain pre-release suffixes like 2.1.0b1, 2.0.0rc2, or 2.0.5.post1. The original code splits on "." and calls int() on each segment, which raises ValueError for segments like 0b1 or post1 that contain non-numeric characters.

   The fix uses re.match(r"(\d+)\.(\d+)\.(\d+)", SA_VERSION) to extract only the three leading numeric version components, ignoring any suffix. This works because the regex captures digits up to the first non-digit in each segment, and stops after the third group, so 2.1.0b1 yields (2, 1, 0), 2.0.5.post1 yields (2, 0, 5), etc.

   It is worth noting that this can be achieved more simply (and correctly) using an additional dependency, packaging, by doing

   from packaging.version import parse as parse_version
   
   IS_VERSION_20 = parse_version(SA_VERSION) >= parse_version("2.0.0")

   With the key difference being that packaging correctly treats pre-releases as less than the release (2.0.0b1 < 2.0.0), while the regex treats them as equal.

[github-actions]

All contributors have signed the CLA ✍️ ✅
_{Posted by the CLA Assistant Lite bot.}

[Dev-iL]

I have read the CLA Document and I hereby sign the CLA

[sfc-gh-mraba]

Thanks for for the contribution.
Pls take a look at the comments.

[potiuk]

Nice one!