Add support for environment context

[ebyhr]

Rationale for this change

Adds EnvironmentContext utility class to track engine metadata (name and version) and automatically includes it in snapshot summaries. This provides visibility into which engine version created/modified each snapshot.

Are these changes tested?

Yes

Are there any user-facing changes?

No - it's internal change

[abnobdoss]

This is great! This looks very helpful for debugging.

[kevinjqliu]

Thanks for the PR! This will be very useful to users. I got a couple of nits

[kevinjqliu]

this will add EnvironmentContext to all ViewVersion which might not be the desired behavior.

Similar to Table Metadata, can we only add EnvironmentContext in the write path?

[kevinjqliu]

Could we find a way to avoid repeating env context everywhere? Perhaps a small test helper like this would keep the exact assertions while localizing this new default behavior:

def with_environment_context(summary: dict[str, str]) -> dict[str, str]:
    return {**summary, **EnvironmentContext.get()}

assert summaries[0] == with_environment_context({
    "added-data-files": "3",
    "added-records": "5",
    ...
})

[kevinjqliu]

Suggested change

	def test_default_value() -> None:
	actual = EnvironmentContext.get()
	assert len(actual) == 2
	assert actual["engine-name"] == "pyiceberg"
	assert re.match(r"^\d+\.\d+\.\d+", actual["engine-version"])
	def test_default_value() -> None:
	assert EnvironmentContext.get() == {
	"engine-name": "pyiceberg",
	"engine-version": __version__,
	}
	def test_get_returns_copy() -> None:
	actual = EnvironmentContext.get()
	actual["test-key"] = "test-value"
	assert "test-key" not in EnvironmentContext.get()

EnvironmentContext might have been mutated, this is a better way to test that the returned value is a copy

[rambleraptor]

+1 on this. The regex is really brittle and won't work for RC builds.

[ebyhr]

Ugh, I was missing the RC build.

I intentionally avoided using the same logic between environment_context.py and this test. I generally don't believe it's a good idea to write tests.

[kevinjqliu]

Suggested change

	def test_put_and_remove() -> None:
	EnvironmentContext.put("test-key", "test-value")
	assert EnvironmentContext.get()["test-key"] == "test-value"
	EnvironmentContext.remove("test-key")
	assert "test-key" not in EnvironmentContext.get()
	def test_put_and_remove() -> None:
	try:
	EnvironmentContext.put("test-key", "test-value")
	assert EnvironmentContext.get()["test-key"] == "test-value"
	assert EnvironmentContext.remove("test-key") == "test-value"
	assert "test-key" not in EnvironmentContext.get()
	finally:
	EnvironmentContext.remove("test-key")

use try/finally to avoid mutating EnvironmentContext

[kevinjqliu]

Suggested change

	for key, value in EnvironmentContext.get().items():
	summary.__setitem__(key, value)
	for key, value in EnvironmentContext.get().items():
	summary[key] = value

nit

[kevinjqliu]

Suggested change

	"engine-version": version("pyiceberg"),
	"engine-version": __version__,

nit

[rambleraptor]

This looks great! My biggest note is to remove the version regexes, since they won't work on RCs.