refactor(tpc): add tpc-ds tests

[cpcloud]

First 27 TPC-DS queries running against DuckDB, Trino, Snowflake, and DataFusion, sans the ones that requires ROLLUP.

More fail on DataFusion than the others. Those are marked with appropriate xfail marker.

[cpcloud]

If this PR is too large I can split up the things into a separate PR for each backend, and disable the runs until the refactoring is done.

[gforsyth]

Nah, I'm most of the way through reviewing it

[gforsyth]

I think the changeset here is good -- do we want to keep the tpch and tpc-ds snapshots around? Since we're comparing results against hand-written SQL, I don't know that we're getting much out of them

[cpcloud]

They're gone.

[cpcloud]

The empty result sets are questionable, but I couldn't find a SF (up to sf=10) where all queries were non-empty. I believe that q17 is supposed to empty, based on the fact that it's always empty for the larger scale factors.

So, basically what I did was selectively allow empty queries in the tpc_test marker.

[gforsyth]

I believe that q17 is supposed to empty, based on the fact that it's always empty for the larger scale factors.

lol, I have an optimization for O(1) compute for query 17 at any scale factor

[gforsyth]

snowflake is passing:

🐚 pytest -m snowflake ibis/backends/tests/tpc/ds/test_queries.py -v
================================ test session starts ================================
platform linux -- Python 3.10.14, pytest-8.2.2, pluggy-1.5.0 -- /nix/store/1lj814h1vbfi6p7m6vr2rvcvkdqxc426-python3-3.10.14-env/bin/python3.10
cachedir: .pytest_cache
hypothesis profile 'dev' -> deadline=None, max_examples=50, suppress_health_check=[HealthCheck.too_slow], database=DirectoryBasedExampleDatabase(PosixPath('/home/gil/github.com/ibis-project/ibis/.hypothesis/examples'))
Using --randomly-seed=3164630463
benchmark: 4.0.0 (defaults: timer=time.perf_counter disable_gc=False min_rounds=5 min_time=0.000005 max_time=1.0 calibration_precision=10 warmup=False warmup_iterations=100000)
rootdir: /home/gil/github.com/ibis-project/ibis
configfile: pyproject.toml
plugins: hypothesis-6.104.2, snapshot-0.9.0, anyio-4.4.0, randomly-3.15.0, mock-3.14.0, benchmark-4.0.0, timeout-2.3.1, cov-5.0.0, repeat-0.9.3, clarity-1.0.1, pytest_httpserver-1.0.10, xdist-3.6.1
collected 540 items / 513 deselected / 27 selected                                  

ibis/backends/tests/tpc/ds/test_queries.py::test_15[snowflake] PASSED    [  3%]
ibis/backends/tests/tpc/ds/test_queries.py::test_04[snowflake] PASSED    [  7%]
ibis/backends/tests/tpc/ds/test_queries.py::test_07[snowflake] PASSED    [ 11%]
ibis/backends/tests/tpc/ds/test_queries.py::test_23[snowflake] XFAIL     [ 14%]
ibis/backends/tests/tpc/ds/test_queries.py::test_03[snowflake] PASSED    [ 18%]
ibis/backends/tests/tpc/ds/test_queries.py::test_20[snowflake] PASSED    [ 22%]
ibis/backends/tests/tpc/ds/test_queries.py::test_24[snowflake] PASSED    [ 25%]
ibis/backends/tests/tpc/ds/test_queries.py::test_18[snowflake] XFAIL     [ 29%]
ibis/backends/tests/tpc/ds/test_queries.py::test_10[snowflake] PASSED    [ 33%]
ibis/backends/tests/tpc/ds/test_queries.py::test_16[snowflake] PASSED    [ 37%]
ibis/backends/tests/tpc/ds/test_queries.py::test_26[snowflake] PASSED    [ 40%]
ibis/backends/tests/tpc/ds/test_queries.py::test_05[snowflake] XFAIL     [ 44%]
ibis/backends/tests/tpc/ds/test_queries.py::test_17[snowflake] PASSED    [ 48%]
ibis/backends/tests/tpc/ds/test_queries.py::test_02[snowflake] PASSED    [ 51%]
ibis/backends/tests/tpc/ds/test_queries.py::test_08[snowflake] PASSED    [ 55%]
ibis/backends/tests/tpc/ds/test_queries.py::test_12[snowflake] PASSED    [ 59%]
ibis/backends/tests/tpc/ds/test_queries.py::test_01[snowflake] PASSED    [ 62%]
ibis/backends/tests/tpc/ds/test_queries.py::test_25[snowflake] PASSED    [ 66%]
ibis/backends/tests/tpc/ds/test_queries.py::test_27[snowflake] PASSED    [ 70%]
ibis/backends/tests/tpc/ds/test_queries.py::test_21[snowflake] PASSED    [ 74%]
ibis/backends/tests/tpc/ds/test_queries.py::test_13[snowflake] PASSED    [ 77%]
ibis/backends/tests/tpc/ds/test_queries.py::test_06[snowflake] PASSED    [ 81%]
ibis/backends/tests/tpc/ds/test_queries.py::test_22[snowflake] XFAIL     [ 85%]
ibis/backends/tests/tpc/ds/test_queries.py::test_09[snowflake] PASSED    [ 88%]
ibis/backends/tests/tpc/ds/test_queries.py::test_19[snowflake] PASSED    [ 92%]
ibis/backends/tests/tpc/ds/test_queries.py::test_11[snowflake] PASSED    [ 96%]
ibis/backends/tests/tpc/ds/test_queries.py::test_14[snowflake] XFAIL     [100%]

============= 22 passed, 513 deselected, 5 xfailed in 124.88s (0:02:04) =============

[gforsyth]

Boy do I not understand waht asceding is about, but it's clearly intentional.

This looks good, and makes it easy for other contributors to tackle individual tpc-ds queries.