Test triage: XPASS/XFAIL/skip summary on feat/where-clause-executor @062d8a0c

[lmeyerov]

Context

• Branch: feat/where-clause-executor
• Commit: 062d8a0c
• Command:
  • .venv/bin/python -B -m pytest -q -ra --ignore=graphistry/tests/test_bolt_util.py --ignore=graphistry/tests/test_gremlin.py --ignore=graphistry/tests/test_ipython.py --ignore=graphistry/tests/test_nodexl.py --ignore=graphistry/tests/test_tigergraph.py --ignore=graphistry/tests/test_feature_utils.py --ignore=graphistry/tests/test_umap_utils.py --ignore=graphistry/tests/test_dgl_utils.py --ignore=graphistry/tests/test_embed_utils.py --ignore=graphistry/tests/test_kusto.py --ignore=graphistry/tests/test_spanner.py
• Python: 3.12.11

Results summary

• 1614 passed, 337 skipped, 9 xfailed, 4 xpassed

XPASS (previously expected to fail)

All are labeled Issue #786 in graphistry/tests/compute/test_astcall_chains.py:

• test_two_edge_filters_sequential
• test_edge_then_node_filter
• test_filter_then_hop_forward
• test_contradictory_filters

XFAIL (still expected to fail)

• Issue GFQL: Chained filter_edges_by_dict operations only apply first filter #786 still failing:
  • test_three_edge_filters_sequential
  • test_two_node_filters_sequential
  • test_chained_filters_with_predicates
  • test_filter_order_should_not_matter
• Missing optional deps (xfail on missing module):
  • graphistry/tests/test_plotter.py::TestPlotterConversions::test_igraph2pandas
  • graphistry/tests/test_plotter.py::TestPlotterConversions::test_pandas2igraph
  • graphistry/tests/test_plotter.py::TestPlotterConversions::test_networkx2igraph
  • graphistry/tests/test_plotter.py::TestPlotterArrowConversions::test_api3_plot_from_igraph
• Oracle limitation:
  • tests/gfql/ref/test_df_executor_core.py::TestOracleLimitations::test_edge_alias_on_multihop

Skips (337)

Mostly expected due to missing optional deps/flags:

• TEST_CUDF=1 cuDF-only, GPU availability, TEST_DASK=1, TEST_DASK_CUDF=1, TEST_CUGRAPH=1.
• Extras not installed: umap/ai, igraph, cugraph.

Skips that look like TODOs / known gaps:

• graphistry/tests/compute/test_id_column_restriction.py:426 (let/ref returns empty result - needs investigation)
• graphistry/tests/compute/test_chain_let.py:834 (Runtime KeyError in ASTRef chain execution with query parameter)
• graphistry/tests/compute/test_id_column_restriction.py:657 (collapse() has complex API - needs topology-aware traversal setup)
• graphistry/tests/compute/test_call_schema_validation.py (schema effects not yet implemented in safelist)
• graphistry/tests/compute/test_astcall_chains.py:176 (Enrichment ops not yet in safelist)

Ask

• Should we convert the XPASS entries into regular passing tests (remove xfail) or keep until Issue GFQL: Chained filter_edges_by_dict operations only apply first filter #786 fully resolved?
• Triage the TODO-skipped cases above: root cause, scope, and whether they should be fixed or reclassified.

[lmeyerov]

Triaged the Issue #786 xfail/xpass set on master: failing cases were due to invalid predicate syntax (raw {'gte': ...} dicts in ASTCall filters) or incorrect predicate imports (Gt/Lt). I updated to use supported predicates (, , /) and removed the outdated xfail markers. Branch: , commit . Targeted tests now pass on that branch. Happy to open a PR if desired.

[lmeyerov]

Follow-up (previous comment was mangled by shell expansion):

Triaged the Issue #786 xfail/xpass set on master: the failing cases were due to invalid predicate syntax (raw {'gte': ...} dicts in ASTCall filters) or incorrect predicate imports (Gt/Lt). I updated graphistry/tests/compute/test_astcall_chains.py to use supported predicates (ge, Between, gt/lt) and removed the outdated xfail markers. Branch: triage/issue-893, commit d416cec0. Targeted tests now pass on that branch.

[lmeyerov]

Update on triage/issue-893:

• Fixed ASTRef traversal to seed initial wavefront from binding nodes while traversing on original graph edges. Commit 0957e30.
• Removed skip in TestProblematicColumnNames.test_let_ref_with_problematic_columns; now passes.
• Tests: graphistry/tests/compute/test_chain_let.py (65 passed), plus targeted tests for let/ref + mixed matchers (3 passed).

[lmeyerov]

Additional fix: registered pytest mark 'tier2' in pytest.ini to silence PytestUnknownMarkWarning (commit 6bd49d6).

[lmeyerov]

Update: enabled ASTCall schema validation via safelist schema_effects + _validate_call_op in compute/validate/validate_schema.py; unskipped call schema validation tests (7 passed). Commit 71db84f. Note: running pytest under uv requires PYTHONPATH=. (used: ============================= test session starts ==============================
platform linux -- Python 3.12.11, pytest-9.0.1, pluggy-1.6.0
rootdir: /home/lmeyerov/Work/pygraphistry
configfile: pytest.ini
plugins: xdist-3.8.0, hypothesis-6.148.8
collected 0 items

============================ no tests ran in 0.09s =============================).

[lmeyerov]

Update: enabled ASTCall schema validation via safelist schema_effects + _validate_call_op in compute/validate/validate_schema.py; unskipped call schema validation tests (7 passed). Commit 71db84f. Note: uv run needs PYTHONPATH=. for local imports (used: uv run env PYTHONPATH=. pytest graphistry/tests/compute/test_call_schema_validation.py -q).

[lmeyerov]

Update: unskipped collapse() problematic-node test with a minimal topology; passes on triage/issue-893 (commit 2b72cb9). Ran: uv run env PYTHONPATH=. pytest graphistry/tests/compute/test_id_column_restriction.py -q (47 passed, 1 skipped). Warnings are DeprecationWarnings from chain() plus a pandas FutureWarning in chain labeling.

[lmeyerov]

Update: removed pandas FutureWarning on label bool fillna in chain.combine_steps by casting to pandas BooleanDtype before fill (commit 6292db7). Re-ran test_id_column_restriction.py: warnings now just DeprecationWarning for chain() usage.

[lmeyerov]

Update: removed all xfails tied to cudf issue #7735 in graphistry/tests/test_hyper_dask.py (commit e70f871). Not validated locally since TEST_CUDF/TEST_DASK_CUDF not set.

[lmeyerov]

Update: removed neo4j driver #789 xfail in graphistry/tests/test_bolt_util.py (commit e442886). Not validated locally since WITH_NEO4J/neo4j env not set.

[lmeyerov]

Update: migrated compute tests from deprecated chain() to gfql() to eliminate DeprecationWarnings. Commit 1b477a7. Example run: uv run env PYTHONPATH=. pytest graphistry/tests/compute/test_id_column_restriction.py -q (no warnings).

[lmeyerov]

Update: removed DeprecationWarnings from boundary-call tests by using internal _chain_impl in boundary handling (commit 5700d78). Re-ran: uv run env PYTHONPATH=. pytest graphistry/tests/compute/test_call_boundaries.py graphistry/tests/compute/test_call_engine_coercion.py graphistry/tests/compute/test_chain_concat.py graphistry/tests/compute/test_id_column_restriction.py graphistry/tests/compute/test_identifier_conflicts.py -q (70 passed, 33 skipped, no warnings).