Arrow IPC binary fetch path for DataFrame execution

[Martozar]

Summary

Adds a native Arrow IPC binary fetch path to gooddata-pandas, providing a faster alternative to the existing JSON-paged AFM path for large result sets.

What changed

gooddata-sdk — binary fetch

• BareExecutionResponse.read_result_arrow() fetches execution results from the server's binary IPC endpoint and returns a pyarrow.Table.

gooddata-pandas — Arrow→DataFrame conversion

• DataFrameFactory.for_exec_def_arrow() — new public method that mirrors for_exec_def() but uses the binary path.
• for_arrow_table() — pure conversion from pa.Table to (pd.DataFrame, DataFrameMetadata), enabling callers to bring their own Arrow data.
• convert_arrow_table_to_dataframe() — low-level converter that reconstructs row/column MultiIndex, subtotals, primary labels, and types from Arrow field metadata.

Why

The JSON paging path serialises every result to JSON and pages it in chunks — it is CPU-heavy and slow for wide or deep result sets. Arrow IPC transfers binary columnar
data in a single round-trip. End-to-end benchmarks against the GoodData demo workspace show 1.3×–33× speedup depending on table shape, with larger tables benefiting most .

Test coverage

• 140 unit tests covering: missing metadata keys (all three required keys), self_destruct mode, _build_field_index edge cases (subtotal padding, asymmetric depth), compute_row_totals_indexes with empty dimensions, for_arrow_table correctness across flat/transposed/subtotals/both-dim-totals cases.
• 47 ground-truth fixture cases generated against the live API and committed to tests/dataframe/fixtures/arrow/, including 3-metric tables, 3-level nested subtotals, multi-aggregation multi-metric tables, and asymmetric totals (different levels/aggregations per metric).
• IPC test fixture updated to use ipc.new_file to match the server format.

[codecov]

Codecov Report

❌ Patch coverage is 91.55405% with 25 lines in your changes missing coverage. Please review.
✅ Project coverage is 78.38%. Comparing base (49ea0d5) to head (7dddd3a).
⚠️ Report is 3 commits behind head on master.

Files with missing lines	Patch %	Lines
...s/gooddata-pandas/src/gooddata_pandas/dataframe.py	46.42%	15 Missing ⚠️
...ta-sdk/src/gooddata_sdk/compute/model/execution.py	76.47%	4 Missing ⚠️
...es/gooddata-pandas/src/gooddata_pandas/__init__.py	50.00%	3 Missing ⚠️
...data-pandas/src/gooddata_pandas/arrow_convertor.py	99.12%	2 Missing ⚠️
...ddata-sdk/src/gooddata_sdk/compute/model/filter.py	0.00%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1489      +/-   ##
==========================================
+ Coverage   78.13%   78.38%   +0.25%     
==========================================
  Files         228      230       +2     
  Lines       14926    15212     +286     
==========================================
+ Hits        11662    11924     +262     
- Misses       3264     3288      +24     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[hkad98]

Nitpick: this is a new dependency. Consider setting the threshold higher e.g., pyarrow>=23.0.1