Fix invalid integration comparisons and reference-generation failures

[cheerly-pku]

What's changed?

Integration comparisons can currently accept extra unmatched data lines or
non-finite values. Reference generation can also record a failed comparison as
an expected result, or report success after a collector failed. This change
rejects those invalid comparisons and makes reference generation fail with a
nonzero status while removing incomplete result.ref files.

The collectors preserve failed checks as status entries in ordinary
result.out output. Numerical tolerances and the existing Wannier90 entry-wise
comparison policy are unchanged. Trailing blank lines are ignored; missing or
shifted data and interior blank lines remain significant.

Linked Issue

No separate issue. Extracted from #8042 following review so changes affecting
all integration tests can be reviewed independently of PPCG. This PR contains
no solver implementation or reference-value changes.

Unit Tests and/or Case Tests for my changes

Commands rerun for merge commit 657582441 (merging develop at
8b546bdf3, without rebasing or force-pushing):

cd tests/integrate
PATH=/usr/bin:/bin OMP_NUM_THREADS=1 python3 -m unittest -v validation_tools/test_compare_file.py
PATH=/usr/bin:/bin OMP_NUM_THREADS=1 bash validation_tools/test_catch_properties.sh
cd ../..
PATH=/usr/bin:/bin OMP_NUM_THREADS=1 python3 -m unittest discover -s tests/integrate/validation_tools -p test_librpa_wavefunctions.py -v
bash -n tests/integrate/Autotest.sh tests/integrate/validation_tools/*.sh
python3 tools/03_code_analysis/agent_governance_check.py --base deepmodeling/develop --head HEAD --format text
git diff --check deepmodeling/develop...HEAD

Results: 7/7 comparison tests and 8/8 upstream LibRPA parser tests pass;
the expanded shell regression and syntax check exit 0; the PR-diff governance
check reports no findings; diff whitespace check passes. The shell tests
exercise missing outputs, failed comparisons, missing/malformed descriptors,
cleanup of partial references, and run_check.sh/Autotest.sh failure status.
System Python with NumPy was selected for the malformed-descriptor test.

Conflict resolution preserves upstream vxc.ref / OUT.autotest/vxc.txt
naming while retaining failed-comparison handling. New shell coverage checks
both Gamma-only and multi-k VXC success/mismatch/missing-output cases, plus
upstream's -p dedicated-validator mode, custom MPI launcher and pre/post
flags, OFML single-rank override, and commented/blank case-list entries.
No reference values, tolerances, or PPCG solver changes are included.

The actual tests/integrate/CMakeLists.txt was also exercised through a
temporary CMake driver (project(... NONE), enable_testing(),
ABACUS_TEST_DIR=<checkout>/tests, and add_subdirectory of integrate):

PATH=/usr/bin:/bin cmake -S /tmp/pr8138-conflict-check.5hibZt -B /tmp/pr8138-conflict-check.5hibZt/with-python -DPython3_EXECUTABLE=/usr/bin/python3
PATH=/usr/bin:/bin OMP_NUM_THREADS=1 ctest --test-dir /tmp/pr8138-conflict-check.5hibZt/with-python -V -R '^(compare_file_test|catch_properties_test)$'
PATH=/usr/bin:/bin cmake -S /tmp/pr8138-conflict-check.5hibZt -B /tmp/pr8138-conflict-check.5hibZt/no-python -DCMAKE_DISABLE_FIND_PACKAGE_Python3=TRUE
ctest --test-dir /tmp/pr8138-conflict-check.5hibZt/no-python -N

Both registered regression tests pass (CTest 2/2). Python-disabled configuration
also succeeds and explicitly skips those two tests, preserving upstream's
removal of the hard Python configure dependency. The ordinary integration-test
entry remains registered; it was not run by this temporary driver.

The full ABACUS integration suite was not rerun on this standalone branch;
these focused regressions exercise the validation scripts without an ABACUS
build. CI should additionally verify their behavior with the existing cases.

Governance Notes

• INPUT/docs changes: none; no INPUT parameter behavior changes.
• Core module impact: none; changes are confined to tests/integrate.
• Exceptions requested: none.

Reminder

• I have read AGENTS.md and docs/developers_guide/agent_governance.md.
• I have linked an issue or explained why this PR does not need one.
• I have added adequate unit tests and/or case tests, or explained why not.
• I have listed the exact verification commands run and their results.
• I have described user-visible behavior changes, including INPUT parameter changes.
• I have explained core-module impact for ESolver, HSolver, ElecState, Hamilt, Operator, Psi, or other source/ changes.
• I have requested any needed governance exception below.

[mohanchen]

some conflicts need to be solved. @cheerly-pku

[cheerly-pku]

some conflicts need to be solved. @cheerly-pku

Thanks for the reminder! The conflicts have been resolved. @mohanchen

[ABACUS-PR-BOT]

Static review of 6575824415cf3edf3a7c56e6d13be2a5b44799e5, with independent static verification. One issue is noted inline. Pinned governance and whitespace checks passed; builds and runtime tests were not run. CI diagnostics were unavailable through the selected account connector.

[ABACUS-PR-BOT]

[P2] Propagate collector failure before checking partial results

Please check the collector exit status in both normal and debug modes before calling check_out. For 101_PW_W90, a missing OUT.autotest/diamond.amn now makes bash -e stop after the energy entries; if those match, check_out visits only that partial result.out, skips AMN/MMN/EIG validation, and can finish successfully. Return nonzero on collection failure, as the reference branch does, and add ordinary/debug regression coverage.

[cheerly-pku]

Thanks for catching this. Fixed in cd55ddc: both normal and debug modes now exit nonzero if the collector fails, before calling check_out. Added regression coverage for missing Wannier90 output after a valid partial result.out is written. The regression catches the old behavior, and the updated shell regression, comparison tests, syntax, whitespace, and governance checks all pass.