style: fix two wrong @file tags and convert leading tab indentation to spaces

[Critsium-xy]

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.

Linked Issue

No issue; this is hygiene found by running
tools/03_code_analysis/code_quality_score.py over the whole source/ tree.
Happy to open a tracking issue if maintainers prefer one.

Unit Tests and/or Case Tests for my changes

No test is added: no behaviour changes. Apart from two comment lines, the whole
diff is whitespace. What matters is the evidence that nothing
else moved.

Commands run

# the scan that produced the two findings
python3 tools/03_code_analysis/code_quality_score.py source/ --format json -o cq.json

# the conversion
expand -i -t4 <file>          # for each of the 286 selected files

# proof that only whitespace moved
git diff -w --stat            # -> empty

# compile check (Linux, 15 cores, Libxc 7.1.2, LibRI 2.1.1)
cmake -B build -G Ninja -DBUILD_TESTING=OFF -DENABLE_LCAO=ON -DENABLE_MPI=ON \
      -DENABLE_OPENMP=ON -DENABLE_LIBXC=ON -DENABLE_LIBRI=ON
cmake --build build -j15 -- -k 0

python3 tools/03_code_analysis/agent_governance_check.py --base upstream/develop --head HEAD

Result summary

• git diff -w --stat is empty: not one non-whitespace character changed
  anywhere in the 246-file whitespace commit.
• Of the 42 raw string literals in the changed files (222 lines, all in
  source_io/module_parameter/read_inp_out.cpp), none has a line this PR
  touches. Leading whitespace inside R"(...)" is part of the string, so that
  was the one place where rewriting leading whitespace could have changed
  behaviour.
• No changed line follows a line ending in a backslash, so no
  backslash-continued string literal is affected either.
• No file gained a CR.
• Compile: 817 of 820 targets built. The 3 that failed
  (module_bse/hamilt_bse.cpp, esolver_factory.cpp,
  esolver_lr_lcao_bse.cpp) all stop at
  fatal error: RI/physics/LR.h: No such file or directory. That header is
  absent from the LibRI 2.1.1 copy on this machine, which ships only
  Exx.h/GW.h/RPA.h. The same three targets fail identically on the
  unpatched baseline, which I confirmed by reverting source/ and rebuilding
  them — so this is an environment gap, not a regression from this PR.
• Governance checker: no blockers; two warnings, both answered below.

Checks not run, with reason

• The Check documentation consistency step (regenerate docs/parameters.yaml
  and diff) was not run locally: the local build could not link an executable
  because of the LibRI gap above. This PR touches
  source_io/module_parameter/read_inp_out.cpp, but only outside its raw string
  literals, so the generated docs cannot change. CI's own doc-consistency step is
  the authoritative check and will confirm it.
• No integration or unit test run: nothing executable changed.

What's changed?

1. Two @file tags that name a file which does not exist. Doxygen then
attributes the block to the wrong file, or to none.

file	@file said
source_base/ndarray.h	NDArray.h
source_lcao/module_rt/band_energy.h	bandenegy.h (also a typo)

The scorer's doc_file_mismatch rule originally flagged a third file,
source_pw/module_pwdft/radial_proj.h (said radial_projection.h). #8007 has
since moved that header to source_pw/module_proj/radial_proj.h with the tag
already corrected, so it was dropped from this PR on rebase.

2. Leading tab indentation converted to spaces in 246 files.
.editorconfig has required indent_style = space with indent_size = 4 for
the whole repository, yet 339 files under source/ still indent with tabs
(18,846 lines). expand -i -t4 rewrites the initial whitespace of a line and
nothing else.

Three groups were deliberately left out:

• 48 files touched by the PRs open when this was prepared (Refactor density matrix module. #8000, feat(xc): add gga_grad 1/2 with variationally consistent mode 2 #7924,
  Refactor occupy.h/cpp and update MPI kmesh_world #7906, Add CUDA NEP backend and GPU neighbor lists #8005 and others), so this does not force a conflict on work in flight.
  Of the 286 files selected, 253 actually had leading tabs; the rest had only
  alignment tabs. Since then develop has deleted source_lcao/setup_dm.h
  (Refactor density matrix module. #8000) and module_stodft/hamilt_sdft_pw.cpp (Move HSolverPW_SDFT into module_stodft as StoHSolverPW; rename HamiltSdftPW to StoHamiltPW #8012), and already converted
  module_out/binstream.{h,cpp} (fix: guard Binstream write path against unopened files and failed fwrite #8025), module_stodft/sto_hamilt_pw.h
  (Move HSolverPW_SDFT into module_stodft as StoHSolverPW; rename HamiltSdftPW to StoHamiltPW #8012), module_ctrl/ctrl_output_pw.h and module_pwdft/op_pw_nl.cpp
  (Feature: Plane-wave RT-TDDFT with CPU/GPU and mixed-precision support #8043), leaving 246. Files that develop moved are converted at their new
  path, e.g. module_proj/onsite_proj_tools_stress.cpp (Refactor projection in pw basis #8007).
• 5 vendored files. source_base/libm/ is ported from glibc-2.36 and carries
  its own LICENCE; source_base/mcd.c is Softpixel MemCheckDeluxe under a BSD
  licence. Reformatting vendored sources makes future syncs with their upstream
  harder.
• Alignment tabs after the first non-blank character (1,193 lines). Only
  leading indentation is converted.

Effect on code_quality_score.py over source/: average 82.15 -> 82.58,
passing files 1560 -> 1562, 204 files up and 6 down.

The 6 that go down lose 1 to 3 points, all through line_too_long, because the
scorer counts a tab as one character while it renders as up to four columns.
Those lines were already past 120 columns on screen; the tab was hiding it. One
of them, module_ri/exx_abfs_ctor_orbs.cpp, moves from 60 to 59 and so slips
just under the tool's pass line. Wrapping those lines would mean editing code in
a commit that is otherwise whitespace-only, so I left it for a follow-up rather
than muddy the diff.

Reviewers: git diff -w on this branch is empty, so reviewing with
git diff -w or GitHub's "Hide whitespace changes" shows only the two
comment lines from the first commit.

Rebase notes

The branch has been rebased onto develop four times to resolve conflicts with PRs
merged in the meantime (#8000, #8007, #8012, #8023, #8025, #8026, #7924, #8043 and others). Because the
second commit is whitespace-only, conflicts were not merged by hand: each
conflicting file was taken from develop as-is and expand -i -t4 was run on
it again. After the latest rebase (base 43a4a2afb):

• git diff -w --stat upstream/develop lists only ndarray.h and
  band_energy.h (one comment line each);
• none of the 248 changed files (246 whitespace + 2 comment) has a leading tab
  left; git ls-files --eol reports all of them as i/lf w/lf, and no added
  line ends in CR;
• read_inp_out.cpp (the only changed file with raw string literals) was
  changed on develop again (Update out_stru parameter and its output context #8026, feat(xc): add gga_grad 1/2 with variationally consistent mode 2 #7924), so the raw-string check was re-run
  on the rebased file: 42 literals covering 222 lines, 47 lines changed by this PR,
  overlap none; no changed line follows a backslash-continued line;
• PYTHONUTF8=1 python tools/03_code_analysis/agent_governance_check.py --base upstream/develop --head HEAD --format text
  exits 0 with no errors; its warnings are the INPUT-docs warning answered
  below plus "new default parameters" warnings on existing declarations
  (e.g. blas_connector.h) whose lines were only re-indented.

The full compile check above was run before the rebases; CI builds the rebased
branch.

Governance Notes

• INPUT/docs changes: none. The governance checker raises an INPUT
  documentation warning because source_io/module_parameter/read_inp_out.cpp is
  in the diff, but no Input_Item registration, annotation, description,
  default_value or availability changes. The parameter descriptions in that
  file live inside R"(...)" raw string literals, and none of the 222 lines
  covered by those literals is touched by this PR, so docs/parameters.yaml and
  docs/advanced/input_files/input-main.md need no regeneration.
• Core module impact: none. No declaration, definition, signature or control
  flow changes anywhere. ESolver, HSolver, ElecState, Hamilt, Operator and Psi
  are affected only by the indentation of lines in files they own.
• Exceptions requested: none.

🤖 Generated with Claude Code