refactor: clean up source_cell/source_estate/source_io reverse dependencies

[Critsium-xy]

Reminder

• Have you linked an issue with this pull request?
• Have you added adequate unit tests and/or case tests for your pull request?
• Have you noticed possible changes of behavior below or in the linked issue?
• Have you explained the changes of codes in core modules of ESolver, HSolver, ElecState, Hamilt, Operator or Psi? (ignore if not applicable)

Linked Issue

N/A — standalone refactor of the module dependency graph.

What's changed?

While auditing the cross-module #include graph under source/source_*, I found several cases where a lower-level module depended upward on a higher-level one, creating unnecessary reverse edges in the module dependency DAG. This PR removes three of them. Each change is an isolated, mechanical move with no behavioral / numerical change — only file locations, #include paths, and CMake source lists are touched. The relevant code is already covered by existing unit tests.

1. Remove a dead source_lcao include from source_cell
source_cell/read_atom_species.cpp carried a dead block:

#ifdef __EXX
#include "source_lcao/module_ri/serialization_cereal.h"
#endif

Nothing in the file uses it — it was a spurious cell → lcao reverse edge. Removed it and added the #include <sstream> the file actually relies on (std::stringstream, previously only available transitively).

bcast_cell.cpp still includes that header legitimately (it uses bcast_data_cereal, defined there), so that include is left untouched.

2. Move Magnetism from source_estate → source_cell
source/{source_estate => source_cell}/magnetism.{h,cpp} (100% rename). The Magnetism class only depends on source_base and is held by value as a member of UnitCell (unitcell.h). Living in source_estate forced a cell → estate reverse edge. Repointed all includers and updated the CMake/test source lists.

3. Move the output print helpers from source_io/module_output → source_base
source/{source_io/module_output => source_base}/output.{h,cpp} (rename). class output is a pure formatter over ModuleBase types (realArray, matrix, matrix3, ComplexMatrix) with no dependency above source_base. Sitting in source_io forced several cell/pw/psi → io reverse edges. Repointed all includers, fixed the header's own internal includes, and updated the main + test CMake source lists.

Test build adjustments (consequence of moving output into base):
Because output.cpp is now part of the base OBJECT library that nearly every test links, two stale patterns had to be cleaned up to avoid duplicate-symbol errors:

• Removed redundant explicit output.cpp entries from base-linked test SOURCES lists (kept where a test hand-compiles source_base/*.cpp and does not link base, e.g. source_md/test).
• Removed now-conflicting no-op output::printM3 / printrm mock stubs from a few tests (symmetry_test_*, to_qo_test, read_atoms_helper_test, psi_initializer_unit_test); the real base implementation now provides those symbols. Each test's still-needed Magnetism / InfoNonlocal mocks were left intact.

Any changes of core modules? (ignore if not applicable)

No functional change to ElecState/Hamilt/Psi/etc. Magnetism (used by ElecState and UnitCell) and the output printer are physically relocated to lower-level modules, but their interfaces and behavior are unchanged; all call sites only have their #include paths updated.

Verification

• Full build with -DBUILD_TESTING=ON (gcc/OpenMPI, LCAO enabled): 0 errors, all targets + unit-test executables link, including the full abacus_* binary.
• Verified no residual references to the old source_estate/magnetism.h or source_io/module_output/output.{h,cpp} paths.
• The legacy Makefile build is unaffected: Makefile.Objects uses VPATH, so object names resolve regardless of which source_* directory a file lives in.

🤖 Generated with Claude Code

[mohanchen]

Great! LGTM.