fix(abacus): preserve scalar spin-constraint separators (deepmodeling#1022)

Fixes deepmodeling#998.

Emit an explicit `sc 0` token for scalar false constraints and update
the golden fixture that encoded the malformed output.

Tests: `cd tests && python -m unittest test_abacus_stru_dump`

Why existing tests missed it: The golden file was generated from the
buggy output, so the old test institutionalized the defect.

Coding agent: Codex
Codex version: codex-cli 0.144.4
Model: gpt-5.6-sol
Reasoning effort: xhigh

Co-authored-by: njzjz-bot <njzjz.bot@gmail.com>

Author: njzjz-bot

dpdata/formats/abacus/stru.py

@@ -802,7 +802,9 @@ def process_file_input(file_input, atom_names, input_name):
                         ["1" if ii else "0" for ii in sc[iatomtype]]
                     )
                 elif isinstance(ndarray2list(sc[iatomtype]), (int, float, bool)):
-                    iout += " sc " + "1" if sc[iatomtype] else "0"
+                    # Parenthesize the conditional expression so both True
+                    # and False values retain the ``sc`` field separator.
+                    iout += " sc " + ("1" if sc[iatomtype] else "0")
 
             if lambda_ is not None:
                 if isinstance(ndarray2list(lambda_[iatomtype]), (list, tuple)) and len(

tests/abacus.scf/stru.ref

@@ -25,5 +25,5 @@ H
 4
 5.416431453540 4.011298860305 3.511161492417 1 1 1 mag 1.000000000000 1.000000000000 1.000000000000 angle2 90.000000000000 sc 1 lambda 0.400000000000 0.500000000000 0.600000000000
 4.131588222365 4.706745191323 4.431136645083 1 1 1 mag 1.000000000000 angle1 100.000000000000 angle2 80.000000000000 sc 1 0 1 lambda 0.700000000000 0.800000000000 0.900000000000
-5.630930319126 5.521640894956 4.450356541303 0 0 0 mag 1.000000000000 angle1 90.000000000000 angle2 70.0000000000000
+5.630930319126 5.521640894956 4.450356541303 0 0 0 mag 1.000000000000 angle1 90.000000000000 angle2 70.000000000000 sc 0
 5.499851012568 4.003388899277 5.342621842622 0 0 0 mag 1.000000000000 angle1 80.000000000000 sc 1