Fix mismatch in ibz_kpoints and add symmetry info in intergrate tests

[maki49]

fix #2472, and the symmetry-part error in #2474

The inaccuracy of STRU file sometimes makes the result of optimized-lattice-search mismatch between real and reciprocal space. That’s why I used reciprocal relation to force-match it in #2441 . But it destroyed the optimization and then cause the mismatch in other cases (like 002-4MoS2, 003-12Pt, 006-Ti15).

So in this PR I try to search the optimized lattice again after using reciprocal relation.
It can repair both of the previous-failed cases in my local test.

For details, taking case 006-Ti15 as an example:

reciprocal-opt-latt from a direct search: (before #2441 ) (some cell-relax-generated cases failed)
0.170254 0.0982965 0
0 -0.196593 0
0 0 -0.107643
reciprocal-opt-latt from a direct use of reciprocal relation: (after #2441 ) (002-4MoS2, 003-12Pt, 006-Ti15 failed)
0.170254 -0.0982962 0
0 -0.196593 0
0 -0 -0.107643
reciprocal-opt-latt from searching optimized lattice again after using reciprocal relation: (now) (both are ok)
0.170254 0.0982965 0
-0.170254 0.0982962 -0
0 0 0.107643

[pxlxingliang]

@maki49 Please add this example to the integration test and modify the settings to make the time cost less than 30s or less

[maki49]

The test failed because other symmetry-on cases' result.ref files have no symmetry infomation.

I will add add symmetry info in all the result.refs in those cases.