Datasets:
Formats:
csv
Languages:
English
Size:
1K - 10K
Tags:
open-science
superconductivity
superconductor-discovery
strongly-correlated
materials-discovery
2d-materials
License:
Commit History
-
pipeline/downfold.py: treat band occupations as per-spin fractions (n = 2·o) and select the band nearest half filling — fixes #15
Verified signatureFollow-up to #15 with a concrete patch and a reproducible before/after table. **What changes (`pipeline/downfold.py`, `active_band()`), two lines:** - `n_band = 2 * o` instead of `o` — in a non-spin-polarized pw.x run the per-band `<occupations>` are per-spin fractions in [0, 1]; the spin factor sits in the k-point weights, which this function renormalizes to 1. (`dos_at_ef()` in the same file already multiplies by 2 for this reason.) - select the crossing band by `abs(2*o − 1.0)` (nearest half filling) instead of `abs(o − 1.0)`, which in [0, 1] picks the fullest band. `build_final.py` needs no change: its particle–hole map (`Ne > 36 → 72 − Ne`) starts working once `n_band` can exceed 1. **Two separable effects.** 1. *Density interpretation* — reproducible from the published table alone. `recompute_filling.py` reads the existing `n_band` column as f, sets Ne = round(72 f) with the particle–hole map, and re-applies the published A_d(δ). Output: `active_challenge_corrected…027154a -
validation: multi-size ED table (2x4..4x4, U8/U0) by Ku-ku — absolute values unconverged, sign/rank only
Verified signaturecae607f
-
accuracy: ratio wording + uncorrelated note
Verified signaturec667a3d -
add ED-vs-VMC 4x4 validation table (ED by Ku-ku)
Verified signaturec056d76
-
publish 4x4 d-wave vs Fermi-sea control
Verified signature0911bd6