assembling_pseudo.x

Description

assembling_pseudo is a TurboRVB utility that assembles per-species pseudopotential (ECP) files into a single file, pseudo.dat, used by TurboRVB VMC/DMC and tools such as readforward.

  • Input: fort.10 (ion list and valence/core info) and one pseudopotential file per atomic species (e.g. Filippi format), placed in a fixed directory.

  • Output: pseudo.dat in ECP format, read by read_pseudo in the main code and by other tools.

  • For systems with multiple species, you place each species’ pseudopotential file under the pseudo/ directory, run the program from a directory containing fort.10, and enter the file extension (e.g. filippi) when prompted. The program writes one pseudo.dat ordered by ion index in fort.10.

Run with --help, -help, or help to show online help (help_online('assembling_pseudo')).

Input and output

Input

  • fort.10 (required): Wave function file. Read via read_fort10 to get nion, zetar (number of valence electrons for the pseudo), atom_number (atomic number), rion (ion coordinates), iespbc (periodic boundary flag). Must exist in the current working directory.

  • Pseudopotential files: One file per species. Search directory is fixed at build time as NAME_DIR/pseudo/ (NAME_DIR is replaced by ${CMAKE_SOURCE_DIR}). File name must be Zzetar_atomnumberatom_number.extension (e.g. Ne with 2 valence electrons: Z10_atomnumber2.filippi). zetar and atom_number are taken from fort.10 for each ion.

  • Standard input: One line giving the extension (e.g. filippi). This is the suffix used for pseudopotential file names. You can use the template template/assembling_pseudo.input as reference.

Which ions are pseudo: For ion index j, if zetar(j) ≠ atom_number(j) and atom_number(j) > 0, that ion is treated as a pseudopotential ion and its corresponding file is read and appended to pseudo.dat. Ions with zetar(j) = atom_number(j) or atom_number(j) ≤ 0 are skipped.

Output

  • pseudo.dat: ECP-format pseudopotential data. First line is ECP. Then, for each pseudo ion: ion index j, cutoff, number of shells, shell lengths (Gaussians per shell), and parameter lines (three reals per line). This file is overwritten if it already exists.

Special case — extension ``distance``: No pseudopotential files are opened. The program prints pairwise distances between all ions (with PBC applied if iespbc is true) to stdout. Useful to get distances only; for assembling pseudos, use the real extension (e.g. filippi).

Notes

Pseudopotential file format

Each per-species input file (e.g. Z10_atomnumber2.filippi) must have the following structure:

  1. Line 1: Pseudopotential name (string). Only the first file’s name is read; the assembled pseudo.dat always starts with ECP.

  2. Line 2: Three values: fake_num (integer), ps_cut (real), num_sh (integer). Cutoff and number of angular momentum shells.

  3. Line 3: num_sh integers: shwork(1:num_sh) — number of Gaussians in each shell.

  4. Next lines: Exactly sum(shwork) lines. Each line has three reals: ps_par1, ps_par2, ps_par3 (e.g. Gaussian exponents/coefficients).

These are written to pseudo.dat as: ion index j, ps_cut, num_sh, the shwork integers, then the parameter lines. TurboRVB’s read_pseudo expects this ECP layout (kindion, rcutoff, pshell, shell lengths, parameters).

Troubleshooting

Fatal errors (program stops)

  • Cannot open fort.10 — fort.10 is missing in the current directory. Put the wave function file in the run directory or run from the directory that contains fort.10.

  • Error while reading fort.10 — fort.10 is corrupted or not in TurboRVB format. Regenerate it with makefort10 or convertfort10 and check the reported line/field.

  • Error while reading a pseudopotential file — The file has too few lines, wrong number of values per line, or wrong type (e.g. real where an integer is expected). Check that the file matches the format above: line 1 (name), line 2 (three values), line 3 (num_sh integers), then sum(shwork) lines of three reals each.

Pseudopotential file not found (message printed, run continues)

If the extension is not distance and a pseudopotential file cannot be opened, the program prints:

##################################
Error opening the pseudo potential file
File   /path/to/pseudo/Z10_atomnumber2.filippi not found
Program ends
##################################

It does not stop: that ion is skipped and the loop continues. You may end up with an incomplete pseudo.dat (missing some species) if you overlook this.

Causes and fixes:

  • Wrong directory — Files are looked up under NAME_DIR/pseudo/ (build-time path, usually the TurboRVB source tree’s pseudo/). Put the files there, or change the source (name_dir) / use a symlink for pseudo/.

  • Wrong file name — Name must be Zzetar_atomnumberatom_number.extension. Example: Z10_atomnumber2.filippi for Ne with 2 valence electrons. Check spelling, case, and that the extension matches what you type on stdin.

  • Wrong extension on stdin — The extension you type must match the file suffix (e.g. type filippi if files are *.filippi).

  • Wrong zetar/atom_number in fort.10 — For pseudo ions, zetar(j) ≠ atom_number(j) and atom_number(j) > 0. Verify with makefort10/convertfort10 that valence and atomic numbers are correct.

Tip: Check the printed dir_pseudo = ... and Extension read = ..., then list that directory to confirm the expected Z<zetar>_atomnumber<atom_number>.<extension> files exist.

Extension distance

When you enter distance as the extension, no pseudopotential files are opened; only pairwise distances are printed. Use a real extension (e.g. filippi) when you want to assemble pseudo.dat.

Other notes

  • Atomic number or zetar ≥ 1000: The code builds file names with at most 3 digits for Z and atom_number. For Z or atom_number ≥ 1000, the file name may be wrong and the file will not be found.

  • pseudo.dat already exists: It is opened with status='unknown' and overwritten. Back it up first if needed.