Goals

[ALF-Import-Bot]

In GitLab by @fassaad on May 4, 2020, 06:16 UTC:

Testing and running ALF.

The aim is to provide a user friendly way of Running ALF.

Run.py contains a complete default set of parameters for the 4X4 finite T Hubbard model at U/t=4.

Sims will overwrite these parameters. For example the line
Model = Hubbard, Lattice_type = Honeycomb, L1 = 3 , L2 = 3, ham_T2 = 0.0
will create the directory Hubbard_Honeycomb_L1=3_L2=3_T2=0.0/ as well as the parameter file, seeds and legacy out_to_in.sh and analysis.sh.

Here is the planned usage of Run.py

Usage: python3 Run.py Type= Alfdir= Runbranch= Program= Testbranch= Config=

Type : R+A, R or A, T with R=Run, A=Analyse, T=Test. Default=R+A
Alfdir : Path to ALF directory. Mandatory
Runbranch : Will run git checkout "Runbranch". Default=master
Executable: Name of executable Default="Model".out
Config : Will run ./configureHPC "Config" Default=Intel
Testbranch: Branch to test against Runbranch Default=master

Migrated from GitLab: https://git.physik.uni-wuerzburg.de/ALF/pyALF/-/issues/1

[ALF-Import-Bot]

In GitLab by @jonasschwab on May 11, 2020, 10:38 UTC:

I've restructured Run.py somewhat. It is no more dependant on out_to_in.sh and analysis.sh, a line of Sims is now in JSON format, meaning that
Model = Hubbard, Lattice_type = Honeycomb, L1 = 3 , L2 = 3, ham_T2 = 0.0
now is
{"Model": "Hubbard", "Lattice_type": "Honeycomb", "L1": 3 , "L2": 3, "ham_T2": 0.0}
and arguments get parsed by argparse, which also supplies a nice help message:

$ ./Run.py -h
usage: Run.py [-h] --alfdir ALFDIR [--type TYPE] [--branch_R BRANCH_R]
              [--branch_T BRANCH_T] [--config CONFIG]
              [--executable_R EXECUTABLE_R] [--executable_T EXECUTABLE_T]

Helper script for compiling, running and testing ALF.

optional arguments:
  -h, --help            show this help message and exit
  --alfdir ALFDIR       Path to ALF directory
  --type TYPE           Options R+T, R, T with R=Run, T=Test. (default: R)
  --branch_R BRANCH_R   Git branch to checkout for run. (default: master)
  --branch_T BRANCH_T   Branch to test against Runbranch. (default: master)
  --config CONFIG       Will run ./configureHPC.sh CONFIG (default: Intel)
  --executable_R EXECUTABLE_R
                        Name of ref executable. (default: <Model>.out)
  --executable_T EXECUTABLE_T
                        Name of test executable. (default: <Model>.out)

[ALF-Import-Bot]

In GitLab by @fassaad on May 12, 2020, 10:27 UTC:

Hi Jonas, looks good. As mentioned in the Wiki, for the type we would need to add A just for analysis, and C, just to create the run directory.