aiida-vibroscopy

[Rijan Karkee]

Dear Developers,

I am a new user to AiiDA.  I want to use aiida-vibroscopy. I have never been to live workshop and I have been following tutorial available on website. 

I have installed it in HPC and verdi status looks like, I hope this is ready to use.

(aiidaENV) rkarkee@ch-fe1:~> verdi status

 ✔ version:     AiiDA v2.5.1

 ✔ config:      /users/rkarkee/.aiida

 ✔ profile:     rkarkee

 ✔ storage:     Storage for 'rkarkee' [open] @ postgresql://aiida_qs_rkarkee_20554bcc4bead70a3479c4ef8d5f1f4e:***@localhost:5432/rkarkee_rkarkee_20554bcc4bead70a3479c4ef8d5f1f4e / DiskObjectStoreRepository: 7ef673c3acf444d2ac6ff9f6c839a15f | /users/rkarkee/.aiida/repository/rkarkee/container

 ✔ rabbitmq:    Connected to RabbitMQ v3.8.14 as amqp://guest:gu...@127.0.0.1:5672?heartbeat=600

 ✔ daemon:      Daemon is running with PID 196463

(aiidaENV) rkarkee@ch-fe1:~>

My first question is how do I submit job in HPC cluster for AiiDA? 

For example, whenever I run pw.x in HPC, I have the following job script. How do I run AiiDA in a job script? 

Is it like I need to create a python file and run it inside the job script?

#!/bin/bash

#SBATCH --time=6:30:00   # walltime

#SBATCH --nodes=1   # number of nodes

#SBATCH --ntasks-per-node=256   # number of tasks per node

#SBATCH --job-name=HfTe5   # job name

#SBATCH --partition=standard   # partition name

#SBATCH --no-requeue   # do not requeue when preempted and on node failure

#SBATCH --signal=30@20  # send signal to job at [seconds] before end

module swap PrgEnv-cray PrgEnv-intel

module load cmake

#srun -n 256   /users/rkarkee/q-e-qe-7.3/bin/pw.x -pd .true.    < scf.in &> scf.out

And how can I create a python file for the following pw.x input (scf.in)?

I want to use pseudopotential from /lustre/scratch4/turquoise/rkarkee/pbe_pseudo

&CONTROL

    calculation = 'scf'

    pseudo_dir  = '/lustre/scratch4/turquoise/rkarkee/pbe_pseudo'

        outdir       = './'

        prefix = 'scf'

        tprnfor = .true.

    tstress = .true.

   wf_collect=.true.

/

&SYSTEM

    ecutwfc     =  70,

    ibrav       = 0,

    nat         = 12,

    ntyp        = 2,

    lspinorb=.true.

noncolin = .true.

  vdw_corr='Grimme-D2'

/

&ELECTRONS

    conv_thr         =  1.00000e-10

    electron_maxstep = 200

    mixing_beta      =  0.2

    diagonalization  = "david"

/

K_POINTS {automatic}

4 4 2  1 1 1

ATOMIC_SPECIES

Hf     178.49  Hf.upf

Te      127.60  Te.upf

CELL_PARAMETERS (angstrom)

 1.98619742   7.132231003  0.

 -1.986197815  7.132230891  0.

  0.           0.          13.586526936

ATOMIC_POSITIONS (crystal)

Hf            0.3184997032        0.3184997302        0.7499999470

Hf            0.6815000238        0.6815000678        0.2499999640

Te            0.6625100716        0.6625100596        0.7499999470

Te            0.3374897664        0.3374897594        0.2499999640

Te            0.9351348779        0.9351349339        0.8531524632

Te            0.0648648561        0.0648648571        0.3531525342

Te            0.9351348779        0.9351349339        0.6468473588

Te            0.0648648561        0.0648648571        0.1468474118

Te            0.7919052147        0.7919052377        0.4336792513

Te            0.2080945513        0.2080945543        0.9336791443

Te            0.7919052147        0.7919052377        0.0663207087

Te            0.2080945513        0.2080945543        0.5663206417

[Lorenzo Bastonero]

Dear Rijan,

Thanks a lot for the interest in using AiiDA and aiida-vibroscopy.

There is also this other tutorial webpage for aiida tutorials, with more extended tutorials specific to aiida-quantumespresso: https://aiida-tutorials.readthedocs.io/en/latest/sections/running_processes/index.html

So, I'd suggest you to follow this one in case you didn't find it before, as it would answer most of your questions.

In the following some pointers that may be useful after you would follow the tutorials, but then feel free to come back again for any further explanations.

- Aiida-vibroscopy uses workflows provided by aiida-quantumespresso, which handles `pw.x` workflows with error handling features - namely the `PwBaseWorkChain`. This is the one you should use. To use it, you need to define a `computer` and a `code`.

- The  `computer`, which in you case would be your HPC, contains info regarding how to connect to that computer (e.g. ssh info, host name, etc). Please have a look here for a more details: https://aiida.readthedocs.io/projects/aiida-core/en/latest/howto/run_codes.html#computer-setup

You should achieve this with the following yaml file (save it as e.g. `computer.yaml`):

```

label: hpc # PLEASE CHANGE HERE AS YOU WISH
description: A DESCRIPTION # PLEASE CHANGE HERE AS YOU WISH
hostname: <PLEASE ADD HOSTNAME> # for example lumi.csc.fi for LUMI-G supercomputer; in your case it can also be a resolved address like 111.222.33.44
transport: core.ssh
scheduler: core.slurm
work_dir: "/scratch/path/to/my_aiida_runs/" # PLEASE CHANGE HERE AS YOU WISH
mpirun_command: "srun -n {tot_num_mpiprocs}" # leave it like this, so that `{tot_num_mpiprocs}` can be changed dynamically, e.g. in case you have other partitions on the same cluster
mpiprocs_per_machine: 256 # your default number of mpiprocs on the default partition

```

And then from the console:

```

verdi computer setup --config computer.yaml

```

Verify it work via 

```

verdi computer test hpc

```

- Second, you have to create a `code`, which instead contains the information regarding the code only itself _on a specific `computer`_, e.g. how to load the needed modules before the execution, where is the binary, etc. Please have a look here instead for how to set up a code on top of a computer: https://aiida.readthedocs.io/projects/aiida-core/en/latest/howto/run_codes.html#how-to-create-a-code

You should achieve this with the following yaml file (save it as e.g. `code.yaml`):

```

label: 'qe-7.3'
description: 'quantum_espresso v7.3'
default_calc_job_plugin: 'quantumespresso.pw'
filepath_executable: '/users/rkarkee/q-e-qe-7.3/bin/pw.x'
computer: 'hpc'
prepend_text: |

    module swap PrgEnv-cray PrgEnv-intel
    module load cmake

append_text: ' '

```

And then install it in AiiDA:

```

verdi code create core.code.installed --config code.yaml

```

You should obtain after these steps a code, in this example it would be called `pw-7.3@hpc`. You can verify you have your code using `verdi code list`. The preparation of sbatch script, input file, parsing etc is all done by AiiDA, so you don't have to worry anymore about all these details. You just need to specify the code `pw-7.3@hpc`.

 

Once you have this code, then all the other info regarding the scheduler would come later. In AiiDA, these information (e.g. walltime, number of nodes, ...) are passed to the `CalcJob` (in this case `PwCalculation`) via the `metadata.options` namespace. In your case, this can be achieved by specifying in the `PwBaseWorkChain` the `metadata.options` for the `PwCalculation`. If you are familiar with the builder, then you would have something like the following:

```

from aiida_quantumespresso.workflows.pw.base import PwBaseWorkChain

builder = PwBaseWorkChain.get_builder()

builder.pw.metadata.options = {

  'resources': {'num_machines':1, 'num_mpiprocs_per_machine': 256, 'num_cores_per_mpiproc': 1},

  'max_wallclock_seconds': 43200, # walltime in seconds

  'queue_name': 'standard',

}

```

Other options for the scheduler can be found here: https://aiida.readthedocs.io/projects/aiida-core/en/latest/reference/apidoc/aiida.schedulers.html#module-aiida.schedulers.datastructures

The code would be specified via:

```

from aiida.orm import *

builder.pw.code = load_code('pw-7.3@hpc')

```

- For specific pw parallelization options, these are specified in the settings:

```

builder.pw.settings = Dict({'cmdline':['-pd', '.true.', '-nk', '4']}) # mind that everything should be a string, even numbers!

```

- For the structure, you need to create a `StructureData` instance. A possible way would be to use:

```

from aiida import load_profile

load_profile()

from ase.io import read

from aiida.orm import *

atoms = read('/path/to/input.pwi') # here I mean the QE pw.x input you just showed

structure = StructureData(ase=atoms)

structure.store()

```

This last step would print a number (the "PK"); save it somewhere so you can load again the structure information. For instance, in this way:

```

from aiida import load_profile

load_profile()

from aiida.orm import *

structure = load_node(PK) # PK is an integer number

```

- The pseudopotentials in AiiDA are not taken from some folder in the HPC, instead they are each time loaded ("copy/pasted") on the `computer`. As such, you can define a group where you store (in AiiDA) all you pseudos, like if you had them in a folder. Please, have a look here: https://aiida-pseudo.readthedocs.io/en/latest/howto.html#installing-from-archive-or-folder

For instance:

```

aiida-pseudo install family /path/to/pseudo_folder/ MyPseudos -P pseudo.upf

```

Then you would load the pseudos from you structure as follows:

```

from aiida.orm import *

structure = load_node(PK)

pseudo_family = load_group('MyPseudos')

pseudos = pseudo_family.get_pseudos(structure=structure)

builder = PwBaseWorkChain.get_builder()

builder.pw.structure = structure

builder.pw.pseudos = pseudos

# continue here https://aiida-tutorials.readthedocs.io/en/latest/sections/running_processes/basics.html#preparing-the-calculation

```

A possible final python script would be:

```

from aiida import load_profile

from aiida.orm import *

from aiida.engine import submit

load_profile()

builder = PwBaseWorkChain.get_builder()

structure = load_node(PK)

code = load_code('pw-7.3@hpc')

pseudo_family = load_group('MyPseudos')

pseudos = pseudo_family.get_pseudos(structure=structure)

kpoints = KpointsData()

kpoints.set_kpoints_mesh([4,4,2],[0.5,0.5,0.5])

settings = {'cmdline':['-pd', '.true.', '-nk', '4']}

parameters = { # note: prefix, pseudos, and others are controlled by the plugin automatically!

  'CONTROL': {

    'tprnfor': True,
    'tstress': True,

  },
  'SYSTEM': { # note: ibrav, nat, ntyp are controlled by the plugin automatically!

    'ecutwfc': 70,

    'lspinorb': True,
    'noncolin': True,

    'vdw_corr': 'Grimme-D2',

  },
  'ELECTRONS': {
    'conv_thr': 1.0e-10
    'electron_maxstep': 200,
    'mixing_beta':  0.2,
    'diagonalization': 'david',

  },

}

builder.clean_workdir = Bool(False) # whether to remove the work directory after the run is finished

builder.kpoints = kpoints

builder.pw.structure = structure

builder.pw.pseudos = pseudos

builder.pw.code = code

builder.pw.parameters = Dict(parameters)

builder.pw.settings = Dict(settings)

builder.pw.metadata.options = {

  'resources': {'num_machines':1, 'num_mpiprocs_per_machine': 256, 'num_cores_per_mpiproc': 1},

  'max_wallclock_seconds': 43200, # walltime in seconds

  'queue_name': 'standard',

}

calc = submit(builder) # submit to aiida and run the workflow!

print(f'Launched a PwBaseWorkChain with PK = {calc.pk}') # register the PK and save it (e.g. in the same script); after it finishes, you can inspect the outputs

```

You can launch this script by simply `python script.py`, where `script.py` is a file containing the previous lines.

This can also be done via the `get_builder_from_protocol` method, which provides an easy way to fill the builder automatically, and then overrides the inputs you need (and very useful for more complex and nested inputs of aiida-vibroscopy - for vibroscopy workflows, have a look at its docs: https://aiida-vibroscopy.readthedocs.io/en/latest/howto/overrides.html):

```

from aiida import load_profile

from aiida.orm import *

from aiida.engine import submit

from aiida_quantumespresso.common.types import ElectronicType

import yaml

load_profile()

with open('/path/to/overrides.yaml', 'r') as handle:

  overrides = yaml.safe_load(handle)

structure = load_node(PK)

code = load_code('pw-7.3@hpc')

pseudo_family = load_group('MyPseudos')

pseudos = pseudo_family.get_pseudos(structure=structure)

kpoints = KpointsData()

kpoints.set_kpoints_mesh([4,4,2],[0.5,0.5,0.5])

overrides['pw']['pseudos'] = pseudos # need to specifically override pseudos

overrides['kpoints'] = kpoints # need to specifically override kpoints

kwargs = {'electronic_type': ElectronicType.INSULATOR} # this puts fixed occupations

builder = PwBaseWorkChain.get_builder_from_protocol(code=code, structure=structure, overrides=overrides, **kwargs)

submit(builder)

print(f'Launched a PwBaseWorkChain with PK = {calc.pk}') # register the PK and save it (e.g. in the same script); after it finishes, you can inspect the outputs

```

Where `overrides.yaml` is a yaml file containing your inputs :

```

clean_workdir: False
max_iterations: 10
meta_parameters:
conv_thr_per_atom: 1.e-11 
pw:
    parameters:
        SYSTEM:
            ecutwfc: 70
            ecutrho: 280
            lspinorb: True
            noncolin: True
            vdw_corr: Grimme-D2
        ELECTRONS:          
            mixing_beta: 0.2
            electron_maxstep: 200
            diagonalization: davidson
    settings:
        cmdline: ['-pd', '.true.', '-nk', '4']
    metadata:  
        options:
            resources:
                num_machines: 1
                num_mpiprocs_per_machine: 256
            max_wallclock_seconds: 43200  # Twelve hours
            queue_name: standard

```

As you see, the script will be much shorter, and you can separate better the inputs, so it's easier to change and resubmit the calculations.

Side note: please, consider to repost the same question on the Aiida Discourse channel, now the favoured communication channel, so that also other people can see this thread, thanks!

Hope this is useful.

Lorenzo

[Rijan Karkee]

Hi Lorenzo,

Thanks for the detailed instructions. 

I was unable to configure the computer. I followed the setup process as you suggested but it looks like it is looking for ssh. 

I am working in the home directory of HPC where I have installed AiiDA, Python etc in the home directory. So can AiiDA be run in the home directory of HPC or we install in local computer and try to run via there? Any suggestions on this?

Thanks

Best

Rijan

Creating the setup:

####################################################

(aiidaENV) rkarkee@ch-fe1:~> verdi computer setup

Report: enter ? for help.

Report: enter ! to ignore the default and set no value.

Computer label: hpc

Hostname: ch-fe.lanl.gov

Description []: LANL_HPC

Transport plugin: core.ssh

Scheduler plugin: core.slurm

Shebang line (first line of each script, starting with #!) [#!/bin/bash]: #!/bin/bash

Work directory on the computer [/scratch/{username}/aiida/]: /lustre/scratch4/turquoise/rkarkee/AiiDA

Mpirun command [mpirun -np {tot_num_mpiprocs}]: srun -n {tot_num_mpiprocs}

Default number of CPUs per machine: 256

Default amount of memory per machine (kB).:

Default amount of memory per machine (kB).: 51765

Escape CLI arguments in double quotes [y/N]: y

Success: Computer<2> hpc created

Report: Note: before the computer can be used, it has to be configured with the command:

Report:   verdi -p rkarkee computer configure core.ssh hpc

##################################################################################

Configuring the setup:

##########################################################################

(aiidaENV) rkarkee@ch-fe1:~> verdi -p rkarkee computer configure core.ssh hpc

Report: enter ? for help.

Report: enter ! to ignore the default and set no value.

User name [rkarkee]:

Port number [22]:

Look for keys [Y/n]: n

SSH key file []:

Connection timeout in s [60]:

Allow ssh agent [Y/n]: y

SSH proxy jump []:

SSH proxy command []:

Compress file transfers [Y/n]: y

GSS auth [False]:

GSS kex [False]:

GSS deleg_creds [False]:

GSS host [ch-fe.lanl.gov]:

Load system host keys [Y/n]:

Key policy (RejectPolicy, WarningPolicy, AutoAddPolicy) [RejectPolicy]:

Use login shell when executing command [Y/n]:

Connection cooldown time (s) [30.0]:

Report: Configuring computer hpc for user rka...@ucmerced.edu.

Success: hpc successfully configured for rka...@ucmerced.edu

######################################################################

Testing if it works:

#############################################################

(aiidaENV) rkarkee@ch-fe1:~> verdi computer test hpc

Report: Testing computer<hpc> for user<rka...@ucmerced.edu>...

* Opening connection... Error: Error connecting to 'ch-fe.lanl.gov' through SSH: [SshTransport] Server 'ch-fe.lanl.gov' not found in known_hosts, connect_args were: {'username': 'rkarkee', 'port': 22, 'look_for_keys': False, 'key_filename': '', 'timeout': 60, 'allow_agent': True, 'proxy_jump': '', 'proxy_command': '', 'compress': True, 'gss_auth': False, 'gss_kex': False, 'gss_deleg_creds': False, 'gss_host': 'ch-fe.lanl.gov'}

[FAILED]: Error while trying to connect to the computer

  Use the `--print-traceback` option to see the full traceback.

Warning: 1 out of 0 tests failed

##############################################

--
AiiDA is supported by the NCCR MARVEL (http://nccr-marvel.ch/), funded by the Swiss National Science Foundation, and by the European H2020 MaX Centre of Excellence (http://www.max-centre.eu/).
 
Before posting your first question, please see the posting guidelines at https://www.aiida.net/sections/mailing_list.html#posting-guidelines
---
You received this message because you are subscribed to the Google Groups "aiidausers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to aiidausers+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/84236b32-5c08-4eeb-a26c-2fcc6b95a8c0n%40googlegroups.com.

[Lorenzo Bastonero]

Dear Rijan,

If you have AiiDA already on the cluster, then you don’t have to configure an ssh computer configuration, but a local computer instead.

It is very simple, you should just change the following two lines:

```

hostname: localhost

transport: core.local

```

This is also shown here: https://aiida.readthedocs.io/projects/aiida-core/en/latest/howto/run_codes.html#computer-setup

Best,

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvt7H5Rd4XSG5QocaX2vKD%3DEbmqqAJOsphXLsV0LG8xK2g%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

I did that but now got a new error:

(aiidaENV) rkarkee@ch-fe2:~> verdi computer setup --config computer.yaml

Report: enter ? for help.

Report: enter ! to ignore the default and set no value.

Shebang line (first line of each script, starting with #!) [#!/bin/bash]: #!/bin/bash

Default amount of memory per machine (kB).: 51765

Escape CLI arguments in double quotes [y/N]: N

Critical: Could not connect to database: (psycopg2.OperationalError) connection to server at "localhost" (127.0.0.1), port 5432 failed: Connection refused

        Is the server running on that host and accepting TCP/IP connections?

(Background on this error at: https://sqlalche.me/e/20/e3q8)

(aiidaENV) rkarkee@ch-fe2:~>

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/ACE69D52-45A4-4CF9-8A15-BCCB81308ED1%40gmail.com.

[Lorenzo Bastonero]

Hello,

This seems to be Postgres that has been shut down. Could you please post this question on AiiDA Discourse (https://aiida.discourse.group/)?

This is the recommended communication channel, so that people can see also this question.

In the meantime, you might try this command:

```

pg_ctl start -l logfile

```

If you installed AiiDA in conda, then have a look here: https://aiida.readthedocs.io/projects/aiida-core/en/latest/intro/install_conda.html#intro-get-started-conda-install

Since this is more of an aiida-core questions, and since I am not directly an aiida-core developer, but a plugin developer (i.e. developer of aiida-quantumespresso, aiida-vibroscopy, etc that use aiida-core), then I would recommend if you open a new thread, e.g. on AiiDA Discourse. You will get much quicker answers there as well!

Best,

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvt13T82cbq3Wx7%3D5Fh9T4kh7kp0MaUeq1vWof09K1zMfQ%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

I talked to aiida.discourse  group and figured out the solution to that issue. 

I configured computer.yaml and code.yaml

I am creating a python file script.py

I have a few questions regarding your earlier suggestions.

I did run the following first:

First I created pk.py as

from aiida import load_profile

load_profile()

from ase.io import read

from aiida.orm import *

atoms = read('/path/to/input.pwi') # here I mean the QE pw.x input you just showed

structure = StructureData(ase=atoms)

pk=structure.store()

print(pk)

I got pk:3 as output

So I need to use this pk number right? Also do I need to do this step everytime even for same input?

Then 

I tried to setup pseudo potential using aiida-pseudo install family /path/to/pseudo_folder/ MyPseudos -P pseudo.upf

I always got some error on this (may I am doing this wrong). 

I have inside /users/rkarkee/pseudo the following pseudopotentials:

Hf.upf Te.upf Hf_FR.upf  Te_Fr.upf

I tried following but in both case got errors as:

(aiidaENV) rkarkee@ch-fe1:~/pseudo> aiida-pseudo install family /users/rkarkee/pseudo/ LDA Hf.upf

Usage: aiida-pseudo install family [OPTIONS] ARCHIVE LABEL

Try 'aiida-pseudo install family -h' for help.

Error: Got unexpected extra argument (Hf.upf)

(aiidaENV) rkarkee@ch-fe1:~/pseudo> aiida-pseudo install family /users/rkarkee/pseudo/ LDA

Report: creating a pseudopotential family from directory `/users/rkarkee/pseudo`... [FAILED]

Critical: `<class 'aiida_pseudo.data.pseudo.pseudo.PseudoPotentialData'>` constructor did not define the element and could not parse a valid element symbol from the filename `{filename}` either. It should have the format `ELEMENT.EXTENSION`

How do I do this correctly?

Thanks

Best

Rijan

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/681E255D-CFC1-423B-A215-EB67C2FDA066%40gmail.com.

[Lorenzo Bastonero]

So I need to use this pk number right?

Exactly, that is now stored in your database with that PK. So anytime you do `load_node(3)`, this will return the structure data you just stored.

Also do I need to do this step everytime even for same input?

Thanks to the database, and the provenance, you will not have to. You can just record somewhere the PK, and reuse it in other calculations. Of course, this will _stay_ as it is. So, if you perform a vc-relax with Quantum ESPRESSO for example, you may want to take the output structure instead, which will have a new PK.

Just for you information: you can put relevant nodes that you need in “Group”s, that work more or less as folders. So, in case you lost your PK, you can still look up into the group. Moreover, it is usually wise to give some proper `label` and `description`. You can even add extra information via `node.base.extras.set(‘key’, value)` if you need to keep something. Note that extras can be modified, unlike attributes (that you can get using e.g. node.base.attributes.all).

Then 

I tried to setup pseudo potential using aiida-pseudo install family /path/to/pseudo_folder/ MyPseudos -P pseudo.upf

I always got some error on this (may I am doing this wrong). 

I have inside /users/rkarkee/pseudo the following pseudopotentials:

Hf.upf Te.upf Hf_FR.upf  Te_Fr.upf

I tried following but in both case got errors as:

(aiidaENV) rkarkee@ch-fe1:~/pseudo> aiida-pseudo install family /users/rkarkee/pseudo/ LDA Hf.upf

Usage: aiida-pseudo install family [OPTIONS] ARCHIVE LABEL

Try 'aiida-pseudo install family -h' for help.

Error: Got unexpected extra argument (Hf.upf)

(aiidaENV) rkarkee@ch-fe1:~/pseudo> aiida-pseudo install family /users/rkarkee/pseudo/ LDA

Report: creating a pseudopotential family from directory `/users/rkarkee/pseudo`... [FAILED]

Critical: `<class 'aiida_pseudo.data.pseudo.pseudo.PseudoPotentialData'>` constructor did not define the element and could not parse a valid element symbol from the filename `{filename}` either. It should have the format `ELEMENT.EXTENSION`

How do I do this correctly?

As you can see from the error message, your files need to be in the format `ELEMENT.EXTENSION`, and of course you should not have more than one pseudo per element. You should create two pseudo families, hence two folders as well. So, for instance, just rename `Hf_FR.upf` to `Hf.FR.upf` and move it in an other folder (same operation for Te). Then you should be good to go.

Lorenzo

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvtzLvEcHFVOHQF7uZqOedFChZpeU6a2jBPm%3DMBnNhEDpw%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

Thanks for a quick reply. 

I got confused. 

When we do the actual code, we call pseudopotential through 

pseudo_family = load_group('MyPseudos') 

In this case, MyPseudos is the label that we use in command aiida-pseudo install family <Folder_to_Hf_pseudopotential>  MyPseudo 

Then in my input file I also need to use Te.upf but 

when I did command aiida-pseudo install family <Folder_to_Te_pseudopotential>  MyPseudo

It says MyPseudo already exists. 

I am confused what exactly we're doing here. Are we copying all required pseudopotentials in a common label or do we need to call different label in  here?

pseudo_family = load_group('MyPseudos_2')  here MyPsedos_2 is for Te.upf

  

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/A9D83350-13AF-432F-BA3A-34C3336FE72A%40gmail.com.

[Lorenzo Bastonero]

So, the idea is that you have a *group* of pseudo potentials, i.e. a *pseudofamily*. In that group, in principle all the elements will have its own pseudo potential (that should be generated in the same way technically). Then, all the pseudos, e.g. needed for a structure, can be loaded from it as I already wrote:

```

structure = load_node(PK)

pseudo_family = load_group('MyPseudos')

pseudos = pseudo_family.get_pseudos(structure=structure)

```

In your case, create two folders:

- - PseudosSR

    |_ Te.upf

    |_ Hf.upf

    |_ (… in principle for all the elements, but not necessarily)

- - PseudosFR

    |_ Te.FR.upf

    |_ Hf.FR.upf 

    |_ (… in principle for all the elements, but not necessarily)

And do aiida-pseudo for each of them, e.g. calling them with the same name of the folder, or as you wish (possibly something significant to where you took the pseudos).

If they are from PseudoDojo, consider the fact that there is already the possibility of downloading and creating the pseudo family group directly. 

Please, see instruction on the aiida-pseudo docs in case.

Lorenzo

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvuSELkyTZYzkdjKHp4gv8ntKdQGiy_T2YR6_GkGDQJkHg%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

Like you said I did create a folder and was successful in doing aiida-pseudo install .... 

as shown below.

(aiidaENV) rkarkee@ch-fe1:~> aiida-pseudo install family PseudosSR LDA_pseudo

Report: creating a pseudopotential family from directory `PseudosSR`... [OK]

Success: installed `LDA_pseudo` containing 2 pseudopotentials

Then I ran script.py (description shown below) but I think it has problem finding pseudopotential.

(aiidaENV) rkarkee@ch-fe1:/lustre/scratch4/turquoise/rkarkee> python script.py

Traceback (most recent call last):

  File "/lustre/scratch4/turquoise/.mdt5/rkarkee/script.py", line 40, in <module>

    builder.pw.pseudos = pseudos

    ^^^^^^^^^^^^^^^^^^

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida/engine/processes/builder.py", line 120, in __setattr__

    raise ValueError(f'invalid attribute value {validation_error.message}')

ValueError: invalid attribute value Invalid type <class 'aiida_pseudo.data.pseudo.pseudo.PseudoPotentialData'> for dynamic port value: expected (<class 'aiida.orm.nodes.data.upf.UpfData'>, <class 'aiida_pseudo.data.pseudo.upf.UpfData'>)

script.py file is as:

from aiida import load_profile

from aiida.orm import *

from aiida.engine import submit

from aiida_quantumespresso.workflows.pw.base import PwBaseWorkChain

load_profile()

builder = PwBaseWorkChain.get_builder()

structure = load_node(3)

code = load_code('qe-7.3@newhpc')

pseudo_family = load_group('LDA_pseudo')

pseudos = pseudo_family.get_pseudos(structure=structure)

kpoints = KpointsData()

kpoints.set_kpoints_mesh([4,4,2],[0.5,0.5,0.5])

settings = {'cmdline':['-pd', '.true.', '-nk', '4']}

parameters = { # note: prefix, pseudos, and others are controlled by the plugin automatically!

  'CONTROL': {

    'tprnfor': True,

    'tstress': True,

  },

  'SYSTEM': { # note: ibrav, nat, ntyp are controlled by the plugin automatically!

    'ecutwfc': 70,

    'lspinorb': True,

    'noncolin': True,

    'vdw_corr': 'Grimme-D2',

  },

  'ELECTRONS': {

    'conv_thr': 1.0e-10,

    'electron_maxstep': 200,

    'mixing_beta':  0.2,

    'diagonalization': 'david',

  },

}

builder.clean_workdir = Bool(False) # whether to remove the work directory after the run is finished

builder.kpoints = kpoints

builder.pw.structure = structure

builder.pw.pseudos = pseudos

builder.pw.code = code

builder.pw.parameters = Dict(parameters)

builder.pw.settings = Dict(settings)

builder.pw.metadata.options = {

  'resources': {'num_machines':1, 'num_mpiprocs_per_machine': 256, 'num_cores_per_mpiproc': 1},

  'max_wallclock_seconds': 43200, # walltime in seconds

  'queue_name': 'standard',

}

calc = submit(builder) # submit to aiida and run the workflow!

print(f'Launched a PwBaseWorkChain with PK = {calc.pk}') # register the PK and save it (e.g. in the same script); after it finishes, you can inspect the outputs

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/EC94C041-267B-4DBD-834D-F200061271F1%40gmail.com.

[Lorenzo Bastonero]

This is because Quantum ESPRESSO accepts only UPF format, and aiida-quantumespresso expects the corresponding data type, i.e. UpfData, and not PseudoPotentialData.

This is easily achieved by adding `-P pseudo.upf` when generating your family. See also here: https://aiida-pseudo.readthedocs.io/en/latest/howto.html#pseudopotential-format

I hope this time everything goes smoothly.

Lorenzo

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvskRez%2B4%3DyV53dYn9NSB9D3o38cwiGFT2Pp3WMZAVxSiQ%40mail.gmail.com.

[Rijan Karkee]

Yes Thanks.

I think it did work. 

after python script.py

It says:

Launched a PwBaseWorkChain with PK=14

I dont see any queues in hpc. 

Where do I check if it ran successfully/ outputs?

Best

Rijan

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/D09D9F3F-46CA-43D9-AA63-310A89435DEB%40gmail.com.

[Lorenzo Bastonero]

Very good!

Please, have a look at this tutorial: https://aiida-tutorials.readthedocs.io/en/latest/sections/running_processes/workflows.html#monitoring-the-process

It will give you all the information needed about running and inspecting *any* AiiDA WorkChain.

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvv-2dqFZeuR92Po_UUbMMzM5uQ%2BxM_8TgFgdQJY35WCEA%40mail.gmail.com.

[Lorenzo Bastonero]

PS: you can find the following commands useful to inspect `PwCalculation` (but in general any CalcJob)

verdi calcjob gotocomputer (goes to the folder where the calculation was submitted and run)

verdi calcjob outputcat (it prints the output of the main calculation, if any)

verdi calcjob inputcat (it prints the input of the main calculation - useful to see if there is what you expect)

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvv-2dqFZeuR92Po_UUbMMzM5uQ%2BxM_8TgFgdQJY35WCEA%40mail.gmail.com.

[Rijan Karkee]

Thanks,

one final question.

The main idea of using aiida was to do Raman (in my case) (both polarized and unpolarized at different temperature at different functionals.

In the same script.py,

what else do I need to add to achieve this?

I read the tutorial on aiida-vibroscopy. Can I add the following in the same script.py for Raman?

And with that Ramanbuilder (shown below) does it calculate polarized Raman as well as temperature dependent Raman?

from aiida.plugins import WorkflowFactory
from aiida.engine import run_get_node

IRamanSpectraWorkChain = WorkflowFactory("vibroscopy.spectra.iraman")

builder = IRamanSpectraWorkChain.get_builder_from_protocol(
    code=data.pw_code,
    structure=structure,
    protocol="fast",
)
results, calc = run_get_node(builder)
Do I run the following in separate .py file or it can be in same script.py

from aiida_vibroscopy.utils.plotting import get_spectra_plot

vibro = calc.outputs.vibrational_data.numerical_accuracy_4
polarized_intensities, unpolarized_intensities, frequencies_pbesol, labels = vibro.run_powder_raman_intensities(frequency_laser=532, temperature=300)

total_intensities_pbesol =  polarized_intensities + unpolarized_intensities
plt = get_spectra_plot(frequencies_pbesol, total_intensities_pbesol)
plt.show()

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/46947362-299A-44E3-9A97-B836666B94CB%40gmail.com.

[Lorenzo Bastonero]

I suggest you to run the main script where you have `results, calc = run_get_node(builder)` separately, save the PK, and use the other part of the code to post-process the generated data. The IRamanSpectraWorkChain will take a while to complete.

Some suggestions:

- probably you need to use your pseudos. Hence, in the script you will need to override the pseudos. This should be done with the following:

```

[…code we already saw in the previous email to load pseudos]

builder.dielectric.scf.pw.pseudos = pseudos

builder.phonon.scf.pw.pseudos = pseudos

```

- remember to put your code, so change the line `code=data.pw_code` with `code=load_code(‘your_pw_code@somewhere’)`

- if you need other  functionals a part from LDA/GGA, then you need also to specify in the pw parameters. See e.g. here how to customise your inputs: https://aiida-vibroscopy.readthedocs.io/en/latest/howto/overrides.html

For the functional, the QE input is SYSTEM -> input_dft (see here https://www.quantum-espresso.org/Doc/INPUT_PW.html#idm447)

- use the submit as in the script I sent you, namely:

```

from aiida.engine import submit

[…]

calc = submit(builder)

print(f‘Launched WorkChain with pk {calc.pk}’)

```

Record the pk, and if it finished correctly, then proceed to the post-processing, and add `calc = load_node(PK)`, PK of the IRamanSpectraWorkChain that you just run.

- other post-processing, once the IRamanSpectraWorkChain is finished correctly, can be found here: https://aiida-vibroscopy.readthedocs.io/en/latest/howto/postprocess.html

HTH

--------------------------------------

Lorenzo Bastonero

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvuZPgAj5ynvo_kQ86%3Dd50Z3X79-S0yWz3s8-J-bAgPOSg%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

Just as I understood correctly, for running a Raman would be similar to scf calculation

For example, would the following code works?

I want to do Raman calculation with my desired K-mesh, some other parameters like Grimme-D2 etc (like the one I showed in scf.in) earlier. How do I tell this code to do this? 

Does the following look right? 

Also I think this calculation may take longer time and I may have to restart time to time until everything is finished. How can I do that?

I am also confused because there are two builder ( one for Raman and one for Pw).

from aiida import load_profile

from aiida.orm import *

from aiida.engine import submit

from aiida_quantumespresso.workflows.pw.base import PwBaseWorkChain

from aiida.plugins import WorkflowFactory
from aiida.engine import run_get_node


IRamanSpectraWorkChain = WorkflowFactory("vibroscopy.spectra.iraman")

structure = load_node(3)
code = load_code('qe-7.3@newhpc')

pseudo_family = load_group('LDA_pseudo')

pseudos = pseudo_family.get_pseudos(structure=structure)

builder = IRamanSpectraWorkChain.get_builder_from_protocol(
    code=code,

    structure=structure,
    protocol="fast",
    
)

builder.dielectric.scf.pw.pseudos = pseudos

builder.phonon.scf.pw.pseudos = pseudos

results, calc = run_get_node(builder)

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/80310959-2E10-4457-905C-A57F7E791E12%40gmail.com.

[Lorenzo Bastonero]

Hello,

Indeed, you don't need the second one, i.e. the PwBaseWorkChain. the IRamanSpectraWorkChain calls the PhononWorkChain and the DielectricWorkChain, which call themselves nested PwBaseWorkChain. The `get_builder_from_protocol` fills all the inputs for you, so that you just need to readjust some of them.

I now pushed some examples scripts on the code repository, so you can try use one of them, or try mixing them as well. They are at this link: https://github.com/bastonero/aiida-vibroscopy/tree/main/examples/workflows/spectra

Before submitting, you can comment `calc = submit(builder)` and write `print(builder)` to verify all the inputs look as you expect, or just copy/paste the code in the `verdi shell` or in a `jupyter notebook`.

The nested input structure of each available workchain (aka workflow) of the plugin are available here: https://aiida-vibroscopy.readthedocs.io/en/latest/topics/workflows/index.html

Just be aware that these are more advanced workflows, so it would maybe good if you start e.g. with a PwRelaxWorkChain (look here for the input structure: https://aiida-quantumespresso.readthedocs.io/en/latest/topics/workflows/pw/relax.html), so that you get accustomed with running simpler AiiDA workflows first.

Hope you get the gist of it. Eventually, the builder is needed to fill all the inputs that you can find where I pointed you previously. The `get_builder_from_protocol` is a quicker way to get you started with the calculations, filling all the inputs for you. But you can always customize all the inputs via `builder.[...]`. Try it yourself in a `verdi shell` or better in a jupyter notebook.

HTH

Lorenzo

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvuSNg4N_Uo1mFpS9bMH3KGz_Nd5J0JXJXqVCuSM4S%3D4tA%40mail.gmail.com.

[Rijan Karkee]

Hi Lorenzo,

I see those examples are very useful and instructive.

One thing I missed there.

If I dont use overrides.yaml (other method)

where or how do I tell  the job submission? Like we did  in previous conversations?

Also, I guess in my machine  mpirun does not work and has to start with srun. How can I fix so that it submit job with srun?

builder.pw.metadata.options = {

  'resources': {'num_machines':1, 'num_mpiprocs_per_machine': 256, 'num_cores_per_mpiproc': 1},

  'max_wallclock_seconds': 43200, # walltime in seconds

  'queue_name': 'standard',

}

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CALawgCcxCQ%3DbxwUKqYdSxeq-PrSswBhthX75nCCHYQ2%3DfqS6Mg%40mail.gmail.com.

[Lorenzo Bastonero]

Happy that those help.

I would suggest you use the overrides, and simply add for each section you see the options you want to specify, and then remove everything else (although I bet you need to overrides to e.g. specify ecutwfc, ecutrho, vdw_corr, etc).

Regarding the command srun: this should be instructed at the `computer` level. So you need to set the computer node specifications accordingly.

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvvX6L2quL6TGLmJ6n2qJiKgygCxcNv7O06HgZCHQn0JGg%40mail.gmail.com.

[Rijan Karkee]

Ok.

In overrides.yaml, If I need to specify K-points as 4 4 2 1 1 1, 

How do I add ? Does the following would work?

And I do not see the way to tune ph.x parameters here.  For example, tr2_ph, etc

 scf:
    kpoints_distance: 0.4 # kpoints distance in Angstrom^-1 to sample the BZ
    kpoints_force_parity: false
    max_iterations: 5
    pw:
      metadata:
        options:
          max_wallclock_seconds: 43200
          resources:
            num_machines: 1
            num_mpiprocs_per_machine: 1
          # queue_name: partition_name # for SLURM
          # account: account_name # for SLURM, also for project etc
          withmpi: true
      parameters:
        ELECTRONS:
          conv_thr: 2.0e-12
          electron_maxstep: 80
          mixing_beta: 0.4
        SYSTEM:
          ecutrho: 240.0
          ecutwfc: 30.0

        KPOINTS:

         4 4 2 1 1 1

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CALawgCepduwHYnKrwfWUVwALiF0LR0gkrXO%3DFVHZDbsmKW-u8w%40mail.gmail.com.

[Lorenzo Bastonero]

You can have a look to the other scripts where I show how to use custom kpoint grids. It should be fairly easy to merge the two files. 

The ph.x is not used in this workflow, only pw.x. I suggest you maybe to have a look to our preprint paper: https://arxiv.org/abs/2308.04308

It exploits finite difference and finite electric fields. It doesn't use DFPT (aka ph.x), since ph.x can only do LDA and NC. 

The methodology implemented allow you to use any functional, but not via ph.x.

Hope this clarifies the confusion. If you want to use ph.x with LDA+NC, then you can use PwBaseWorkChain --> PhBaseWorkChain, but it's not yet implemented such workflow in aiida-vibroscopy as it has limited applicability (only LDA and NC pseudo).

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvvnhykfKPMJcOoiXSpeD8jQMmzeEc%3Du6toXNUt6WgdGqg%40mail.gmail.com.

[Rijan Karkee]

Understood.

I am trying to run the code. I got the error I guess it did not recognize the pseudopotential.  I have pseudo potential label as pbe_psp. I just combined other codes into submit_overrides.py (please find below).

(aiidaENV) rkarkee@ch-fe1:/lustre/scratch5/rkarkee> python submit_with_overrides.py

Traceback (most recent call last):

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida_quantumespresso/workflows/pw/base.py", line 161, in get_builder_from_protocol

    pseudo_family = orm.QueryBuilder().append(pseudo_set, filters={'label': pseudo_family}).one()[0]

                    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida/orm/querybuilder.py", line 1105, in one

    raise NotExistent('No result was found')

aiida.common.exceptions.NotExistent: No result was found

The above exception was the direct cause of the following exception:

Traceback (most recent call last):

  File "/lustre/scratch5/.mdt0/rkarkee/submit_with_overrides.py", line 85, in <module>

    main()

  File "/lustre/scratch5/.mdt0/rkarkee/submit_with_overrides.py", line 49, in main

    builder = IRamanSpectraWorkChain.get_builder_from_protocol(

              ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida_vibroscopy/workflows/spectra/iraman.py", line 113, in get_builder_from_protocol

    phonon = PhononWorkChain.get_builder_from_protocol(

             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida_vibroscopy/workflows/phonons/base.py", line 258, in get_builder_from_protocol

    scf = PwBaseWorkChain.get_builder_from_protocol(

          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

  File "/users/rkarkee/conda/envs/aiidaENV/lib/python3.11/site-packages/aiida_quantumespresso/workflows/pw/base.py", line 163, in get_builder_from_protocol

    raise ValueError(

ValueError: required pseudo family `SSSP/1.3/PBEsol/efficiency` is not installed. Please use `aiida-pseudo install` toinstall it.

%%%%%%%%%%%%%

overrides.yaml

%%%%%%%%%%%%%%%%%%%%%%%%

clean_workdir: false # whether to clean the working directiories

dielectric:

  clean_workdir: false

  kpoints_parallel_distance: 0.2 # kpoints distance in Angstrom^-1 to sample the BZ parallel to the electric field. If used, it should help in converging faster the final results

  property: raman

  # central_difference: # if you know what you are doing, custom numerical derivatives with respect to electric field

  #   accuracy: 2

  #   electric_field_step: 0.0005

  scf:

    kpoints_distance: 0.4 # kpoints distance in Angstrom^-1 to sample the BZ

    kpoints_force_parity: false

    max_iterations: 5

    pw:

      metadata:

        options:

          max_wallclock_seconds: 43200

          resources:

            num_machines: 1

            num_mpiprocs_per_machine: 256

            queue_name: standard # for SLURM

          # account: account_name # for SLURM, also for project etc

          withmpi: true

      parameters:

        ELECTRONS:

          conv_thr: 2.0e-10

          electron_maxstep: 100

          mixing_beta: 0.2

        SYSTEM:

          ecutrho: 280.0

          ecutwfc: 70.0

          vdw_corr: Grimme-D2

  settings:

    sleep_submission_time: 1.0

phonon:

  clean_workdir: false

  displacement_generator:

    distance: 0.01 # atomic displacements for phonon calculation, in Angstrom

  scf:

    kpoints_distance: 0.15 # kpoints distance in Angstrom^-1 to sample the BZ

    kpoints_force_parity: false

    max_iterations: 5

    pw:

      metadata:

        options:

          max_wallclock_seconds: 43200

          resources:

            num_machines: 1

            num_mpiprocs_per_machine: 1

          # queue_name: partition_name # for SLURM

          # account: account_name # for SLURM, also for project etc

          withmpi: true

      settings:

        cmdline: ['-nk', '8']

        # gamma_only: True # to use only if KpointsData has only a mesh 1 1 1 0 0 0 (i.e. Gamma not shifted)

      parameters:

        ELECTRONS:

          conv_thr: 2.0e-12

          electron_maxstep: 80

          mixing_beta: 0.4

        SYSTEM:

          ecutwfc: 70.0

          vdw_corr: Grimme-D2

  settings:

    sleep_submission_time: 1.0 # waiting time in seconds between different submission of SCF calculation. Recommended to be at least 1 second, to not overload.

settings:

  run_parallel: true

  use_primitive_cell: false

symmetry:

  distinguish_kinds: false

  is_symmetry: true

  symprec: 1.0e-05

%%%%%%%%%%%%%%%%%%

submit.py

%%%%%%%%%%%%%%%%%%%%%

# -*- coding: utf-8 -*-

# pylint: disable=line-too-long,wildcard-import,pointless-string-statement,unused-wildcard-import

"""Submit an IRamanSpectraWorkChain via the get_builder_from_protocol using the overrides."""

from pathlib import Path

from aiida import load_profile

from aiida.engine import submit

from aiida.orm import *

from aiida_quantumespresso.common.types import ElectronicType

from aiida_vibroscopy.workflows.spectra.iraman import IRamanSpectraWorkChain

load_profile()

# =============================== INPUTS =============================== #

# Please, change the following inputs.

mesh = [[4, 4, 2], [0.5, 0.5, 0.5]]

pseudo_family_name='pbe_psp'

pw_code_label = 'qe-7.3@newhpc'

structure_id = 21  # PK or UUID of your AiiDA StructureData

protocol = 'fast'  # also 'moderate' and 'precise'; 'moderate' should be good enough in general

overrides_filepath = './overrides.yaml'  # should be a path, e.g. /path/to/overrides.yaml. Format is YAML

# Consult the documentation for HOW-TO for how to use properly the overrides.

# !!!!! FOR FULL INPUT NESTED STRUCTURE: https://aiida-vibroscopy.readthedocs.io/en/latest/topics/workflows/spectra/iraman.html

# You can follow the input structure provided on the website to fill further the overrides.

# ====================================================================== #

# If you don't have a StructureData, but you have a CIF or XYZ, or similar, file

# you can import your structure uncommenting the following:

# from ase.io import read

# atoms = read('/path/to/file.cif')

# structure = StructureData(ase=atoms)

# structure.store()

# structure_id =  structure.pk

# print(f"Your structure has been stored in the database with PK={structure_id}")

def main():

    """Submit an IRamanSpectraWorkChain calculation."""

    code = load_code(pw_code_label)

    structure = load_node(structure_id)

    kwargs = {'electronic_type': ElectronicType.INSULATOR}

    kpoints = KpointsData()

    kpoints.set_kpoints_mesh(mesh[0], mesh[1])

    pseudo_family = load_group(pseudo_family_name)

    pseudos = pseudo_family.get_pseudos(structure=structure)

    builder = IRamanSpectraWorkChain.get_builder_from_protocol(

        code=code,

        structure=structure,

        protocol=protocol,

        overrides=Path(overrides_filepath),

        **kwargs,

    )

    builder.dielectric.scf.kpoints = kpoints

    builder.dielectric.pop('kpoints_parallel_distance', None)

    builder.dielectric.scf.pop('kpoints_distance', None)

    builder.phonon.scf.kpoints = kpoints

    builder.dielectric.scf.pw.pseudos = pseudos

    builder.phonon.scf.pw.pseudos = pseudos

    calc = submit(builder)

    print(f'Submitted IRamanSpectraWorkChain with PK={calc.pk} and UUID={calc.uuid}')

    print('Register *at least* the PK number, e.g. in you submit script.')

    print('You can monitor the status of your calculation with the following commands:')

    print('  * verdi process status PK')

    print('  * verdi process list -L IRamanSpectraWorkChain # list all running IRamanSpectraWorkChain')

    print(

        '  * verdi process list -ap1 -L IRamanSpectraWorkChain # list all IRamanSpectraWorkChain submitted in the previous 1 day'

    )

    print('If the WorkChain finishes with exit code 0, then you can inspect the outputs and post-process the data.')

    print('Use the command')

    print('  * verdi process show PK')

    print('To show further information about your WorkChain. When finished, you should see some outputs.')

    print('The main output can be accessed via `load_node(PK).outputs.vibrational_data.numerical_accuracy_*`.')

    print('You have to complete the remaning `*`, which depends upond the accuracy of the calculation.')

    print('See also the documentation and the reference paper for further details.')

if __name__ == '__main__':

    """Run script."""

    main()

To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CALawgCfqZXGddtSCayQsjVrhOzNbZhQhNSg5j7JFOcEHM%3DuLmA%40mail.gmail.com.

[Nicola Marzari]

Dear all,

Just checking - I think aiida has moved to discourse, correct?

Nicola

Sent from a tiny keyboard... Contact info:

http://theossrv1.epfl.ch/Main/Contact

On 13 Feb 2024, at 19:16, Rijan Karkee <rijan...@gmail.com> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/CAMaUnvtaEjgxJm8DZtMFK5vT2b8RFQX00%2BrQdJ33ANOSDuCTAw%40mail.gmail.com.

[Lorenzo]

Dear Nicola,

That’s correct. 

So I guess it would be now better to move the discussion there. Please, repost any further question there.

Thanks for your understanding.

Best,

Lorenzo

Il giorno 13 feb 2024, alle ore 19:32, 'Nicola Marzari' via aiidausers <aiida...@googlegroups.com> ha scritto:



To view this discussion on the web visit https://groups.google.com/d/msgid/aiidausers/7AC8DDE4-89A2-4451-82A7-1D13318C2C7E%40epfl.ch.