Coverage for biobb_cp2k / cp2k / cp2k_run.py: 83%
76 statements
« prev ^ index » next coverage.py v7.13.4, created at 2026-03-05 12:58 +0000
1#!/usr/bin/env python3
3"""Module containing the Cp2kRun class and the command line interface."""
4from typing import Optional
5import shutil
6import os
7from pathlib import Path, PurePath
8from biobb_common.generic.biobb_object import BiobbObject
9from biobb_common.tools import file_utils as fu
10from biobb_common.tools.file_utils import launchlogger
11from biobb_cp2k.cp2k.common import check_input_path, check_output_path
14class Cp2kRun(BiobbObject):
15 """
16 | biobb_cp2k Cp2kRun
17 | Wrapper of the `CP2K QM tool <https://www.cp2k.org/>`_ module.
18 | Runs atomistic simulations of solid state, liquid, molecular, periodic, material, crystal, and biological systems using CP2K QM tool.
20 Args:
21 input_inp_path (str): Input configuration file (CP2K run options). File type: input. `Sample file <https://github.com/bioexcel/biobb_cp2k/raw/master/biobb_cp2k/test/data/cp2k/cp2k_energy.inp>`_. Accepted formats: inp (edam:format_2330), in (edam:format_2330), txt (edam:format_2330), wfn (edam:format_2333).
22 output_log_path (str): Output log file. File type: output. `Sample file <https://github.com/bioexcel/biobb_cp2k/raw/master/biobb_cp2k/test/reference/cp2k/cp2k_run_out.log>`_. Accepted formats: log (edam:format_2330), out (edam:format_2330), txt (edam:format_2330), o (edam:format_2330).
23 output_outzip_path (str): Output files. File type: output. `Sample file <https://github.com/bioexcel/biobb_cp2k/raw/master/biobb_cp2k/test/reference/cp2k/cp2k_run_out.zip>`_. Accepted formats: zip (edam:format_3987), gzip (edam:format_3987), gz (edam:format_3987).
24 output_rst_path (str): Output restart file. File type: output. `Sample file <https://github.com/bioexcel/biobb_cp2k/raw/master/biobb_cp2k/test/reference/cp2k/cp2k_run_out.wfn>`_. Accepted formats: wfn (edam:format_2333).
25 properties (dict - Python dictionary object containing the tool parameters, not input/output files):
26 * **binary_path** (*str*) - ("cp2k.sopt") CP2K binary path to be used.
27 * **param_path** (*str*) - (None) Path to the CP2K parameter data files (BASIS_SET, POTENTIALS, etc.). If not provided, the parameter data files included in the package will be used.
28 * **mpi_bin** (*str*) - (None) Path to the MPI runner. Usually "mpirun" or "srun".
29 * **mpi_np** (*int*) - (0) [0~1000|1] Number of MPI processes. Usually an integer bigger than 1.
30 * **mpi_flags** (*str*) - (None) Path to the MPI hostlist file.
31 * **remove_tmp** (*bool*) - (True) [WF property] Remove temporal files.
32 * **restart** (*bool*) - (False) [WF property] Do not execute if output files exist.
33 * **sandbox_path** (*str*) - ("./") [WF property] Parent path to the sandbox directory.
35 Examples:
36 This is a use example of how to use the building block from Python::
38 from biobb_cp2k.cp2k.cp2k_run import cp2k_run
39 prop = {
40 'binary_path': 'cp2k.sopt'
41 }
42 cp2k_run(input_inp_path='/path/to/cp2k_input.inp',
43 output_log_path='/path/to/cp2k_log.log',
44 output_outzip_path='/path/to/cp2k_out.zip',
45 output_rst_path='/path/to/cp2k_rst.wfn',
46 properties=prop)
48 Info:
49 * wrapped_software:
50 * name: CP2K
51 * version: >=7.1.0
52 * license: GPL-2.0
53 * multinode: mpi
54 * ontology:
55 * name: EDAM
56 * schema: http://edamontology.org/EDAM.owl
58 """
60 def __init__(self, input_inp_path: str, output_log_path: str,
61 output_outzip_path: str, output_rst_path: str,
62 properties: Optional[dict] = None, **kwargs) -> None:
64 properties = properties or {}
66 # Call parent class constructor
67 super().__init__(properties)
68 self.locals_var_dict = locals().copy()
70 # Input/Output files
71 self.io_dict = {
72 'in': {'input_inp_path': input_inp_path},
73 'out': {'output_log_path': output_log_path,
74 'output_outzip_path': output_outzip_path,
75 'output_rst_path': output_rst_path}
76 }
78 # Properties specific for BB
79 self.properties = properties
80 self.binary_path = properties.get('binary_path', 'cp2k.sopt')
81 self.param_path = properties.get('param_path', None)
83 # Properties for MPI
84 self.mpi_bin = properties.get('mpi_bin')
85 self.mpi_np = properties.get('mpi_np')
86 self.mpi_flags = properties.get('mpi_flags')
88 # Check the properties
89 self.check_properties(properties)
90 self.check_arguments()
92 def check_data_params(self, out_log, out_err):
93 """ Checks input/output paths correctness """
95 # Check input(s)
96 self.io_dict["in"]["input_inp_path"] = check_input_path(self.io_dict["in"]["input_inp_path"], "input_inp_path", False, out_log, self.__class__.__name__)
98 # Check output(s)
99 self.io_dict["out"]["output_log_path"] = check_output_path(self.io_dict["out"]["output_log_path"], "output_log_path", False, out_log, self.__class__.__name__)
100 self.io_dict["out"]["output_outzip_path"] = check_output_path(self.io_dict["out"]["output_outzip_path"], "output_outzip_path", False, out_log, self.__class__.__name__)
101 self.io_dict["out"]["output_rst_path"] = check_output_path(self.io_dict["out"]["output_rst_path"], "output_rst_path", False, out_log, self.__class__.__name__)
103 @launchlogger
104 def launch(self):
105 """Launches the execution of the Cp2kRun module."""
107 # check input/output paths and parameters
108 self.check_data_params(self.out_log, self.err_log)
110 # Setup Biobb
111 if self.check_restart():
112 return 0
113 self.stage_files()
115 # Creating temporary folder
116 tmp_folder = fu.create_unique_dir()
117 fu.log('Creating %s temporary folder' % tmp_folder, self.out_log)
119 shutil.copy2(self.io_dict["in"]["input_inp_path"], tmp_folder)
121 # set path to the CP2K parameter data files
122 if not self.param_path:
123 # os.environ["CP2K_DATA_DIR"] = str(PurePath(myself.__file__).parent.joinpath('cp2k_data'))
124 os.environ["CP2K_DATA_DIR"] = str(Path(os.getenv("CONDA_PREFIX", "")).joinpath('cp2k_aux').joinpath('cp2k_data'))
125 else:
126 if not Path(PurePath(self.param_path)).exists():
127 fu.log(self.__class__.__name__ + ': Unexisting %s folder, exiting' % self.param_path, self.out_log)
128 raise SystemExit(self.__class__.__name__ + ': Unexisting %s folder' % self.param_path)
129 os.environ["CP2K_DATA_DIR"] = self.param_path
131 # Command line
132 # cp2k.sopt -i benzene_dimer.inp -o mp2_test.out
133 self.cmd = ['cd', tmp_folder, ';',
134 self.binary_path,
135 '-i', PurePath(self.io_dict["in"]["input_inp_path"]).name,
136 '-o', PurePath(self.io_dict["out"]["output_log_path"]).name]
138 # general mpi properties
139 if self.mpi_bin:
140 mpi_cmd = [self.mpi_bin]
141 if self.mpi_np:
142 mpi_cmd.append('-n')
143 mpi_cmd.append(str(self.mpi_np))
144 if self.mpi_flags:
145 mpi_cmd.extend(self.mpi_flags)
146 self.cmd = mpi_cmd + self.cmd
148 # Run Biobb block
149 self.run_biobb()
151 # Gather output files in a single zip file
152 self.output = PurePath(tmp_folder).joinpath("cp2k_out.zip")
153 out_files = []
154 restart = ''
155 for root, dirs, files in os.walk(tmp_folder):
156 for file in files:
157 # fu.log('FILE %s ' % file, self.out_log)
158 # if file.endswith('.out'):
159 # out_files.append(file)
160 # elif file.endswith('.wfn'):
161 # restart = file
163 if file.endswith('.wfn'):
164 restart = tmp_folder + '/' + file
165 else:
166 out_files.append(tmp_folder + '/' + file)
168 fu.zip_list(str(self.output), out_files, self.out_log)
170 # Copy outputs from temporary folder to output path
171 shutil.copy2(self.output, PurePath(self.io_dict["out"]["output_outzip_path"]))
172 shutil.copy2(PurePath(tmp_folder).joinpath(PurePath(self.io_dict["out"]["output_log_path"]).name), PurePath(self.io_dict["out"]["output_log_path"]))
173 if restart:
174 shutil.copy2(restart, PurePath(self.io_dict["out"]["output_rst_path"]))
176 # Copy files to host
177 self.copy_to_host()
179 # remove temporary folder(s)
180 self.tmp_files.append(tmp_folder)
181 self.remove_tmp_files()
183 self.check_arguments(output_files_created=True, raise_exception=False)
185 return self.return_code
188def cp2k_run(input_inp_path: str,
189 output_log_path: str, output_outzip_path: str, output_rst_path: str,
190 properties: Optional[dict] = None, **kwargs) -> int:
191 """Create :class:`Cp2kRun <cp2k.cp2k_run.Cp2kRun>`cp2k.cp2k_run.Cp2kRun class and
192 execute :meth:`launch() <cp2k.cp2k_run.Cp2kRun.launch>` method"""
193 return Cp2kRun(**dict(locals())).launch()
196cp2k_run.__doc__ = Cp2kRun.__doc__
197main = Cp2kRun.get_main(cp2k_run, "Running atomistic simulations of solid state, liquid, molecular, periodic, material, crystal, and biological systems using CP2K QM tool.")
199if __name__ == '__main__':
200 main()