Coverage for biobb_analysis/ambertools/cpptraj_strip.py: 81%
85 statements
« prev ^ index » next coverage.py v7.6.4, created at 2024-11-14 14:38 +0000
1#!/usr/bin/env python3
3"""Module containing the Cpptraj Strip class and the command line interface."""
4import argparse
5from typing import Optional
6from pathlib import PurePath
7from biobb_common.generic.biobb_object import BiobbObject
8from biobb_common.configuration import settings
9from biobb_common.tools import file_utils as fu
10from biobb_common.tools.file_utils import launchlogger
11from biobb_analysis.ambertools.common import get_default_value, check_top_path, check_traj_path, check_out_path, get_binary_path, get_in_parameters, get_out_parameters, copy_instructions_file_to_container, get_mask
14class CpptrajStrip(BiobbObject):
15 """
16 | biobb_analysis CpptrajStrip
17 | Wrapper of the Ambertools Cpptraj module for stripping a defined set of atoms (mask) from a given cpptraj compatible trajectory.
18 | Cpptraj (the successor to ptraj) is the main program in Ambertools for processing coordinate trajectories and data files. The parameter names and defaults are the same as the ones in the official `Cpptraj manual <https://amber-md.github.io/cpptraj/CPPTRAJ.xhtml>`_.
20 Args:
21 input_top_path (str): Path to the input structure or topology file. File type: input. `Sample file <https://github.com/bioexcel/biobb_analysis/raw/master/biobb_analysis/test/data/ambertools/cpptraj.parm.top>`_. Accepted formats: top (edam:format_3881), pdb (edam:format_1476), prmtop (edam:format_3881), parmtop (edam:format_3881), zip (edam:format_3987).
22 input_traj_path (str): Path to the input trajectory to be processed. File type: input. `Sample file <https://github.com/bioexcel/biobb_analysis/raw/master/biobb_analysis/test/data/ambertools/cpptraj.traj.dcd>`_. Accepted formats: mdcrd (edam:format_3878), crd (edam:format_3878), cdf (edam:format_3650), netcdf (edam:format_3650), nc (edam:format_3650), restart (edam:format_3886), ncrestart (edam:format_3886), restartnc (edam:format_3886), dcd (edam:format_3878), charmm (edam:format_3887), cor (edam:format_2033), pdb (edam:format_1476), mol2 (edam:format_3816), trr (edam:format_3910), gro (edam:format_2033), binpos (edam:format_3885), xtc (edam:format_3875), cif (edam:format_1477), arc (edam:format_2333), sqm (edam:format_2033), sdf (edam:format_3814), conflib (edam:format_2033).
23 output_cpptraj_path (str): Path to the output processed trajectory. File type: output. `Sample file <https://github.com/bioexcel/biobb_analysis/raw/master/biobb_analysis/test/reference/ambertools/ref_cpptraj.strip.netcdf>`_. Accepted formats: mdcrd (edam:format_3878), crd (edam:format_3878), netcdf (edam:format_3650), nc (edam:format_3650), rst7 (edam:format_3886), ncrst (edam:format_2033), dcd (edam:format_3878), pdb (edam:format_1476), mol2 (edam:format_3816), binpos (edam:format_3885), trr (edam:format_3910), xtc (edam:format_3875), sqm (edam:format_2033).
24 properties (dic - Python dictionary object containing the tool parameters, not input/output files):
25 * **start** (*int*) - (1) [1~100000|1] Starting frame for slicing.
26 * **end** (*int*) - (-1) [-1~100000|1] Ending frame for slicing.
27 * **steps** (*int*) - (1) [1~100000|1] Step for slicing.
28 * **mask** (*str*) - ("all-atoms") Mask definition. Values: c-alpha (All c-alpha atoms; protein only), backbone (Backbone atoms), all-atoms (All system atoms), heavy-atoms (System heavy atoms; not hydrogen), side-chain (All not backbone atoms), solute (All system atoms except solvent atoms), ions (All ion molecules), solvent (All solvent atoms), AnyAmberFromatMask (Amber atom selection syntax like `@*`).
29 * **format** (*str*) - ("netcdf") Output trajectory format. Values: crd (AMBER trajectory format), cdf (Format used by netCDF software library for writing and reading chromatography-MS data files), netcdf (Format used by netCDF software library for writing and reading chromatography-MS data files), nc (Format used by netCDF software library for writing and reading chromatography-MS data files), restart (AMBER coordinate/restart file with 6 coordinates per line), ncrestart (AMBER coordinate/restart file with 6 coordinates per line), restartnc (AMBER coordinate/restart file with 6 coordinates per line), dcd (AMBER trajectory format), charmm (Format of CHARMM Residue Topology Files (RTF)), cor (Charmm COR), pdb (Protein Data Bank format), mol2 (Complete and portable representation of a SYBYL molecule), trr (Trajectory of a simulation experiment used by GROMACS), gro (GROMACS structure), binpos (Translation of the ASCII atom coordinate format to binary code), xtc (Portable binary format for trajectories produced by GROMACS package), cif (Entry format of PDB database in mmCIF format), arc (Tinker ARC), sqm (SQM Input), sdf (One of a family of chemical-data file formats developed by MDL Information Systems), conflib (LMOD Conflib).
30 * **binary_path** (*str*) - ("cpptraj") Path to the cpptraj executable binary.
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.
34 * **container_path** (*str*) - (None) Container path definition.
35 * **container_image** (*str*) - ('afandiadib/ambertools:serial') Container image definition.
36 * **container_volume_path** (*str*) - ('/tmp') Container volume path definition.
37 * **container_working_dir** (*str*) - (None) Container working directory definition.
38 * **container_user_id** (*str*) - (None) Container user_id definition.
39 * **container_shell_path** (*str*) - ('/bin/bash') Path to default shell inside the container.
41 Examples:
42 This is a use example of how to use the building block from Python::
44 from biobb_analysis.ambertools.cpptraj_strip import cpptraj_strip
45 prop = {
46 'start': 1,
47 'end': -1,
48 'steps': 1,
49 'mask': 'c-alpha',
50 'format': 'netcdf'
51 }
52 cpptraj_strip(input_top_path='/path/to/myTopology.top',
53 input_traj_path='/path/to/myTrajectory.dcd',
54 output_cpptraj_path='/path/to/newTrajectory.netcdf',
55 properties=prop)
57 Info:
58 * wrapped_software:
59 * name: Ambertools Cpptraj
60 * version: >=20.0
61 * license: GNU
62 * ontology:
63 * name: EDAM
64 * schema: http://edamontology.org/EDAM.owl
66 """
68 def __init__(self, input_top_path, input_traj_path, output_cpptraj_path,
69 properties=None, **kwargs) -> None:
70 properties = properties or {}
72 # Call parent class constructor
73 super().__init__(properties)
74 self.locals_var_dict = locals().copy()
76 # Input/Output files
77 self.io_dict = {
78 "in": {"input_top_path": input_top_path, "input_traj_path": input_traj_path},
79 "out": {"output_cpptraj_path": output_cpptraj_path}
80 }
82 # Properties specific for BB
83 self.instructions_file = get_default_value('instructions_file')
84 self.start = properties.get('start', 1)
85 self.end = properties.get('end', -1)
86 self.steps = properties.get('steps', 1)
87 self.mask = properties.get('mask', 'all-atoms')
88 self.format = properties.get('format', 'netcdf')
89 self.properties = properties
90 self.binary_path = get_binary_path(properties, 'binary_path')
92 # Check the properties
93 self.check_properties(properties)
94 self.check_arguments()
96 def check_data_params(self, out_log, err_log):
97 """ Checks all the input/output paths and parameters """
98 self.io_dict["in"]["input_top_path"], self.input_top_path_orig = check_top_path(self.io_dict["in"]["input_top_path"], out_log, self.__class__.__name__)
99 self.io_dict["in"]["input_traj_path"] = check_traj_path(self.io_dict["in"]["input_traj_path"], out_log, self.__class__.__name__)
100 self.io_dict["out"]["output_cpptraj_path"] = check_out_path(self.io_dict["out"]["output_cpptraj_path"], out_log, self.__class__.__name__)
101 self.in_parameters = {'start': self.start, 'end': self.end, 'step': self.steps, 'mask': self.mask}
102 self.out_parameters = {'format': self.format}
104 def create_instructions_file(self, container_io_dict, out_log, err_log):
105 """Creates an input file using the properties file settings"""
106 instructions_list = []
107 # different path if container execution or not
108 if self.container_path:
109 self.instructions_file = str(PurePath(self.container_volume_path).joinpath(self.instructions_file))
110 else:
111 self.instructions_file = str(PurePath(fu.create_unique_dir()).joinpath(self.instructions_file))
112 fu.create_name(prefix=self.prefix, step=self.step, name=self.instructions_file)
114 # parm
115 instructions_list.append('parm ' + container_io_dict["in"]["input_top_path"])
117 # trajin
118 in_parameters = self.in_parameters
119 in_params = ''
120 if in_parameters:
121 in_params = get_in_parameters(self.in_parameters, out_log, 'strip')
122 instructions_list.append('trajin ' + container_io_dict["in"]["input_traj_path"] + ' ' + in_params)
124 # mask
125 mask = self.in_parameters.get('mask', '')
126 if not mask or mask == 'None':
127 fu.log('No mask provided, exiting', out_log, self.global_log)
128 raise SystemExit('Mask parameter is mandatory')
129 strip_mask = get_mask(mask, out_log)
130 instructions_list.append('strip ' + strip_mask)
132 # trajout
133 out_params = get_out_parameters(self.out_parameters, out_log)
134 instructions_list.append('trajout ' + container_io_dict["out"]["output_cpptraj_path"] + ' ' + out_params)
136 # create .in file
137 with open(self.instructions_file, 'w') as mdp:
138 for line in instructions_list:
139 mdp.write(line.strip() + '\n')
141 return self.instructions_file
143 @launchlogger
144 def launch(self) -> int:
145 """Execute the :class:`CpptrajStrip <ambertools.cpptraj_strip.CpptrajStrip>` ambertools.cpptraj_strip.CpptrajStrip object."""
147 # check input/output paths and parameters
148 self.check_data_params(self.out_log, self.err_log)
150 # Setup Biobb
151 if self.check_restart():
152 return 0
153 self.stage_files()
155 # create instructions file
156 self.create_instructions_file(self.stage_io_dict, self.out_log, self.err_log)
158 # if container execution, copy intructions file to container
159 if self.container_path:
160 copy_instructions_file_to_container(self.instructions_file, self.stage_io_dict['unique_dir'])
162 # create cmd and launch execution
163 self.cmd = [self.binary_path, '-i', self.instructions_file]
165 # Run Biobb block
166 self.run_biobb()
168 # Copy files to host
169 self.copy_to_host()
171 # remove temporary folder(s)
172 self.tmp_files.extend([
173 self.stage_io_dict.get("unique_dir", ""),
174 str(PurePath(self.instructions_file).parent)
175 ])
176 self.remove_tmp_files()
178 self.check_arguments(output_files_created=True, raise_exception=False)
180 return self.return_code
183def cpptraj_strip(input_top_path: str, input_traj_path: str, output_cpptraj_path: str, properties: Optional[dict] = None, **kwargs) -> int:
184 """Execute the :class:`CpptrajStrip <ambertools.cpptraj_strip.CpptrajStrip>` class and
185 execute the :meth:`launch() <ambertools.cpptraj_strip.CpptrajStrip.launch>` method."""
187 return CpptrajStrip(input_top_path=input_top_path,
188 input_traj_path=input_traj_path,
189 output_cpptraj_path=output_cpptraj_path,
190 properties=properties, **kwargs).launch()
193def main():
194 """Command line execution of this building block. Please check the command line documentation."""
195 parser = argparse.ArgumentParser(description="Strips a defined set of atoms (mask) from a given cpptraj compatible trajectory.", formatter_class=lambda prog: argparse.RawTextHelpFormatter(prog, width=99999))
196 parser.add_argument('--config', required=False, help='Configuration file')
198 # Specific args of each building block
199 required_args = parser.add_argument_group('required arguments')
200 required_args.add_argument('--input_top_path', required=True, help='Path to the input structure or topology file. Accepted formats: top, pdb, prmtop, parmtop, zip.')
201 required_args.add_argument('--input_traj_path', required=True, help='Path to the input trajectory to be processed. Accepted formats: crd, cdf, netcdf, restart, ncrestart, restartnc, dcd, charmm, cor, pdb, mol2, trr, gro, binpos, xtc, cif, arc, sqm, sdf, conflib.')
202 required_args.add_argument('--output_cpptraj_path', required=True, help='Path to the output processed trajectory.')
204 args = parser.parse_args()
205 args.config = args.config or "{}"
206 properties = settings.ConfReader(config=args.config).get_prop_dic()
208 # Specific call of each building block
209 cpptraj_strip(input_top_path=args.input_top_path,
210 input_traj_path=args.input_traj_path,
211 output_cpptraj_path=args.output_cpptraj_path,
212 properties=properties)
215if __name__ == '__main__':
216 main()