Coverage for biobb_flexserv / pcasuite / pcz_unzip.py: 90%
42 statements
« prev ^ index » next coverage.py v7.13.4, created at 2026-03-05 13:10 +0000
1#!/usr/bin/env python3
3"""Module containing the PCZunzip class and the command line interface."""
4from typing import Optional
5import shutil
6from pathlib import PurePath
7from biobb_common.tools import file_utils as fu
8from biobb_common.generic.biobb_object import BiobbObject
9from biobb_common.tools.file_utils import launchlogger
12class PCZunzip(BiobbObject):
13 """
14 | biobb_flexserv PCZunzip
15 | Wrapper of the pcaunzip tool from the PCAsuite FlexServ module.
16 | Uncompress Molecular Dynamics (MD) trajectories compressed using Principal Component Analysis (PCA) algorithms.
18 Args:
19 input_pcz_path (str): Input compressed trajectory. File type: input. `Sample file <https://github.com/bioexcel/biobb_flexserv/raw/master/biobb_flexserv/test/data/pcasuite/pcazip.pcz>`_. Accepted formats: pcz (edam:format_3874).
20 output_crd_path (str): Output uncompressed trajectory. File type: output. `Sample file <https://github.com/bioexcel/biobb_flexserv/raw/master/biobb_flexserv/test/reference/pcasuite/traj.crd>`_. Accepted formats: crd (edam:format_3878), mdcrd (edam:format_3878), inpcrd (edam:format_3878), pdb (edam:format_1476).
21 properties (dict - Python dictionary object containing the tool parameters, not input/output files):
22 * **binary_path** (*str*) - ("pcaunzip") pcaunzip binary path to be used.
23 * **verbose** (*bool*) - (False) Make output verbose
24 * **pdb** (*bool*) - (False) Use PDB format for output trajectory
25 * **remove_tmp** (*bool*) - (True) [WF property] Remove temporal files.
26 * **restart** (*bool*) - (False) [WF property] Do not execute if output files exist.
27 * **sandbox_path** (*str*) - ("./") [WF property] Parent path to the sandbox directory.
29 Examples:
30 This is a use example of how to use the building block from Python::
32 from biobb_flexserv.pcasuite.pcz_unzip import pcz_unzip
33 prop = {
34 'pdb': False
35 }
36 pcz_unzip( input_pcz_path='/path/to/pcazip_input.pcz',
37 output_crd_path='/path/to/pcazip_traj.crd',
38 properties=prop)
40 Info:
41 * wrapped_software:
42 * name: FlexServ PCAsuite
43 * version: >=1.0
44 * license: Apache-2.0
45 * ontology:
46 * name: EDAM
47 * schema: http://edamontology.org/EDAM.owl
49 """
51 def __init__(self, input_pcz_path: str,
52 output_crd_path: str, properties: Optional[dict] = None, **kwargs) -> None:
54 properties = properties or {}
56 # Call parent class constructor
57 super().__init__(properties)
58 self.locals_var_dict = locals().copy()
60 # Input/Output files
61 self.io_dict = {
62 'in': {'input_pcz_path': input_pcz_path},
63 'out': {'output_crd_path': output_crd_path}
64 }
66 # Properties specific for BB
67 self.properties = properties
68 self.binary_path = properties.get('binary_path', 'pcaunzip')
69 self.verbose = properties.get('verbose', False)
70 self.pdb = properties.get('pdb', False)
72 # Check the properties
73 self.check_properties(properties)
74 self.check_arguments()
76 @launchlogger
77 def launch(self):
78 """Launches the execution of the FlexServ pcaunzip module."""
80 # Setup Biobb
81 if self.check_restart():
82 return 0
83 # self.stage_files()
85 # Internal file paths
86 # try:
87 # # Using rel paths to shorten the amount of characters due to fortran path length limitations
88 # input_pcz = str(Path(self.stage_io_dict["in"]["input_pcz_path"]).relative_to(Path.cwd()))
89 # output_crd = str(Path(self.stage_io_dict["out"]["output_crd_path"]).relative_to(Path.cwd()))
90 # except ValueError:
91 # # Container or remote case
92 # input_pcz = self.stage_io_dict["in"]["input_pcz_path"]
93 # output_crd = self.stage_io_dict["out"]["output_crd_path"]
95 # Manually creating a Sandbox to avoid issues with input parameters buffer overflow:
96 # Long strings defining a file path makes Fortran or C compiled programs crash if the string
97 # declared is shorter than the input parameter path (string) length.
98 # Generating a temporary folder and working inside this folder (sandbox) fixes this problem.
99 # The problem was found in Galaxy executions, launching Singularity containers (May 2023).
101 # Creating temporary folder
102 tmp_folder = fu.create_unique_dir()
103 fu.log('Creating %s temporary folder' % tmp_folder, self.out_log)
105 shutil.copy2(self.io_dict["in"]["input_pcz_path"], tmp_folder)
107 # Command line
108 # pcaunzip -i infile [-o outfile] [--pdb] [--verbose] [--help]
109 # self.cmd = [self.binary_path,
110 # "-i", input_pcz,
111 # "-o", output_crd
112 # ]
114 self.cmd = ['cd', tmp_folder, ';',
115 self.binary_path,
116 "-i", PurePath(self.io_dict["in"]["input_pcz_path"]).name,
117 "-o", PurePath(self.io_dict["out"]["output_crd_path"]).name
118 ]
120 if self.verbose:
121 self.cmd.append('-v')
123 if self.pdb:
124 self.cmd.append('--pdb')
126 # Run Biobb block
127 self.run_biobb()
129 # Copy outputs from temporary folder to output path
130 shutil.copy2(PurePath(tmp_folder).joinpath(PurePath(self.io_dict["out"]["output_crd_path"]).name), PurePath(self.io_dict["out"]["output_crd_path"]))
132 # Copy files to host
133 # self.copy_to_host()
135 # Remove temporary folder(s)
136 self.tmp_files.append(tmp_folder)
137 self.remove_tmp_files()
139 self.check_arguments(output_files_created=True, raise_exception=False)
141 return self.return_code
144def pcz_unzip(input_pcz_path: str,
145 output_crd_path: str,
146 properties: Optional[dict] = None, **kwargs) -> int:
147 """Create :class:`PCZunzip <flexserv.pcasuite.PCZunzip>`flexserv.pcasuite.PCZunzip class and
148 execute :meth:`launch() <flexserv.pcasuite.PCZunzip.launch>` method"""
149 return PCZunzip(**dict(locals())).launch()
152pcz_unzip.__doc__ = PCZunzip.__doc__
153main = PCZunzip.get_main(pcz_unzip, "Uncompress Molecular Dynamics (MD) compressed trajectories using Principal Component Analysis (PCA) algorithms.")
155if __name__ == '__main__':
156 main()