Source code for mppi.Calculators.YamboCalculator

"""
This module manages calculations performed with Yambo.
The run of the computation is performed by the python subprocess package (direct scheduler)
or by the slurm scheduler.
"""

from .Runner import Runner
from mppi.Calculators.Tools import find_string_file
from mppi.Calculators.RunRules import build_slurm_header, mpi_command
import os

[docs]def type_identifier(file): """ Found the type of the o-file according to its name. Different extension like 'hf' or 'hf_01' are recognized with the same type, that is 'hf' in this example. Args: file (:py:class:`string`) : file name Return: :py:class:`string` : string with the type of the file """ key = file.split('.')[-1] key_split = key.split('_') if key_split[-1].isdigit(): key = key.rsplit('_',1)[0] return key
[docs]def get_output_files(path): """ Scan the path with the output files and build a dictionary in which the keys are type of output (hf,qp,carriers,...) and the values are the names of the file for each type. If the path contains several replica of the output files (due to the fact the Yambo has been exectuted many times without cleaning the output folder) the function identifies the files associated to the last run. Args: path (:py:class:`string`) : folder where the o-* files are stored Return: :py:class:`dict` : Dictionary with the types (keys) and names values, including the path, of the files o-* produced by the run """ data = {} for file in os.listdir(path): if 'o-' in file: key = type_identifier(file) if key not in data: data[key] = [os.path.join(path,file)] else: data[key].append(os.path.join(path,file)) output = {} for key in data: data[key].sort(reverse=True) output[key] = data[key][0] return output
[docs]def get_report(path): """ Look for the name of the r-* file(s) produced by the execution of the code. If multiple istances of the report are found, the function selects the one assciated to the last computation. Args: path (:py:class:`string`) : folder with the r-* and the o-* files Return: :py:class:`string` :String with the name, including the path, of the file r-* produced by the run. If no report files is found an empty string is provided as output """ report = [] if os.path.isdir(path): for file in os.listdir(path): if 'r-' in file: report.append(os.path.join(path,file)) if len(report) == 0 : report.append(' ') report.sort(reverse=True) return report[0]
[docs]def get_db_files(dbsPath): """ Look for the files of tht type ndb.* in the dbsPath. Note that the first element of dbsPath is the folder where Yambo writes the ndb database. If a ndb is found in this folder its (eventual) replica in the other folders of dbsPath are not considered (this behavior protects under erroneuos identification of the correct databas that can happen, for instance, if a ndb given as input is not compliant with the actual computation so that Yambo has to compute it). Args: dbstPath (:py:class:`list`) : list of folders in which the ndb databases created by Yambo are sought Return: :py:class:`dict`: A dictionary in which the keys are the extension of the databases found and the values are their names, including the path """ dbs = {} for dir in dbsPath: if os.path.isdir(dir): for file in os.listdir(dir): if 'ndb' in file and 'fragment' not in file: key = file.split('.')[-1] if key not in dbs : dbs[key] = os.path.join(dir,file) return dbs
[docs]def build_results_dict(run_dir, outputPath, dbsPath, verbose = True): """ Return a dictionary with the names of the o- file(s), the report 'r-' file, the `ns.db1` in the SAVE folder and the names of the ndb database written by yambo in the jobname folder. Databases are sought in the dbsPath. Args: run_dir (:py:class:`string`) : `run_dir` folder of the calculation outputPath (:py:class:`string`) : folder with the 'o-' files dbsPath (:py:class:`list`) : list of folders with the ndb databases verbose (:py:class:`bool`) : set the amount of information provided on terminal Return: :py:class:`dict` : the dictionary {'output' : [o-1,o-2,...],'dft':...,'dipoles':..., ....} """ results = dict(output=get_output_files(outputPath),report=get_report(outputPath)) if verbose and len(results['output']) == 0: print(""" There are no o-* files. Maybe you have performed a computation that does not create any output file or wait_end_run and/or the dry_run option are active. Otherwise a possible error has occured during the computation """) # add the dft database from the SAVE folder dft = os.path.join(run_dir,'SAVE','ns.db1') if os.path.isfile(dft): results['dft'] = dft else: if verbose: print('ns.db1 database not found in SAVE folder') # add the dbs found in the jobname folder results.update(get_db_files(dbsPath)) return results
[docs]class YamboCalculator(Runner): """ Perform a Yambo calculation. The parameters used to define the parellelization strategy are provided in the `runRules` object. Parameters: runRulues (:class:`RunRules`) : instance of the :class:`RunRules` class executable (:py:class:`string`) : set the executable (yambo, ypp, yambo_rt, ...) of the Yambo package skip (:py:class:`bool`) : if True evaluate if the computation can be skipped. This is done by checking that the report file built by yambo exists and contains the string `game_over`, defined as a data member of this class clean_restart (:py:class:`bool`) : if True delete the folder with the output files and the database before running the computation dry_run (:py:class:`bool`) : with this option enabled the calculator setup the calculations and write the script for submitting the job, but the computations are not run wait_end_run (:py:class:`bool`) : with this option disabled the run method does not wait the end of the run. This option may be useful for interacting with the code in particular in _asincronous_ computation managed by the slurm scheduler activate_BeeOND (:py:class:`bool`) : if True set I/O of the run in the BeeOND_dir created by the slurm scheduler. The value of the ``BeeOND_dir`` is written as a data member of the class and can be modified if needed verbose (:py:class:`bool`) : set the amount of information provided on terminal fatlog (:py:class:`bool`) : if True set the `-fatlog` key to provide more information in the report file kwargs : other parameters that are stored in the _global_options dictionary Computations are performed in the folder specified by the ``run_dir`` parameter. The ``name`` parameter is used as name of the yambo input and as the name of the folder where yambo writes the o- `output` files. The ``jobname`` parameter is the name of the folder where yambo writes the .ndb databases. If this parameter is not provided in the run method the assumption jobname=name is made by the calculator. Example: >>> rr = RunRules(scheduler='slurm',ntasks_per_node=4,memory='124GB') >>> code = YamboCalculator(rr,executable='yambo',skip=True,verbose=True) >>> code.run(input = ..., run_dir = ...,name = ...,jobname = ..., **kwargs) When the run method is called the class runs the command: cd run_dir ; `mpirun command` executable_name -F name.in -J jobname -C name - O out_dir where the arguments of the run method are: Args: run_dir (:py:class:`string`) : the folder in which the simulation is performed input (:py:class:`string`) : instance of the :class:`YamboInput` class that define the input objects name (:py:class:`string`) : string with the names associated to the input file (without extension). This string is used also as the name of the folder in which results are written (argument of the -C option of yambo) as well as a part of the name of the output files jobname (:py:class:`list` or :py:class:`string`) : string (or list of strings) with the value(s) of the jobname folders (argument of the -J option of yambo). The first element is the folder name, where yambo writes the database. The other values (if provided) are the folders where yambo seeks for pre existing databases. All the elements of the list are assumed to be located in the ``run_dir`` of the calculator. If this variable is not specified the value of name is attributed to jobname out_dir (:py:class:`string`) : position of the folder in which the $jobname folder is located. This parameter is automatically set by the calculator the value of ``BeeOND_dir`` if the option `activate_BeeOND` is enabled. Otherwise all the folders are written in the ``run_dir`` kwargs : other parameters that are stored in the run_options dictionary """ BeeOND_dir = '/mnt/${SLURM_JOB_USER}-jobid_${SLURM_JOB_ID}' game_over = 'Game Over & Game summary' time_profile = 'Time-Profile' def __init__(self,runRules, executable = 'yambo', skip = True, clean_restart = True, dry_run = False, wait_end_run = True, activate_BeeOND = False, verbose = True, fatlog = False, **kwargs): # Use the initialization from the Runner class (all options inside _global_options) Runner.__init__(self,**runRules,executable=executable,skip=skip,clean_restart=clean_restart, dry_run=dry_run,wait_end_run=wait_end_run,activate_BeeOND=activate_BeeOND, verbose=verbose,fatlog=fatlog,**kwargs) print('Initialize a Yambo calculator with scheduler %s' %self._global_options['scheduler'])
[docs] def pre_processing(self): """ Process local run dictionary. Check that the run_dir exists and that it contains the SAVE folder. If clean_restart = True the results associated to previous computations are deleted before the run. The slurm out script is always deleted before the run, otherwise the calculator can erroneously assume that the run is completed. Note: If the run_dir and/or the SAVE folder do not exist an alert is written but the execution of the run method proceedes. """ run_dir = self.run_options.get('run_dir', '.') input = self.run_options.get('input') name = self.run_options.get('name','default') clean_restart= self.run_options.get('clean_restart') verbose = self.run_options.get('verbose') reformat = self.run_options.get('reformat',True) SAVE = os.path.join(run_dir,'SAVE') self.is_to_run() is_to_run = self.run_options.get('is_to_run') # Check if the run_dir and SAVE folder exists and write the input if not os.path.isdir(run_dir): print('Run_dir %s does not exists'%run_dir) elif not os.path.isdir(SAVE): print('SAVE folder does not exists') else: if input is not None: input.write(run_dir,name+'.in',reformat=reformat) else: print('input not provided') # Clean the folders before the run if is_to_run: self.clean_slurm_out() if clean_restart: self.clean_run() else: if verbose: print('run performed starting from existing results') return {}
[docs] def process_run(self): """ Method associated to the running of the executable. The method runs the computation and wait the end of the computation before passing to the :meth:`post_processing` method. """ is_to_run = self.run_options.get('is_to_run') if is_to_run: job = self.run_job() self.wait(job) return {}
[docs] def post_processing(self): """ Return a dictionary with the names of the o- file(s), the report 'r-' file, the `ns.db1` in the SAVE folder and the names of the ndb databases written by yambo in the jobname folder. The method performs a sanity check of the computation by checking that the `game_over` string is found in the report and shows the time needed to perform the simulation (if the `time_profile` string is found in the report). Return: :py:class:`dict` : the dictionary {'output' : [o-1,o-2,...],'report':...,'dft':...,'dipoles':..., ....} """ run_dir = self.run_options.get('run_dir', '.') name = self.run_options.get('name','default') outputPath = os.path.join(run_dir,name) jobname = self.run_options.get('jobname',name) verbose = self.run_options.get('verbose') if type(jobname) == str : dbsPath = [os.path.join(run_dir,jobname)] if type(jobname) == list : dbsPath = [os.path.join(run_dir,j) for j in jobname] results = build_results_dict(run_dir,outputPath,dbsPath,verbose=verbose) if verbose: report = results['report'] if find_string_file(report,self.game_over) is None: print('game_over string not found in report. Check the computation!') time_sim = find_string_file(report,self.time_profile) if self.run_options['is_to_run'] is True and time_sim is not None: print('Run performed in %s'%time_sim.split()[-1]) return results
[docs] def is_to_run(self): """ The method evaluates if the computation has to be performed. If ``skip`` is False the run is always performed, instead if ``skip`` is True the method checks if the report file exists and contains the string `game_over` defined as a member of the class. The method adds the key `is_to_run` to ``the run_options`` of the class. Note that the first element of the report list is used, so caution is needed if there is more than one report file in the report key of the results dictionary. """ skip = self.run_options.get('skip') run_dir = self.run_options.get('run_dir', '.') name = self.run_options.get('name','default') outputPath = os.path.join(run_dir,name) report = get_report(outputPath) verbose = self.run_options.get('verbose') if not skip: self.run_options['is_to_run'] = True else: if find_string_file(report,self.game_over) is not None: if verbose: print('Skip the run of',name) self.run_options['is_to_run'] = False else: self.run_options['is_to_run'] = True
[docs] def run_job(self): """ Run the computation. The operations performed depend on the scheduler adopted. If the dry_run option is enabled the run is not performed but the slurm script is written on disk. Return: The type of the object depends on the chosen scheduler. For scheduler `direct` job is an instance of Popen, while for `slurm` scheduler job is the name of the slurm script. """ from subprocess import Popen run_dir = self.run_options.get('run_dir', '.') scheduler = self.run_options['scheduler'] dry_run = self.run_options.get('dry_run',False) verbose = self.run_options.get('verbose') if scheduler == 'direct': # Set the OMP_NUM_THREADS variable in the environment os.environ['OMP_NUM_THREADS'] = str(self.run_options['omp_num_threads']) if not dry_run: comm_str = 'cd %s ; %s'%(run_dir,self.run_command()) job = Popen(comm_str, shell = True) else: job = None if verbose: print('Dry_run option active. Computation not performed') elif scheduler == 'slurm': job = self.build_slurm_script() if not dry_run: slurm_submit = 'cd %s ; sbatch %s.sh' %(run_dir,job) if verbose: print('slurm submit: ',slurm_submit ) slurm_run = Popen(slurm_submit, shell = True) else: if verbose: print('Dry_run option active. Script not submitted') else: print('scheduler unknown') return job
[docs] def wait(self,job): """ Wait the end of the job. If the dry_run option is enabled or wait_end_run is False the check is not performed. Args: jobs : The reference to the job to be executed. If the scheduler is `direct` jobs is an instance of Popen of the :py:class:subprocess package. If the scheduler is `slurm` jobs is a string with the name of the slurm script """ import time dry_run = self.run_options.get('dry_run',False) wait_end_run = self.run_options.get('wait_end_run',True) name = self.run_options.get('name','default') verbose = self.run_options.get('verbose') delay = 1 # in second if wait_end_run is True and dry_run is False: message_written = False while not self.run_ended(job): if not message_written: if verbose: print('computation %s is running...'%name) message_written = True time.sleep(delay) if verbose: print('computation %s ended'%name) else: if verbose: print('The wait_end_run is False or the dry_run option is active. The calculator proceedes to the postprocessing')
[docs] def build_slurm_script(self): """ Create the slurm script associated to the run. Return: :py:class:`string`: string with the name of the slurm script """ run_dir = os.path.abspath(self.run_options.get('run_dir', '.')) name = self.run_options.get('name','default') jobname = self.run_options.get('jobname',name) activate_BeeOND = self.run_options.get('activate_BeeOND') job = 'job_'+name comm_str = self.run_command() if type(jobname) == str : dbs_dir = os.path.abspath(os.path.join(run_dir,jobname)) if type(jobname) == list : dbs_dir = os.path.abspath(os.path.join(run_dir,jobname[0])) lines = build_slurm_header(self.run_options) if activate_BeeOND: lines.append('export BEEOND_DIR=%s'%self.BeeOND_dir) lines.append('export RUN_DIR=%s'%run_dir) lines.append('echo "The BeeOND option is activated. The I/O of the .ndb database is performed in $BEEOND_DIR"') lines.append('echo "RUN_DIR path is $RUN_DIR"') lines.append('echo "BEEOND_DIR path is $BEEOND_DIR"') lines.append('if [ ! -d $BEEOND_DIR ]; then') lines.append('echo "$BEEOND_DIR not found!"') lines.append('exit') lines.append('fi') lines.append('echo " "') lines.append('') # If the jobname folder with pre-existing ndb is found, it is copied in the BeeOND_dir path if os.path.isdir(dbs_dir): lines.append('export DBS_DIR=%s'%dbs_dir) lines.append('echo "found DBS_DIR folder $DBS_DIR. Copy the DBS_DIR in the $BEEOND_DIR folder"') lines.append('echo "rsync -rzv $DBS_DIR/ $BEEOND_DIR"') lines.append('rsync -rzv $DBS_DIR/ $BEEOND_DIR') lines.append('echo " "') lines.append('') lines.append('echo "Add the option -O $BEEOND_DIR to the run command of the calculator"') comm_str += ' -O %s'%self.BeeOND_dir lines.append('echo " "') lines.append('') # Do not show the " character in the of the comm_str in the echo line lines.append('echo "execute : %s"'%(comm_str.replace('"',''))) lines.append(comm_str) lines.append('echo " "') lines.append('') if activate_BeeOND: lines.append('echo "Copy the folder with the ndb database in the $RUN_DIR"') lines.append('echo "rsync -rzv $BEEOND_DIR/ $RUN_DIR"') lines.append('rsync -rzv $BEEOND_DIR/ $RUN_DIR') lines.append('echo " "') lines.append('') lines.append('echo "JOB_DONE"') f = open(os.path.join(run_dir,job+'.sh'),'w') f.write('\n'.join(lines)) f.close() return job
[docs] def run_command(self): """ Define the run command used to run the computation. Return: :py:class:`string` : command that runs the computation """ executable = self.run_options.get('executable') run_dir = self.run_options.get('run_dir', '.') name = self.run_options.get('name') jobname = self.run_options.get('jobname',name) fatlog = self.run_options.get('fatlog') verbose = self.run_options.get('verbose') if type(jobname) == str : J = jobname if type(jobname) == list : J = '"'+','.join(jobname)+'"' mpi_run = mpi_command(self.run_options) command = mpi_run + ' ' + executable input_name = name + '.in' if fatlog : command += ' -fatlog ' comm_str = command + ' -F %s -J %s -C %s'%(input_name,J,name) if verbose: print('run command: %s' %comm_str) return comm_str
[docs] def run_ended(self,job): """ Check the status of the running job. Args: job : reference to the actual job. job is an istance of Popen for `direct` scheduler or a string for `slurm` scheduler Return: :py:class:`bool`: return True if the computation is ended and False if it is running """ scheduler = self.run_options.get('scheduler') run_dir = self.run_options.get('run_dir', '.') if scheduler == 'direct': if job.poll() is not None: is_ended = True else: is_ended = False if scheduler == 'slurm': job_out = os.path.join(run_dir,job+'.out') if not os.path.isfile(job_out): is_ended = False else: with open(job_out, 'r') as f: last_line = f.read().splitlines()[-1] if last_line == 'JOB_DONE': is_ended = True else: is_ended = False return is_ended
[docs] def clean_slurm_out(self): """ Delete the job_$name.out file. """ run_dir = self.run_options.get('run_dir','.') name = self.run_options.get('name') verbose = self.run_options.get('verbose') job_out = os.path.join(run_dir,'job_'+name+'.out') if os.path.isfile(job_out): if verbose: print('delete job_out script:',job_out) os.system('rm %s'%job_out)
[docs] def clean_run(self): """ Delete existing results before performing the computation. Delete the out_dir folder and the folder with the databases. If several folders are provided in the jobname field, only the first one is deleted, since the others are assumed to contain databases obtained from other computations. """ run_dir = self.run_options.get('run_dir','.') name = self.run_options.get('name') jobname = self.run_options.get('jobname',name) verbose = self.run_options.get('verbose') out_dir = os.path.join(run_dir,name) if type(jobname) == list: ndb_dir = os.path.join(run_dir,jobname[0]) else: ndb_dir = os.path.join(run_dir,jobname) if os.path.isdir(out_dir): if verbose: print('delete folder:',out_dir) os.system('rm -r %s'%out_dir) if os.path.isdir(ndb_dir): if verbose: print('delete folder:',ndb_dir) os.system('rm -r %s'%ndb_dir)