schrodinger.application.matsci.genetic_optimization module¶

Classes and functions for the genetic optimization module.

class schrodinger.application.matsci.genetic_optimization.PropertyInfo(key, units, is_positive, class_evaluator, class_kwargs)¶

Bases: tuple

class_evaluator¶: Alias for field number 3

class_kwargs¶: Alias for field number 4

is_positive¶: Alias for field number 2

key¶: Alias for field number 0

units¶: Alias for field number 1

class schrodinger.application.matsci.genetic_optimization.ClassEvaluator(structs, properties)¶

Bases: object

Manage a class evaluator.

SEPARATOR = '_'¶

HOST_STR = 'host_str'¶

__init__(structs, properties)¶

Create an instance.

Parameters:

structs (list of schrodinger.structure.Structure) – contains input structures
properties (list) – contains Property

getBaseName(struct, aproperty)¶

Get the base name.

Parameters:

struct (schrodinger.structure.Structure) – the structure
aproperty (Property) – the property

Return type:

str

Returns:

the base name

runIt()¶

Run it.

Raises:: RuntimeError – for any issue

setQueue(job_dj)¶

Set a JobDJ to run jobs.

Parameters:: job_dj (schrodinger.job.queue.JobDJ) – the queue

getQueue(properties, tpp=1, disable_smart_distribution=True)¶

Return a JobDJ to run jobs.

Parameters:

properties (list) – contains Property
tpp (int) – the threads per process
disable_smart_distribution (bool) – if True then force disable smart distribution regardless of the input tpp value

Return type:

JobDJ

Returns:

the queue

static getHostStr(host_str=None)¶

Return the host string.

Parameters:: host_str (str) – the host string, for example ‘localhost:4’
Return type:: str
Returns:: the host string

class schrodinger.application.matsci.genetic_optimization.StructureEvaluator(structs, properties)¶

Bases: ClassEvaluator

Manage structure evaluation.

SMARTS_PATTERN_SEPARATOR = '_'¶

SMARTS_PROP = 'smarts'¶

MOL_WEIGHT_PROP = 'molecular_weight'¶

NATOMS_PROP = 'natoms'¶

NELEMENTS_PROP = 'nelements'¶

PROPERTIES = {'molecular_weight', 'natoms', 'nelements', 'smarts'}¶

SMARTS_KEY = 'i_matsci_SMARTS_property_%s'¶

MOL_WEIGHT_KEY = 'r_m_Molecular_weight'¶

NATOMS_KEY = 'i_m_Number_of_atoms'¶

NELEMENTS_KEY = 'i_m_Number_of_elements'¶

NO_UNITS = 'None'¶

MOL_WEIGHT_UNITS = 'g/mol'¶

PATTERNS = 'patterns'¶

__init__(structs, properties)¶: See parent class for documentation.

static getInfo(key, units, patterns=None, host_str=None)¶

Return a PropertyInfo.

Parameters:

key (str) – the property key
units (str) – the property units
patterns (list) – the SMARTS patterns
host_str (str) – the host string, for example ‘localhost:4’

Return type:

PropertyInfo

Returns:

the property information

runIt()¶: Run it.

class schrodinger.application.matsci.genetic_optimization.ChemInformatics(structs, properties)¶

Bases: ClassEvaluator

Manage cheminformatics jobs.

UNITS = 'unknown'¶

PROP_BASE_NAME = None¶

CHECK_MSG = ''¶

OUT_EXT = '.mae'¶

__init__(structs, properties)¶: See parent class for documentation.

classmethod getInfo(name, units, model_file, tpp=1, host_str=None)¶

Return a PropertyInfo.

Parameters:

cls (type) – the calling class
name (str) – the property name
units (str) – the property units
model_file (str) – the model file
tpp (int) – the threads per process
host_str (str) – the host string, for example ‘localhost:4’

Return type:

PropertyInfo

Returns:

the property information

getModelFile(aproperty)¶

Return the model file from the given property.

Parameters:: aproperty (Property) – the property
Return type:: str
Returns:: the model file

copyModelFiles()¶: Copy the model files to the CWD.

makeMaestroInfile(struct, aproperty)¶

Make Maestro infile.

Parameters:

struct (schrodinger.structure.Structure) – the structure
aproperty (Property) – the property

Return type:

str

Returns:

the Maestro input file name

getPropertyValue(property_outfile, prop_key=None)¶

Get the property value.

Parameters:

property_outfile (str) – the property output file
prop_key (str or None) – if the property output file is a Maestro file then this specifies the structure property key

Raises:

RuntimeError – if property output file doesn’t exist or doesn’t contain the property value

Return type:

float

Returns:

the property value

generateExtraInput(struct)¶

Generate any extra input files needed for the actual property evaluation.

Parameters:: struct (schrodinger.structure.Structure) – the structure

getCmd(mae_infile, model_infile, prop_name, xtra_infile, property_outfile, job_name)¶

Get the command line.

Parameters:

mae_infile (str) – Maestro file containing the structure
model_infile (str) – the cheminformatics model file
prop_name (str) – a property name possibly used to identify the property value
xtra_infile (str) – an additional input file
property_outfile (str) – name of the output file containing the property value
job_name (str) – the job name

Return type:

list

Returns:

the command line

runPrediction(struct)¶

Run the prediction.

Parameters:: struct (schrodinger.structure.Structure) – the structure
Raises:: RuntimeError – if it fails

setUp()¶: Do any necessary set up prior to the actual property prediction.

runIt()¶: Run it.

classmethod getModelFiles(property_lists)¶

Return file names of any model files.

Parameters:: property_lists (list) – contains lists of property specifications
Return type:: list
Returns:: file names of the model files

classmethod checkModelFiles(model_files, host_str=None)¶

Check the given model files.

Parameters:

model_files (list[str]) – the names of the model files
host_str (str) – the host string, for example ‘localhost:4’

Raises:

RuntimeError – if there is an issue

static getAllModelFiles(property_lists, check=False, host_str=None)¶

Return file names of any model files.

Parameters:

property_lists (list) – contains lists of property specifications
check (bool) – whether to check the model files
host_str (str) – the host string, for example ‘localhost:4’

Return type:

list

Returns:

file names of the model files

class schrodinger.application.matsci.genetic_optimization.CanvasKPLS(structs, properties)¶

Bases: ChemInformatics

Manage Canvas KPLS jobs.

EXT = 'kpls.tar.gz'¶

FP_TEXT_FILE = 'fpInfo.txt'¶

MODEL_OPTION = 'kpls_model'¶

KEY = 'r_matsci_KPLS_%s/%s'¶

PATH = '/scr/buildbot/builds/core-suite-build/core-suite-build/build/mmshare-v7.0/data/genetic_optimization/canvas_kpls_models'¶

FP_EXT = '.fp'¶

VALUE_PATTERN = re.compile('\\s*1\\s+unknown.*\\s+(-?\\d+\\.?\\d*)$')¶

ALLOWED_FP_TYPES = ['linear', 'maccs', 'radial', 'molprint2D', 'torsion', 'pairwise', 'triplet', 'quartet', 'dendritic']¶

SINGLE_PRECISION = 32¶

DOUBLE_PRECISION = 64¶

BIT_EXT = '-bit'¶

LABEL = 'canvasKPLS'¶

CHECK_MSG = " Model files must correspond to a single property type and as input take fingerprint files featuring one of the following fingerprint types, ['linear', 'maccs', 'radial', 'molprint2D', 'torsion', 'pairwise', 'triplet', 'quartet', 'dendritic'], as well as either no atom type or one of the 12 known atom types (integers in [1, 12]). See $SCHRODINGER/utilities/canvasFPGen -h for more details."¶

makeFingerPrintInfile(mae_infile, name, job_dj=None)¶

Make fingerprint infile or add the job to do so to the given queue.

Parameters:

mae_infile (str) – the Maestro input file name
name (str) – the property name
job_dj (None or schrodinger.job.queue.JobDJ) – if given then add the fingerprint job to this queue and return

Raises:

RuntimeError – if canvasFPGen fails

Return type:

str

Returns:

the Canvas fingerprint input file name

getPropertyValue(property_outfile, prop_key=None)¶: See parent class.

generateExtraInput(struct)¶: See parent class.

makeFpOptionsDict()¶: Make the fingerprint options dictionary.

getCmd(mae_infile, model_infile, prop_name, xtra_infile, property_outfile, job_name)¶: See parent class.

setUp()¶: See parent class.

static getFpOptions(model_file)¶

Return fingerprint options obtained from the given Canvas KPLS model file.

Parameters:: model_file (str) – the name of the Canvas KPLS model file
Return type:: int, str, int or None
Returns:: contains (1) precision, (2) fingerprint type, and (3) atom type if present
Raises:: RuntimeError – if there is anything wrong with the Canvas KPLS model file

class schrodinger.application.matsci.genetic_optimization.AutoQSAR(structs, properties)¶

Bases: ChemInformatics

Manage AutoQSAR jobs.

EXT = 'qzip'¶

MODEL_OPTION = 'auto_qsar_model'¶

KEY = 'r_matsci_Auto_QSAR_%s/%s'¶

PROP_BASE_NAME = 'r_autoqsar_Pred_{prop_name}'¶

IN_EXT = '.inp'¶

LABEL = 'autoqsar'¶

writeAutoQSARInFile(mae_infile, model_infile, base_name, prop_name)¶

Write the AutoQSAR input file.

Parameters:

mae_infile (str) – the input Maestro file name
model_infile (str) – the input AutoQSAR model file name
base_name (str) – the base name to use for the AutoQSAR input file
prop_name (str) – the property name

Return type:

str

Returns:

the name of the AutoQSAR input file

generateExtraInput(struct)¶: See parent class.

getCmd(mae_infile, model_infile, prop_name, xtra_infile, property_outfile, job_name)¶: See parent class.

class schrodinger.application.matsci.genetic_optimization.DeepChem(structs, properties)¶

Bases: ChemInformatics

Manage DeepChem jobs.

EXT = 'qzip'¶

MODEL_OPTION = 'deep_autoqsar_model'¶

KEY = 'r_matsci_DeepAutoQSAR_%s/%s'¶

PROP_BASE_NAME = 'r_m_{prop_name}_score'¶

LABEL = 'deepautoqsar'¶

getCmd(mae_infile, model_infile, prop_name, xtra_infile, property_outfile, job_name)¶: See parent class.

class schrodinger.application.matsci.genetic_optimization.Jaguar(structs, properties)¶

Bases: ClassEvaluator

Manage Jaguar jobs.

JAGUAR_OPTIONS = 'jaguar_options'¶

TPP = 'tpp'¶

JAGUAR_OUTPUT_ATTR = 'jaguar_output_attr'¶

IN_EXT = '.in'¶

OUT_EXT = '.out'¶

__init__(structs, properties)¶: See parent class for documentation.

loadQueue(job_dj)¶

Load the JobDJ with jobs.

Parameters:: job_dj (JobDJ) – the queue

postProcess(silent=False)¶

Post process the jobs and set the final results.

Parameters:: silent (bool) – if True, don’t raise.
Raises:: RuntimeError – if there is a problem

runIt(silent=False)¶

Run it.

Parameters:: silent (bool) – if True, don’t raise.
Raises:: RuntimeError – if there is a problem

class schrodinger.application.matsci.genetic_optimization.GlassTransitionTemperature(structs, properties)¶

Bases: CanvasKPLS

Manage glass transition temperature jobs.

UNITS = 'C'¶

KEY = 'r_matsci_KPLS_Tg/C'¶

PROP = 'kpls_tg'¶

FILE = 'Tg250.kpls.tar.gz'¶

static getInfo(host_str=None)¶

Return a PropertyInfo.

Parameters:: host_str (str) – the host string, for example ‘localhost:4’
Return type:: PropertyInfo
Returns:: the property information

class schrodinger.application.matsci.genetic_optimization.RefractiveIndex(structs, properties)¶

Bases: CanvasKPLS, Jaguar

Manage refractive index jobs.

UNITS = 'none'¶

KEY = 'r_matsci_Refractive_Index_298K/none'¶

PROP = 'refractive_index'¶

FILE = '01a_kpls5_amorphous_density_HT.kpls.tar.gz'¶

MASS_DENSITY_KEY = 'r_matsci_Mass_Density/g/cm^3'¶

ISOTROPIC_POLARIZABILITY_KEY = 'r_matsci_Isotropic_Polarizability/bohr^3'¶

__init__(structs, properties)¶: See parent classes for documentation.

getQueue(*args, **kwargs)¶: See parent class for documentation.

static getInfo(jaguar_options=None, tpp=1, host_str=None)¶

Return a PropertyInfo.

Parameters:

jaguar_options (dict) – contain Jaguar options
tpp (int) – the threads per process
host_str (str) – the host string, for example ‘localhost:4’

Return type:

PropertyInfo

Returns:

the property information

classmethod setRefractiveIndexProperty(struct, mass_density, polarizability)¶

Set the refractive index property on the given structure.

Parameters:

struct (schrodinger.structure.Structure) – the structure
mass_density (float) – the mass density in g/cm^3
polarizability (float) – the isotropically averaged polarizability in atomic units of bohr^3

Raises:

ValueError – Abnormal small total_weight or large polarizability

runIt()¶: Run it.

class schrodinger.application.matsci.genetic_optimization.CustomClassEvaluator(structs, properties)¶

Bases: ClassEvaluator

Manage a custom class evaluator.

UNITS = 'unknown'¶

KEY = 'r_matsci_{name}/{units}'¶

CUSTOM_CLASS_FILE_OPTION = 'custom_class_file'¶

UNITS_DICT = {'stub': 'unknown'}¶

RUN_DICT = {'stub': 'runStub'}¶

EXTRA_INPUT_FILES_DICT = {'stub': []}¶

__init__(structs, properties)¶: See parent class for documentation.

static getInfo(name, units, custom_class_file, tpp=1, host_str=None)¶

Return a PropertyInfo.

Parameters:

name (str) – the property name
units (str) – the property units
custom_class_file (str) – the Python file containing the custom class
tpp (int) – the threads per process
host_str (str) – the host string, for example ‘localhost:4’

Return type:

PropertyInfo

Returns:

the property information

static getModule(path)¶

Return the module from the given path.

Parameters:: path (str) – the path to the module python file
Raises:: RuntimeError – if there is a problem
Return type:: object
Returns:: the module

static getExtraInputFiles(path, name)¶

Return the extra input files for the given property name.

Parameters:

path (str) – the path to the module python file
name (str) – the name of the property

Raises:

RuntimeError – if there is a problem

Return type:

list

Returns:

the extra input files

static checkDict(path, dict_name)¶

Check dictionary.

Parameters:

path (str) – the path to the module python file
dict_name (str) – the name of the dictionary to check

Raises:

RuntimeError – if there is a problem

static checkModule(path, name)¶

Check the module.

Parameters:

path (str) – the path to the module python file
name (str or None) – if given the name of the property using this module will be checked against the module

Raises:

RuntimeError – if there is a problem

static addInputFiles(job_builder, property_lists)¶

Add input files from the given properties to the job builder.

Parameters:

job_builder (launchapi.JobSpecificationArgsBuilder) – Job specification builder object
property_lists (list) – contains lists of property specifications

Raises:

RuntimeError – if there is a problem

runStub(st, job_dj, tpp)¶

Run stub.

Parameters:

st (schrodinger.structure.Structure) – the structure
job_dj (JobDJ) – the queue
tpp (int) – the threads per process

Raises:

RuntimeError – if there is a problem

Return type:

float

Returns:

the property value

runIt()¶

Run it.

Raises:: RuntimeError – if there is a problem

schrodinger.application.matsci.genetic_optimization.get_script_property_info_dict(host_str=None)¶

Return a (name, PropertyInfo) dict for script based properties.

Parameters:: host_str (str) – the host string, for example ‘localhost:4’
Return type:: dict
Returns:: contains (name, PropertyInfo)

schrodinger.application.matsci.genetic_optimization.get_property_info(name, jaguar_options=None, tpp=None, patterns=None, host_str=None)¶

Return a PropertyInfo for the given name and properties.

Parameters:

name (str) – the property name
jaguar_options (dict) – contains Jaguar options
tpp (int) – the threads per process
patterns (list) – the SMARTS patterns
host_str (str) – the host string, for example ‘localhost:4’

Return type:

PropertyInfo or None

Returns:

the PropertyInfo

schrodinger.application.matsci.genetic_optimization.get_random_csearch_seed(this_random=None)¶

Return a random csearch seed.

Parameters:: this_random (numpy.random.RandomState or None) – random state, if None use the module constant
Return type:: int
Returns:: the seed

exception schrodinger.application.matsci.genetic_optimization.PropertySyntaxError¶: Bases: Exception

exception schrodinger.application.matsci.genetic_optimization.UnknownPropertySuboptionError¶: Bases: Exception

class schrodinger.application.matsci.genetic_optimization.Property(index=1, key=None, name=None, units=None, minimax=None, target=None, comparator=None, error=None, weight=1.0, positive=None, summarize=None, class_kwargs=None)¶

Bases: object

Manage a property to be used in a genetic optimization.

MAX = 'max'¶

MIN = 'min'¶

EQUALS = 'eq'¶

GREATER_THAN = 'gt'¶

LESS_THAN = 'lt'¶

SUB_OPTIONS = ['index', 'key', 'name', 'units', 'minimax', 'target', 'comparator', 'error', 'weight', 'positive', 'patterns', 'summarize', 'kpls_model', 'custom_class_file', 'class_kwargs']¶

__init__(index=1, key=None, name=None, units=None, minimax=None, target=None, comparator=None, error=None, weight=1.0, positive=None, summarize=None, class_kwargs=None)¶

Create an instance.

Parameters:

index (int) – a numeric index used to refer to this Property instance, a default of 1 is used
key (str) – the schrodinger.structure.Structure property key to be optimized
name (str) – specify a name for the property, this name will be, for example used in any *log files, etc.
units (str) – enter the units that the property is in, for example eV, nm, etc.
minimax (str) – to minimize or maximize this property then set this option to the class constants MIN or MAX
target (float) – if instead of maximizing or minimizing the property, the genetic optimization is supposed to handle a specific value then enter that value using this option.
comparator (str) – specify here how the target value and computed values are to be compared, i.e. either the class constants EQUALS for =, GREATER_THAN for >, or LESS_THAN for <.
error (float) – if equality to a target value has been specified then this option allows the user to control the error bounds of the target value, if not specified then a default of 10% of the specified target value will be used.
weight (float) – specify the weight to use for this property, if the genetic optimization is to be run on several properties then the weight allows the user to bias the solution. This option can also be used to control a situation where more than a single property is desired and where those properties are quantified using different physical units such that the numbers might be orders of magnitude apart from one another, for example comparing eV and nm. A default of 1.0 is used.
positive (bool) – True if this property can only take on positive values, for example as in the area of a surface, False otherwise, for example as in temperature in Celcius. The default is False.
summarize (bool) – if True then print a summary of this property, False otherwise
class_kwargs (dict or None) – contains kwargs for class based evaluation of this property

setAttributes(index=1, key=None, name=None, units=None, minimax=None, target=None, comparator=None, error=None, weight=1.0, positive=None, summarize=None, class_kwargs=None)¶

Set some attributes for this class.

Parameters:

index (int) – a numeric index used to refer to this Property instance, a default of 1 is used
key (str) – the schrodinger.structure.Structure property key to be optimized
name (str) – specify a name for the property, this name will be, for example used in any *log files, etc.
units (str) – enter the units that the property is in, for example eV, nm, etc.
minimax (str) – to minimize or maximize this property then set this option to the class constants MIN or MAX
target (float) – if instead of maximizing or minimizing the property, the genetic optimization is supposed to handle a specific value then enter that value using this option.
comparator (str) – specify here how the target value and computed values are to be compared, i.e. either the class constants EQUALS for =, GREATER_THAN for >, or LESS_THAN for <.
error (float) – if equality to a target value has been specified then this option allows the user to control the error bounds of the target value, if not specified then a default of 10% of the specified target value will be used.
weight (float) – specify the weight to use for this property, if the genetic optimization is to be run on several properties then the weight allows the user to bias the solution. This option can also be used to control a situation where more than a single property is desired and where those properties are quantified using different physical units such that the numbers might be orders of magnitude apart from one another, for example comparing eV and nm. A default of 1.0 is used.
positive (bool) – True if this property can only take on positive values, for example as in the area of a surface, False otherwise, for example as in temperature in Celcius. The default is False.
summarize (bool) – if True then print a summary of this property, False otherwise
class_kwargs (dict or None) – contains kwargs for class based evaluation of this property

setClassKwargs(class_kwargs)¶

Set the class kwargs.

Parameters:: class_kwargs (dict or None) – contains kwargs for class based evaluation of this property

parsePropertyString(property_string)¶

Parse the attributes of this class from a string representation of the property specifications. For example, ‘index=1 key=r_matsci_Reduction_Potential_(eV) name=reduction units=eV target=1.28 comparator=eq error=0.05 weight=0.5’ or ‘index=2 key=r_matsci_Oxidation_Potential_(eV) name=oxidation units=eV minimax=max weight=2.5’

Parameters:

property_string (str) – the string representation of the property specifications

Raises:

PropertySyntaxError – if there is something wrong with the property syntax
UnknownPropertySuboptionError – if an unknown property suboption is found

checkProperty()¶: Check this property instance.

isScriptProperty()¶

Return True if this property is a script property, False otherwise.

Return type:: bool
Returns:: return True if this property is a script property, False otherwise

static getPropertyStrings(property_lists)¶

Return property strings from the given property lists.

Parameters:: property_lists (list) – contains lists of property specifications
Return type:: list
Returns:: contains string representations of the property specifications

static getRelPath(file_name)¶

Return the relative path to the given file name.

Parameters:: file_name (str) – the file name
Return type:: str
Returns:: the relative path to the file

static getKwargs(property_string, option_substrings, add_relative_paths=None)¶

Return kwargs of the given property options from the given property string.

Parameters:

property_string (str) – the string representation of the property specifications, containing options as ‘<option_substring>=<value>’
option_substrings (list or str) – contains the option substrings for the needed values, a single occurence or list of occurences may be passed
add_relative_paths (list) – contains options for which relative paths should be added, such relative paths might be needed for correctly parallelizing the evaluation stage of the genetic optimization as they will be needed to copy otherwise shared files into local subdirectories

Return type:

dict, str, or None

Returns:

the extracted dictionary of kwargs or single kwarg depending on the input option_substrings or None if nothing is found

static rmKwargs(property_string, option_substrings)¶

Return a copy of the given property string with all of the given property option substrings removed.

Parameters:

property_string (str) – the string representation of the property specifications, containing options as ‘<option_substring>=<value>’
option_substrings (list) – contains the option substrings to be removed

Return type:

str

Returns:

the string representation of the property specifications less the options substrings that were to be removed

static addKwargs(property_string, kwargs)¶

Add the given options to the given property string.

Parameters:

property_string (str) – the string representation of the property specifications, containing options as ‘<option_substring>=<value>’
kwargs (dict) – key-value option pairs to add to the property string

Return type:

str

Returns:

the string representation of the property specifications containin the new options

static getCustomInfo(property_string, name, tpp=1, host_str=None)¶

Return a PropertyInfo.

Parameters:

property_string (str) – the string representation of the property specifications, containing options as ‘<option_substring>=<value>’
name (str) – the property name
tpp (int) – the threads per process
host_str (str) – the host string, for example ‘localhost:4’

Raises:

RuntimeError – if there is a problem

Return type:

PropertyInfo

Returns:

the property information

schrodinger.application.matsci.genetic_optimization.set_title_to_stoichiometry(astructure, toappend=None, separation='.')¶

Set the structure title to be the stoichiometry of the structure.

Parameters:

astructure (schrodinger.structure.Structure) – the structure
toappend (str) – a string to append to the stoichiometry
separation (str) – used to separate the stoichiometry and the toappend str

class schrodinger.application.matsci.genetic_optimization.StructureGenome¶

Bases: GenomeBase

Manage a genome. The genome, aka chromosome, is the solution to the problem trying to be solved via genetic optimization. It is referred to as being composed of genes that are manipulated by the crossover and mutation operators. In our genetic optimization module this genome is basically just a schrodinger.structure.Structure object.

__init__()¶: Create an instance.

copy(genome)¶

Copy the current genome to the provided genome.

Parameters:: genome (StructureGenome) – a new genome instance to which to copy the current genome

clone()¶

Clone the current genome.

Return type:: StructureGenome
Returns:: genome

updateStructureProperties(index, generation)¶

Update some structure properties.

Parameters:

index (int) – the index of this individual
generation (int) – this generation

resetParentProperties()¶: Reset the crossover and mutation parent structure properties.

removeProperties()¶: Remove some structure properties.

optimizeGeometry()¶: Optimize the geometry of this genome’s structure using OPLS.

addPreviousFreezerFile(freezer_file)¶

Add the given file to the list of previous freezer files.

Parameters:: freezer_file (str) – the name of the file to be added

evaluate(**args)¶

Evaluate the score of this individual.

Parameters:: args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

evaluator¶

initializator¶

mutator¶

crossover¶

internalParams¶

score¶

fitness¶

schrodinger.application.matsci.genetic_optimization.from_initial_population(genome, **args)¶

Draw a unique genome from the initial population.

Parameters:

genome (StructureGenome) – a genome
args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

schrodinger.application.matsci.genetic_optimization.get_num_simple_bonds(astructure)¶

Return the number of simple bonds in the provided structure. The definition of a simple bond follows from that used in the reaction channel module and is an acyclic single order bond that may involve a hydrogen atom.

Parameters:: astructure (schrodinger.structure.Structure) – the structure for which to get the number of simple bonds
Return type:: int
Returns:: the number of simple bonds

schrodinger.application.matsci.genetic_optimization.combine_two_structures(astructure, bstructure, offset=10.0)¶

Combine two structure objects into a single structure object using somewhat arbitrary placement.

Parameters:

astructure (schrodinger.structure.Structure) – the first of the structures to be combined
bstructure (schrodinger.structure.Structure) – the second of the structures to be combined
offset (float) – the final distance between the structures will be the sum of the molecular VDW radii plus this offset in Angstrom

Return type:

schrodinger.structure.Structure

Returns:

the combined structure object

schrodinger.application.matsci.genetic_optimization.bond_crossover(genome, **args)¶

Perform a crossover operation by swapping molecular fragments at two randomly choosen bonds, i.e. a double displacement reaction channel.

Parameters:

genome (StructureGenome) – a genome
args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

Return type:

tuple

Returns:

tuple containing the sister and brother StructureGenome

schrodinger.application.matsci.genetic_optimization.get_element_mutator_dict(astructure)¶

Return a dictionary where the keys contain the indicies of the mutatable atoms and the values contain those elements that the keyed atom may be mutated to.

Parameters:: astructure (schrodinger.structure.Structure) – the structure to be mutated
Return type:: dict
Returns:: keys are atom indicies of those atoms that are mutatable and values are those elements that the atom can be mutated to

schrodinger.application.matsci.genetic_optimization.get_isoelectronic_mutator_indicies(astructure)¶

Return a list of atom indicies that can be mutated by the isoelectronic mutator.

Parameters:: astructure (schrodinger.structure.Structure) – the structure to be mutated
Return type:: list
Returns:: mutatable indicies

schrodinger.application.matsci.genetic_optimization.get_child_like_parent(parent_st, children_sts, definition)¶

Return the child structure that is most like the provided parent.

Parameters:

parent_st (schrodinger.structure.Structure) – the parent structure
children_sts (list of schrodinger.structure.Structure) – the children structures
definition (two-element list) – each sublist contains two atom indicies describing the reactive bonds in parent and fragment structures which created the children

Return type:

schrodinger.structure.Structure

Returns:

the sought child structure

schrodinger.application.matsci.genetic_optimization.elemental_mutator(genome, **args)¶

Perform a random elemental mutation to an element in the same column (as known as group) of the periodic table. Note that hydrogen and the halogens are considered to belong to the same column.

Parameters:

genome (StructureGenome) – a genome
args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

Return type:

int

Returns:

the number of mutations applied, appears to never be used in PyEvolve

schrodinger.application.matsci.genetic_optimization.fragment_mutator(genome, **args)¶

Randomly mutate the genome by swapping a molecular fragement on one side of a bond by a similar fragment from a library.

Parameters:

genome (StructureGenome) – a genome
args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

Return type:

int

Returns:

the number of mutations applied, appears to never be used in PyEvolve

schrodinger.application.matsci.genetic_optimization.isoelectronic_mutator(genome, **args)¶

Perform a random isoelectronic mutation from the following sets of series CH3X, NH2X, OHX, and FX, CH2XY, NHXY, OXY, and CHXYZ and NXYZ, where X, Y, and Z are non-H-bonds.

Parameters:

genome (StructureGenome) – a genome
args (dict) – dictionary of genetic optimization parameters created and used by pyevolve

Return type:

int

Returns:

the number of mutations applied, appears to never be used in PyEvolve

schrodinger.application.matsci.genetic_optimization.get_loggable_float(afloat, num_decimal='%.2f', field_width=10)¶

Return a float as a string with the specified format.

Parameters:

afloat (float) – a float to convert to a string
num_decimal (str) – the format of the string representation
field_width (int) – the field width of the final string

Return type:

str

Returns:

the float as a string

schrodinger.application.matsci.genetic_optimization.uniquify_titles_callback(ga_obj)¶

Callback to uniquify titles of the individuals.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.prepare_next_generation_dirs_callback(ga_obj)¶

Callback to update the generation property of the genomes and to create a subdirectory to hold the next series of evaluations.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.manage_skips_callback(ga_obj)¶

Callback to manage skips in the evaluation.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.manage_failures_callback(ga_obj)¶

Callback to manage failures in the evaluation.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.logging_summary_callback(ga_obj)¶

Callback to log progress.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.molecule_history_callback(ga_obj)¶

Callback to append all structures from all generations to individual log files.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization

schrodinger.application.matsci.genetic_optimization.first_property(ga_obj)¶

Terminate when the first property has been matched.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization
Return type:: bool
Returns:: True to terminate, False otherwise

schrodinger.application.matsci.genetic_optimization.all_properties(ga_obj)¶

Terminate when all properties have been matched.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization
Return type:: bool
Returns:: True to terminate, False otherwise

schrodinger.application.matsci.genetic_optimization.unproductive(ga_obj)¶

Terminate if the maximum number of unproductive generations has been reached.

Parameters:: ga_obj (GSimpleGA.GSimpleGA) – the entire current state of the genetic optimization
Return type:: bool
Returns:: True to terminate, False otherwise

class schrodinger.application.matsci.genetic_optimization.CheckInput¶

Bases: object

Manage checking user input.

checkMaeFile(input_file, logger=None)¶

Check that a file exists and is *mae.

Parameters:

input_file (str) – the name of the input file
logger (logging.Logger) – output logger

checkOperators(operators, logger=None)¶

Check the operators.

Parameters:

operators (list) – contains tuples of the operator functions and their weights
logger (logging.Logger) – output logger

checkRates(crossover_rate, mutation_rate, logger=None)¶

Check the specified rates of crossover and mutation.

Parameters:

crossover_rate (float) – the rate of crossover as a percentage
mutation_rate (float) – the rate of mutation as a percentage
logger (logging.Logger) – output logger

checkInitialPopulation(initial_population, crossover_names, mutator_names, crossover_rate, mutation_rate, no_open_shell, logger=None)¶

Check the initial population.

Parameters:

initial_population (list) – the initial population of schrodinger.structure.Structure
crossover_names (list) – contains the function names of the crossover operators to be used
mutator_names (list) – contains the function names of the mutation operators to be used
crossover_rate (float) – the rate of crossover
mutation_rate (float) – the rate of mutation
no_open_shell (bool) – if True then check for open shell structures otherwise do not
logger (logging.Logger) – output logger

checkPopulationParam(population, num_structures_given, logger=None)¶

Check the population parameter.

Parameters:

population (int) – the size of the population to use in the genetic optimization
num_structures_given (int) – the number of structures provided to the genetic optimization
logger (logging.Logger) – output logger

checkFragmentLibs(fragment_libs, logger=None)¶

Check the specified fragment libraries.

Parameters:

fragment_libs (list) – strings specifying fragment libraries to be used
logger (logging.Logger) – output logger

Return type:

list

Returns:

valid user provided fragment files

checkProperties(properties, logger=None)¶

Check the list of properties.

Parameters:

properties (list) – contains Property instances
logger (logging.Logger) – output logger

checkGenerations(generations, logger=None)¶

Check the specified number of generations.

Parameters:

generations (int) – the number of generations
logger (logging.Logger) – output logger

checkSelection(selection, logger=None)¶

Check the specified selection protocol.

Parameters:

selection (str) – the selection protocol to use.
logger (logging.Logger) – output logger

checkTournamentSize(tournament_size, population, logger=None)¶

Check the specified tournament size.

Parameters:

tournament_size (int) – the size of tournament to use in tournament based selection
population (int) – the size of population to use
logger (logging.Logger) – output logger

checkTerminationParams(terminators, num_unproductive, logger=None)¶

Check the termination parameters.

Parameters:

terminators (list) – the list of terminators to use
num_unproductive (int) – used when the unproductive termination option is active, it is the generation number on which to exit if the score hasn’t improved
logger (logging.Logger) – output logger

Return type:

list and int

Returns:

valid terminators and valid num_unproductive

checkScaling(scaling, properties, logger=None)¶

Check the scaling.

Parameters:

scaling (str) – the scaling protocol to use in the genetic optimization
properties (list) – the properties to be optimized
logger (logging.Logger) – output logger

checkElitism(elitism, population, logger=None)¶

Check the elitism.

Parameters:

elitism (int) – the number of elite individuals to use
population (int) – the size of population to use
logger (logging.Logger) – output logger

checkConformationalSearch(conformational_search, logger=None)¶

Check the conformational search.

Parameters:

conformational_search (bool or str) – specifies whether a conformational search is to be performed, if a string is given specifies a file used to set options
logger (logging.Logger) – output logger

checkFreezers(freezers, pop_size, input_size, logger=None)¶

Check the freezers.

Parameters:

freezers (list) – collection of freezers to use
pop_size (int) – the size of the population
input_size (int) – the number of structures given
logger (logging.Logger) – output logger

Return type:

list

Returns:

collection of freezers to use

checkInoculate(inoculate, logger=None)¶

Check the inoculate.

Parameters:

inoculate (list) – circumstances in which to inoculate
logger (logging.Logger) – output logger

schrodinger.application.matsci.genetic_optimization.print_bad_jobs(all_bad_jobs, logger, bad_type='skip')¶

Log bad jobs, i.e. skips and failures.

Parameters:

all_bad_jobs (dict) – a collection of bad subjobs, keys are genetic optimization generation and values are a list of Skip or Failure objects for bad subjobs
logger (logging.Logger) – output logger
bad_type (str) – specifies either ‘skip’ or ‘fail’ type

class schrodinger.application.matsci.genetic_optimization.GeneticOptimization(initial_population, properties, structure_score_threshold=-50.0, eval_kwargs={}, crossovers=None, mutators=None, fragment_libs=['optoelectronics'], script_evaluator=None, generations=10, population=8, crossover_rate=90.0, mutation_rate=90.0, selection='roulette_wheel', tournament_size=2, terminators=['unproductive', 'all_properties'], num_unproductive=6, scaling='sigma_truncation', elitism=1, random_seed=None, no_minimize=False, file_base_name='genopt', no_open_shell=False, props_to_remove=None, jobbe=None, conformational_search=False, freezers=['remainder', 'previous'], inoculate=['no_child', 'bad_structure'], class_evaluators=None, logger=None)¶

Bases: object

Manage the genetic optimization.

MSGWIDTH = 80¶

__init__(initial_population, properties, structure_score_threshold=-50.0, eval_kwargs={}, crossovers=None, mutators=None, fragment_libs=['optoelectronics'], script_evaluator=None, generations=10, population=8, crossover_rate=90.0, mutation_rate=90.0, selection='roulette_wheel', tournament_size=2, terminators=['unproductive', 'all_properties'], num_unproductive=6, scaling='sigma_truncation', elitism=1, random_seed=None, no_minimize=False, file_base_name='genopt', no_open_shell=False, props_to_remove=None, jobbe=None, conformational_search=False, freezers=['remainder', 'previous'], inoculate=['no_child', 'bad_structure'], class_evaluators=None, logger=None)¶

Create an instance.

Parameters:

initial_population (list) – the initial population of schrodinger.structure.Structure
properties (list of Property) – the properties to be optimized, including structural properties as well as more physical calculable observables
structure_score_threshold (float) – if structure-based properties are being sought and if the base evaluator will be used then subjobs on structures with structure scores below this value will not be launched but rather such structures treated as skips
eval_kwargs (dict) – a dictionary that will be available in all evaluator functions
crossovers (list) – contains two-element tuples each of which holds a crossover operator to be used in the optimization along with a weight
mutators (list) – contains two-element tuples each of which holds a mutation operator to be used in the optimization along with a weight
fragment_libs (list) – strings specifying fragment libraries to be used, can be either module constants from FRAGMENT_LIBS.keys() (or ALL if all of those are desired) or the names of Maestro files (including the file extensions) containing fragments collected by the user
script_evaluator (method) – the evaluator function to be called to score individuals during the optimization, takes a StructureGenome and returns a JobDJ
generations (int) – the number of generations for which to run the optimization
population (int) – the population size to use in the optimization, can be less-than-or-equal-to the length of initial_population
crossover_rate (float) – the rate of crossover as a percentage
mutation_rate (float) – the rate of mutation as a percentage
selection (str) – the selection protocol used to select individuals to the gene pool for the upcoming generation
tournament_size (int) – the size of tournament to use if using tournament based selection, unused if a tournament based selection is not being used
terminators (list) – list of strings that specify the termination protocols to be used to terminate the optimization, typically more than one is specified only if the unproductive protocol is being used
num_unproductive (int) – if the unproductive protocol is being used to terminate the optimization then this integer specifies how many unproductive cycles are allowed before terminating, unused if a different termination protocol is used
scaling (str) – specifies the scaling protocol to use, scaling scales the raw scores of the individuals to produce fitness scores to ease selection in cases where raw scores are nearly equal
elitism (int) – specify the number of elite individuals guaranteed to be added to the gene pool for the upcoming generation, zero disables elitism
random_seed (None or int) – the random seed, if None then system time will be used
no_minimize (bool) – specify that the offspring structures generated by the crossover and mutation operators not be geometry optimized prior to selection
file_base_name (str) – base name to use for output and generation log files
no_open_shell (bool) – if True then do not allow the processing of open shell molecules, False otherwise
props_to_remove (list) – a list of structure property keys to be removed prior to the evaluation stage
jobbe (schrodinger.job.jobcontrol._Backend) – the jobcontrol backend of the driver job
conformational_search (bool or str) – specifies whether a Macromodel conformational search will be performed prior to evaluation, when a string it specifies a simplified Macromodel input file containing extra options
freezers (list) – a collection of freezers containing structures that are used to swap out individuals from the population
inoculate (list) – the list of circumstances under which to use the structure freezers
class_evaluators (dict) – keys are the evaluator classes to be called to score individuals during the optimization, each must inherit ClassEvaluator, values are lists of Property to be passed to the class evaluator
logger (logging.Logger) – output logger

setRootLoggerForPyEvolve()¶: Set up the root logger for PyEvolve.

setOperatorNames()¶: Set the operator names.

checkInputParams()¶: Check the input parameters.

printProperties()¶: Log the set of sought properties and their details.

printParams()¶: Log the parameters.

initializeGenome()¶

Initialize a genome.

Return type:: StructureGenome
Returns:: a genome

initializeGA(genome)¶

Initialize the genetic optimization.

Parameters:: genome (StructureGenome) – a genome

setMonomerGrowAtoms()¶: Set the monomer grow atoms using the mark monomer module convention rather than the polymer builder module convention.

runIt()¶: Run the components of the genetic optimization.

schrodinger.application.matsci.genetic_optimization.get_output_file_name(basename)¶

Get the output file name from the basename.

Parameters:: basename (str) – base name to use
Return type:: str
Returns:: output_file_name, name of output file

schrodinger.application.matsci.genetic_optimization.get_generation_log_file_name(basename, generation)¶

Get the generation log file name.

Parameters:

basename (str) – base name to use
generation (int) – the generation

Return type:

str

Returns:

generation_log_file_name, name of generation log file

schrodinger.application.matsci.genetic_optimization.get_structure_score(astructure, properties, conformational_search, seed=None, this_random=None)¶

Return the structure score for the provided structure.

Parameters:

astructure (schrodinger.structure.Structure) – the structure to score
properties (list of Property) – the properties used in scoring
conformational_search (bool or str) – specifies whether a Macromodel conformational search will be performed prior to evaluation, when a string it specifies a simplified Macromodel input file containing extra options
seed (int or None) – random seed used in conformational search or None if conformational search is not being done
this_random (numpy.random.RandomState or None) – random state, if None use the module constant

Return type:

float

Returns:

the structure score

schrodinger.application.matsci.genetic_optimization.structure_evaluator(genome, this_random=None)¶

This is the structure evaulator.

Parameters:

genome (StructureGenome) – a genome
this_random (numpy.random.RandomState or None) – random state, if None use the module constant

Return type:

float

Returns:

the score for this individual

schrodinger.application.matsci.genetic_optimization.base_evaluator(genome)¶

This is the base evaulator used to wrap all other evaluators.

Parameters:: genome (StructureGenome) – a genome
Return type:: float
Returns:: the score for this individual

schrodinger.application.matsci.genetic_optimization.optoelectronics_evaluator(genome)¶

Run an optoelectronics job.

Parameters:: genome (StructureGenome) – a genome
Return type:: JobDJ
Returns:: the JobDJ object for this individual, it is run in the base evaluator

schrodinger.application.matsci.genetic_optimization.apply_uniform_operator_weights(operators)¶

Set the operator weights uniformly.

Parameters:: operators (list) – a list of two-element tuples, each tuple contains first an operator function and second a weight
Return type:: list
Returns:: list of two-element tuples of operators and uniform weights

schrodinger.application.matsci.genetic_optimization.structure_is_open_shell(astructure, ignore_charge=True)¶

Return True if the provided structure is open shell, i.e. has an odd number of electrons.

Parameters:

astructure (schrodinger.structure.Structure) – the structure in question
ignore_charge (bool) – if True then ignore any structure.formal_charge settings

Return type:

bool

Returns:

True if the provided structure is open shell, False otherwise

schrodinger.application.matsci.genetic_optimization.get_element_histogram(astructure)¶

Return a dictionary where keys are elements and values are the numbers of atoms of a given element.

Parameters:: astructure (schrodinger.structure.Structure) – the structure in question
Return type:: dict
Returns:: dictionary with element histogram, keys are elements (strs) and values are numbers (ints)

schrodinger.application.matsci.genetic_optimization.remove_basename_ext(stoich_ext)¶

Remove the basename extension from the given string and return the remainder which is the stoichiometry. Do this instead of having to recompute the stoichiometry which can be expensive.

Parameters:: stoich_ext (str) – contains the stoichiometry and basename extension
Return type:: str
Returns:: stoichiometry

schrodinger.application.matsci.genetic_optimization.get_low_energy_conformers(astructure_in, macromodel_options_file=None, remove_files=False, overwrite=False, seed=None, this_random=None, host_str=None)¶

Return the lowest energy conformers from a Macromodel conformational search.

Parameters:

astructure_in (schrodinger.structure.Structure) – the structure to search for conformations
macromodel_options_file (str or None) – the name of a simplified Macromodel input file that contains any options to use in addition to those used by default in a conformational search or None if there are none and you just want to use the defaults
remove_files (bool) – if the job is successful, specifies whether to remove all files created for it after it finishes
overwrite (bool) – if True then the coordinates of the input structure will be overwritten by those of the lowest energy conformer and that structure alone returned by this function
seed (int or None) – used to seed the random number generator used in the Macromodel conformational search, should be in CONF_SEARCH_SEED_RANGE, if None then if a CONFSEARCH_SEED has been specified in macromodel_options_file it will be used, otherwise a random int in CONF_SEARCH_SEED_RANGE will be used
this_random (numpy.random.RandomState or None) – random state, if None use the module constant
host_str (str) – the host string, for example ‘localhost:4’

Return type:

list of schrodinger.structure.Structure, int

Returns:

the structures of the lowest energy conformers sorted by increasing energy and the seed used in the conformational search (same as input if input was given either as seed or in macromodel_options_file)

schrodinger.application.matsci.genetic_optimization.get_random_structure(structure_libs, tries_from_libs=3, structure_score_threshold=None, properties=None, conformational_search=False, seed=None, this_random=None)¶

From the given dictionary of libraries return a random structure.

Parameters:

structure_libs (dict) – keys are strings specifying the types of libraries to be used and can be module constants from FREEZER_CHOICES.keys(), values are lists of libraries by type and can be either module constants from FRAGMENT_LIBS.keys(), ALL, or the names of Maestro files (including the file extensions)
tries_from_libs (int) – the number of times to try before giving up
structure_score_threshold (float or None) – specifies that a structure with a structure score greater-than-or-equal-to this threshold is sought, the best of the considered structures will be returned and will contain several structure properties related to the scoring
properties (list of Property or None) – the properties used in structure scoring
conformational_search (bool or str) – specifies whether a Macromodel conformational search will be performed prior to evaluation, when a string it specifies a simplified Macromodel input file containing extra options
seed (int or None) – if not None specifies that random should be reseeded with the given value
this_random (numpy.random.RandomState or None) – random state, if None use the module constant

Return type:

schrodinger.structure.Structure or None

Returns:

the random structure or None if one couldn’t be found

schrodinger.application.matsci.genetic_optimization.get_freezer_structure(structure_libs, tries_from_libs=3, structure_score_threshold=None, properties=None, conformational_search=False, inoculate='no_child', crossover_applied=None, mutation_applied=None, basename_ext=None, seed=None, this_random=None)¶

Return a random structure from the freezer and update that structure’s properties.

Parameters:

structure_libs (dict) – keys are strings specifying the types of libraries to be used and can be module constants from FREEZER_CHOICES.keys(), values are lists of libraries by type and can be either module constants from FRAGMENT_LIBS.keys(), ALL, or the names of Maestro files (including the file extensions)
tries_from_libs (int) – the number of times to try before giving up
structure_score_threshold (float or None) – specifies that a structure with a structure score greater-than-or-equal-to this threshold is sought, the best of the considered structures will be returned and will contain several structure properties related to the scoring
properties (list of Property or None) – the properties used in structure scoring
conformational_search (bool or str) – specifies whether a Macromodel conformational search will be performed prior to evaluation, when a string it specifies a simplified Macromodel input file containing extra options
inoculate (str) – specify the reason for drawing from the freezer, which is an inoculate option from INOCULATE_CHOICES
crossover_applied (str or None) – specify the intended crossover operator or None if there isn’t to be one
mutation_applied (str or None) – specify the intended mutation operator or None if there isn’t to be one
basename_ext (str or None) – specify an extension to append to the stoichiometry which is used to set the title of the returned structure
seed (int or None) – if not None specifies that random should be reseeded with the given value
this_random (numpy.random.RandomState or None) – random state, if None use the module constant

Return type:

schrodinger.structure.Structure or None

Returns:

the random structure or None if one couldn’t be found