aizynthfinder.utils package¶

Submodules¶

aizynthfinder.utils.bonds module¶

Module containing a class to identify broken focussed bonds

class aizynthfinder.utils.bonds.BrokenBonds(focussed_bonds)¶

Bases: object

A class to keep track of focussed bonds breaking in a target molecule.

Parameters:: focussed_bonds (Sequence[Sequence[int]]) – A list of focussed bond pairs. The bond pairs are represented as tuples of size 2. These bond pairs exist in the target molecule’s atom bonds.

aizynthfinder.utils.bonds.sort_bonds(bonds)¶

Parameters:: bonds (Sequence[Sequence[int]])
Return type:: List[Tuple[int, int]]

aizynthfinder.utils.exceptions module¶

Module containing custom exception classes

exception aizynthfinder.utils.exceptions.CostException¶

Bases: Exception

Exception raised by molecule cost classes

exception aizynthfinder.utils.exceptions.ExternalModelAPIError¶

Bases: Exception

Custom error type to signal failure in External model

exception aizynthfinder.utils.exceptions.MoleculeException¶

Bases: Exception

An exception that is raised by molecule class

exception aizynthfinder.utils.exceptions.NodeUnexpectedBehaviourException¶

Bases: Exception

Exception that is raised if the tree search is behaving unexpectedly.

exception aizynthfinder.utils.exceptions.PolicyException¶

Bases: Exception

An exception raised by policy classes

exception aizynthfinder.utils.exceptions.RejectionException¶

Bases: Exception

An exception raised if a retro action should be rejected

exception aizynthfinder.utils.exceptions.ScorerException¶

Bases: Exception

Exception raised by scoring classes

exception aizynthfinder.utils.exceptions.StockException¶

Bases: Exception

An exception raised by stock classes

exception aizynthfinder.utils.exceptions.TreeAnalysisException¶

Bases: Exception

Exception raised when analysing trees

aizynthfinder.utils.files module¶

Module containing routines to work with files and processes.

aizynthfinder.utils.files.read_datafile(filename)¶

Read aizynth output from disc in either .hdf5 or .json format

Parameters:: filename (Union[str, Path]) – the path to the data
Returns:: the loaded data
Return type:: pd.DataFrame

aizynthfinder.utils.files.save_datafile(data, filename)¶

Save the given data to disc in either .hdf5 or .json format

Parameters:

data (pd.DataFrame) – the data to save
filename (Union[str, Path]) – the path to the data

Return type:

None

aizynthfinder.utils.files.cat_hdf_files(input_files, output_name, trees_name=None)¶

Concatenate hdf5 or json datafiles

Parameters:

input_files (List[str])
output_name (str)
trees_name (Optional[str])

Return type:

None

aizynthfinder.utils.files.cat_datafiles(input_files, output_name, trees_name=None)¶

Concatenate hdf5 or json datafiles

if tree_name is given, will take out the trees column from the tables and save it to a gzipped-json file.

Parameters:

input_files (List[str]) – the paths to the files to concatenate
output_name (str) – the name of the concatenated file
trees_name (Optional[str]) – the name of the concatenated trees

Return type:

None

aizynthfinder.utils.files.split_file(filename, nparts)¶

Split the content of a text file into a given number of temporary files

Parameters:

filename (str) – the path to the file to split
nparts (int) – the number of parts to create

Returns:

list of filenames of the parts

Return type:

List[str]

aizynthfinder.utils.files.start_processes(inputs, log_prefix, cmd_callback, poll_freq=5)¶

Start a number of background processes and wait for them to complete.

The standard output and standard error is saved to a log file.

The command to start for each process is given by the cmd_callback function that takes the index of the process and an item of the input as arguments.

Parameters:

inputs (Sequence[Any]) – a sequence of input to the processes
log_prefix (str) – the prefix to the log file of each processes
cmd_callback (Callable) – function that creates the process commands
poll_freq (int) – the polling frequency for checking if processes are completed

Return type:

None

aizynthfinder.utils.image module¶

This module contains a collection of routines to produce pretty images

aizynthfinder.utils.image.molecule_to_image(mol, frame_color, size=300)¶

Create a pretty image of a molecule, with a colored frame around it

Parameters:

mol (Molecule) – the molecule
frame_color (PilColor) – the color of the frame
size (int) – the size of the image

Returns:

the produced image

Return type:

PilImage

aizynthfinder.utils.image.molecules_to_images(mols, frame_colors, size=300)¶

Create pretty images of molecules with a colored frame around each one of them.

The molecules will be resized to be of similar sizes.

Parameters:

mols (Sequence[Molecule]) – the molecules
frame_colors (Sequence[PilColor]) – the color of the frame for each molecule
size (int) – the sub-image size

Returns:

the produced images

Return type:

List[PilImage]

aizynthfinder.utils.image.crop_image(img, margin=20)¶

Crop an image by removing white space around it

Parameters:

img (PilImage) – the image to crop
margin (int) – padding, defaults to 20

Returns:

the cropped image

Return type:

PilImage

aizynthfinder.utils.image.draw_rounded_rectangle(img, color, arc_size=20)¶

Draw a rounded rectangle around an image

Parameters:

img (PilImage) – the image to draw upon
color (PilColor) – the color of the rectangle
arc_size (int) – the size of the corner, defaults to 20

Returns:

the new image

Return type:

PilImage

aizynthfinder.utils.image.save_molecule_images(molecules, frame_colors, size=300)¶

Create images of a list of molecules and save them to disc a globally managed folder.

Parameters:

molecules (Sequence[Molecule]) – the molecules to save as images
frame_colors (Sequence[PilColor]) – the color of the frame around each image
size (int) – the sub-image size for each molecule

Returns:

the filename of the created images

Return type:

Dict[Molecule, str]

aizynthfinder.utils.image.make_visjs_page(filename, molecules, reactions, edges, frame_colors, hierarchical=False)¶

Create HTML code of a bipartite graph of molecules and reactions using the vis.js network library.

Package the created HTML page and all images as tar-ball.

Parameters:

filename (str) – the basename of the archive
molecules (Sequence[Molecule]) – the molecules nodes
reactions (Sequence[FixedRetroReaction]) – the reaction nodes
edges (Union[Sequence[Tuple[Any, Any]], nx.digraph.OutEdgeView]) – the edges of the graph
frame_colors (Sequence[PilColor]) – the color of the frame around each image
hierarchical (bool) – if True, will produce a hierarchical layout

Return type:

None

class aizynthfinder.utils.image.RouteImageFactory(route, in_stock_colors=None, show_all=True, margin=100)¶

Bases: object

Factory class for drawing a route

Parameters:

route (StrDict) – the dictionary representation of the route
in_stock_colors (FrameColors) – the colors around molecules, defaults to {True: “green”, False: “orange”}
show_all (bool) – if True, also show nodes that are marked as hidden
margin (int) – the margin between images

aizynthfinder.utils.loading module¶

Module containing routine to dynamically load a class from a specification

aizynthfinder.utils.loading.load_dynamic_class(name_spec, default_module=None, exception_cls=<class 'ValueError'>)¶

Load an object from a dynamic specification.

The specification can be either:: ClassName, in-case the module name is taken from the default_module argument
or: package_name.module_name.ClassName, in-case the module is taken as package_name.module_name

Parameters:

name_spec (str) – the class specification
default_module (Optional[str]) – the default module
exception_cls (Any) – the exception class to raise on exception

Returns:

the loaded class

Return type:

Any

aizynthfinder.utils.logging module¶

Module containing routines to setup proper logging

aizynthfinder.utils.logging.logger()¶

Returns the logger that should be used by all classes

Returns:: the logger object
Return type:: Logger

aizynthfinder.utils.logging.setup_logger(console_level, file_level=None)¶

Setup the logger that should be used by all classes

The logger configuration is read from the logging.yml file.

Parameters:

console_level (int) – the level of logging to the console
file_level (int | None) – the level of logging to file, if not set logging to file is disabled, default to None

Returns:

the logger object

Return type:

Logger

aizynthfinder.utils.math module¶

Module containing diverse math functions, including neural network-related functions.

aizynthfinder.utils.math.dense_layer_forward_pass(x, weights, bias, activation)¶

Forward pass through a dense neural network layer. :param x: layer input :param weights: layer weights :param bias: layer bias :param activation: layer activation function :return: the layer output

Parameters:

x (ndarray)
weights (ndarray)
bias (ndarray)
activation (Callable)

Return type:

ndarray

aizynthfinder.utils.math.rectified_linear_unit(x)¶

ReLU activation function

Parameters:: x (ndarray)
Return type:: ndarray

aizynthfinder.utils.math.sigmoid(x)¶

Sigmoid activation function

Parameters:: x (ndarray)
Return type:: ndarray

aizynthfinder.utils.math.softmax(x)¶

Compute softmax values for each sets of scores in x.

Parameters:: x (ndarray)
Return type:: ndarray

aizynthfinder.utils.models module¶

Module containing helper routines for using Keras, Tensorflow and Onnx models

aizynthfinder.utils.models.load_model(source, key, use_remote_models)¶

Load model from a configuration specification.

If use_remote_models is True, tries to load:

A Tensorflow server through gRPC
A Tensorflow server through REST API
A local Keras model

otherwise it just loads the local model.

Parameters:

source (str) – if fallbacks to a local model, this is the filename
key (str) – when connecting to Tensorflow server this is the model name
use_remote_models (bool) – if True will try to connect to remote model server

Returns:

a model object with a predict object

Return type:

Union[‘LocalKerasModel’, ‘LocalOnnxModel’, ‘ExternalModelViaGRPC’, ‘ExternalModelViaREST’]

class aizynthfinder.utils.models.LocalKerasModel(filename)¶

Bases: object

A keras policy model that is executed locally.

The size of the input vector can be determined with the len() method.

Variables:

model – the compiled model
output_size – the length of the output vector

Parameters:

filename (str) – the path to a Keras checkpoint file

predict(*args, **_)¶

Perform a forward pass of the neural network.

Parameters:

args (ndarray) – the input vectors
_ (ndarray)

Returns:

the vector of the output layer

Return type:

ndarray

class aizynthfinder.utils.models.LocalOnnxModel(filename)¶

Bases: object

An Onnx model that is executed locally.

The size of the input vector can be determined with the len() method.

Variables:

model – the compiled Onnx model
output_size – the length of the output vector

Parameters:

filename (str) – the path to a Onnx model checkpoint file

predict(*args, **_)¶

Perform a prediction run on the onnx model.

Parameters:

args (ndarray) – the input vectors
_ (ndarray)

Returns:

the vector of the output layer

Return type:

ndarray

class aizynthfinder.utils.models.ExternalModelViaREST(name)¶

Bases: object

A neural network model implementation using TF Serving via REST API.

Parameters:: name (str) – the name of model

predict(*args, **kwargs)¶

Get prediction from model.

If the keys in kwargs agree with the model input names, they will be used in stead of arg

Parameters:

args (ndarray) – the input vectors
kwargs (ndarray) – the named input vectors

Returns:

the vector of the output layer

Return type:

ndarray

class aizynthfinder.utils.models.ExternalModelViaGRPC(name)¶

Bases: object

A neural network model implementation using TF Serving via gRPC.

Parameters:: name (str) – the name of model

predict(*args, **kwargs)¶

Get prediction from model.

If the keys in kwargs agree with the model input names, they will be used in stead of arg

Parameters:

args (ndarray) – the input vectors
kwargs (ndarray) – the named input vectors

Returns:

the vector of the output layer

Return type:

ndarray

aizynthfinder.utils.mongo module¶

Module containing routines to obtain a MongoClient instance

aizynthfinder.utils.mongo.get_mongo_client(host='localhost', port=27017, user=None, password=None, tls_certs_path='')¶

A helper function to create and reuse MongoClient

The client is only setup once. Therefore if this function is called a second time with different parameters, it would still return the first client.

Parameters:

host (str) – the host
port (int) – the host port
user (str | None) – username, defaults to None
password (str | None) – password, defaults to None
tls_certs_path (str) – the path to TLS certificates if to be used, defaults to “”

Raises:

ValueError – if host and port is not given first time

Returns:

the MongoDB client

Return type:

MongoClient | None

aizynthfinder.utils.paths module¶

Module containing routines for returning package paths

aizynthfinder.utils.paths.package_path()¶

Return the path to the package

Return type:: str

aizynthfinder.utils.paths.data_path()¶

Return the path to the data directory of the package

Return type:: str

aizynthfinder.utils.sc_score module¶

Module containing the implementation of the SC-score model for synthetic complexity scoring.

class aizynthfinder.utils.sc_score.SCScore(model_path, fingerprint_length=1024, fingerprint_radius=2)¶

Bases: object

Encapsulation of the SCScore model

Re-write of the SCScorer from the scscorer package

The predictions of the score is made with a sanitized instance of an RDKit molecule

mol = Molecule(smiles="CCC", sanitize=True)
scscorer = SCScorer("path_to_model")
score = scscorer(mol.rd_mol)

The model provided when creating the scorer object should be pickled tuple. The first item of the tuple should be a list of the model weights for each layer. The second item of the tuple should be a list of the model biases for each layer.

Parameters:

model_path (str) – the filename of the model weights and biases
fingerprint_length (int) – the number of bits in the fingerprint
fingerprint_radius (int) – the radius of the fingerprint

forward(x)¶

Forward pass with dense neural network

Parameters:: x (ndarray)
Return type:: ndarray

aizynthfinder.utils.type_utils module¶

Module containing all types and type imports

aizynthfinder

Navigation

Related Topics

aizynthfinder.utils package¶

Submodules¶

aizynthfinder.utils.bonds module¶

aizynthfinder.utils.exceptions module¶

aizynthfinder.utils.files module¶

aizynthfinder.utils.image module¶

aizynthfinder.utils.loading module¶

aizynthfinder.utils.logging module¶

aizynthfinder.utils.math module¶

aizynthfinder.utils.models module¶

aizynthfinder.utils.mongo module¶

aizynthfinder.utils.paths module¶

aizynthfinder.utils.sc_score module¶

aizynthfinder.utils.type_utils module¶

Module contents¶