API Reference

This is the auto-generated reference for LipiDetective’s public API.

Models

class TransformerNetwork(config: dict[str, Any], output_attentions: bool = False)

forward(src: Tensor, tgt: Tensor) → Tensor | tuple[Tensor, list[Tensor]]

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

generate_mask(src: Tensor, tgt: Tensor) → tuple[Tensor, Tensor, Tensor]

predict(src: Tensor) → Tensor

predict_top_3(src: Tensor) → tuple[Tensor, Tensor]

predict_beam_decode(src: Tensor) → Tensor

predict_greedy(src: Tensor) → Tensor

return_encoder_embedding(src: Tensor) → Tensor

greedy_decode(encoder_output: Tensor, tgt: Tensor, memory_padding_mask: Tensor) → None

beam_decode(encoder_output: Tensor, tgt: Tensor, memory_padding_mask: Tensor) → None

get_attention_layers(src: Tensor, src_key_padding_mask: Tensor) → list[Tensor]

class ConvolutionalNetwork(config: dict[str, Any])

forward(x: Tensor) → Tensor

Forward pass of the convolutional network with three convolutional layers and pooling layers, followed by three fully connected linear layers.

Parameters:

x (torch.Tensor) – input tensor of features with shape (batch_size, 2, n_peaks+1). Dimension 1 is size 2 as the tensor contains the m/z and intensity values of each peak. Dimension 2 is size n_peaks + 1 as the measurement mode (-1 for negative and +1 for positive) and the precursor mass are added.

Returns:

output of the convolutional network with shape (batch_size, 3). Corresponds to the three masses of the lipid components (headgroup and two side chains) that are supposed to be predicted.

Return type:

torch.Tensor

calculate_fc1_size(len_input_spectrum: int) → int

Calculates the size of the output after the convolutional layers so that the size of the first fully connected layer can be set accordingly.

Parameters:

len_input_spectrum (int) – length of the input spectrum

Returns:

size of the in_features of the first fully connected layer

Return type:

int

class FeedForwardNetwork(config: dict[str, Any])

forward(x: Tensor) → Tensor

Define the computation performed at every call.

Should be overridden by all subclasses.

Note

Although the recipe for forward pass needs to be defined within this function, one should call the Module instance afterwards instead of this since the former takes care of running the registered hooks while the latter silently ignores them.

class RandomForest(config: dict[str, Any])

run() → None

get_spectrum_data(spectrum: Dataset) → list[Any]

use_single_classifier(train_features: list[Any], train_labels: list[Any], test_features: list[Any]) → tuple[Any, RandomForestClassifier]

plot_decision_tree(decision_tree: Any, name_file: str) → None

use_triple_classifier(train_features: list[Any], train_labels: list[Any], test_features: list[Any]) → tuple[list[list[Any]], RandomForestClassifier, RandomForestClassifier, RandomForestClassifier]

use_triple_regressor(train_features: list[Any], train_labels: list[Any], test_features: list[Any]) → tuple[list[list[Any]], RandomForestRegressor, RandomForestRegressor, RandomForestRegressor]

calculate_accuracy(prediction: Any, labels: list[Any], model: str, task: str) → str

check_classification_accuracy(prediction: object, label: object) → bool

check_regression_accuracy(prediction: Any, label: Any) → bool

write_output_to_file(statistics: dict[str, str]) → None

extract_info_dataset(group: Group | Dataset, val_lipids: list[str], train_set: list[Any], test_set: list[Any]) → None

extract_info_dataset_no_split(group: Group | Dataset, dataset: list[Any]) → None

extract_features_and_labels(dataset: list[Any]) → tuple[Any, Any]

prepare_data() → tuple[list[Any], list[Any], list[Any], list[Any]]

Workflow

class Trainer(config: dict[str, Any])

The trainer class creates the lightning module and executes the specified workflow. It handles processing and splitting of the dataset.

train_with_validation() → None

train_without_validation() → None

test() → None

This loop is for analyzing the models performance on a previously unseen labeled test dataset.

predict() → None

Predicts lipid species for unlabeled data. Input is one or more mzML.

get_pred_files() → list[str]

schedule_tuning() → None

tune_model(config: dict[str, Any], num_epochs: int, train_loader: DataLoader[Any], val_loader: DataLoader[Any], trainset_lipids: list[str], valset_lipids: list[str]) → None

prepare_tune_config() → None

check_parameter_for_tuning(config_section: str, parameter: str) → None

parse_lipid_dataset_name(lipid_name: str) → str

get_unique_lipids(dataset_list: list[str]) → list[str]

perform_data_split() → DataSplit

This method extracts the names of all datasets in the HDF5 input file and saves them in separate lists for the training and validation sets. These lists can than be used to iterate over the dataset using lazy loading if the whole dataset is too big to be loaded at once.

split_data_via_instructions(dataset_names: list[str]) → DataSplit

split_data_by_lipid_species(dataset_names: list[str]) → DataSplit

Sorts the data by lipid type, mode and collision energy so the training and validation sets generated during the splitting can be balanced.

run_random_forest() → None

class H5Dataset(config: dict[str, Any], dataset_names: list[str], lipid_librarian: LipidLibrary, file_path: str)

This class implements a custom PyTorch dataset. It handles reading in the training data from HDF5 files. It overrides the __getitem__ method to return the processed spectrum and its metadata for a sample at a given index.

file_path

The path to the HDF5 file

Type:

str

dataset_names

list of dataset names in the HDF5 file group “all_datasets”, used to access samples by index

Type:

list

dataset_len

An integer count of the samples in the HDF5 file

Type:

int

hdf5_file

the opened HDF5 file, set to None during initialization and set once first sample is requested

Type:

h5py.Dataset

config

Dictionary containing the information from the config.yaml file

Type:

dict

network_type

String of the network type specified in the config.yaml file

Type:

str

decimal_accuracy

Integer indicating the decimal accuracy to which the mass spectra should be binned

Type:

int

lipid_librarian

LipidLibrarian instance used for generating the label for a sample

Type:

LipidLibrary

get_n_highest_peaks(spectrum: ndarray, n_peaks: int) → ndarray

Processes the spectrum for a sample in the H5Dataset to prepare it as input for the model.

Parameters:

• spectrum (np.ndarray) – the spectrum extracted from an HDF5 dataset containing m/z and intensity arrays

• n_peaks (int) – maximum number of peaks to be fed into neural network

Returns:

the spectrum containing the number of peaks specified in n_peaks and sorted by descending intensity

Return type:

np.ndarray

bin_spectrum(spectrum: ndarray) → ndarray

Truncates the m/z values at the decimal position defined in the config.yaml and sums up the intensities.

Parameters:

spectrum (np.ndarray) – the m/z and intensity values of a dataset from the HDF5 file

Returns:

the binned spectrum ordered by intensity

Return type:

np.ndarray

class PredictionDataset(file_path: str, config: dict[str, Any])

process_input() → list[dict[str, Any]]

process_mzml() → list[dict[str, Any]]

process_json() → list[dict[str, Any]]

get_n_highest_peaks(mz_array: ndarray, intensity_array: ndarray, precursor: float) → Tensor

class LightningModule(config: dict[str, Any], evaluator: Evaluator, trainset_lipids: list[str] | None = None, valset_lipids: list[str] | None = None, testset_lipids: list[str] | None = None)

configure_optimizers() → Any

Choose what optimizers and learning-rate schedulers to use in your optimization. Normally you’d need one. But in the case of GANs or similar you might have multiple. Optimization with multiple optimizers only works in the manual optimization mode.

Returns:

Any of these 6 options.

• Single optimizer.

• List or Tuple of optimizers.

• Two lists - The first list has multiple optimizers, and the second has multiple LR schedulers (or multiple lr_scheduler_config).

• Dictionary, with an "optimizer" key, and (optionally) a "lr_scheduler" key whose value is a single LR scheduler or lr_scheduler_config.

• None - Fit will run without any optimizer.

The lr_scheduler_config is a dictionary which contains the scheduler and its associated configuration. The default configuration is shown below.

lr_scheduler_config = {
    # REQUIRED: The scheduler instance
    "scheduler": lr_scheduler,
    # The unit of the scheduler's step size, could also be 'step'.
    # 'epoch' updates the scheduler on epoch end whereas 'step'
    # updates it after a optimizer update.
    "interval": "epoch",
    # How many epochs/steps should pass between calls to
    # `scheduler.step()`. 1 corresponds to updating the learning
    # rate after every epoch/step.
    "frequency": 1,
    # Metric to monitor for schedulers like `ReduceLROnPlateau`
    "monitor": "val_loss",
    # If set to `True`, will enforce that the value specified 'monitor'
    # is available when the scheduler is updated, thus stopping
    # training if not found. If set to `False`, it will only produce a warning
    "strict": True,
    # If using the `LearningRateMonitor` callback to monitor the
    # learning rate progress, this keyword can be used to specify
    # a custom logged name
    "name": None,
}

When there are schedulers in which the .step() method is conditioned on a value, such as the torch.optim.lr_scheduler.ReduceLROnPlateau scheduler, Lightning requires that the lr_scheduler_config contains the keyword "monitor" set to the metric name that the scheduler should be conditioned on.

# The ReduceLROnPlateau scheduler requires a monitor
def configure_optimizers(self):
    optimizer = Adam(...)
    return {
        "optimizer": optimizer,
        "lr_scheduler": {
            "scheduler": ReduceLROnPlateau(optimizer, ...),
            "monitor": "metric_to_track",
            "frequency": "indicates how often the metric is updated",
            # If "monitor" references validation metrics, then "frequency" should be set to a
            # multiple of "trainer.check_val_every_n_epoch".
        },
    }


# In the case of two optimizers, only one using the ReduceLROnPlateau scheduler
def configure_optimizers(self):
    optimizer1 = Adam(...)
    optimizer2 = SGD(...)
    scheduler1 = ReduceLROnPlateau(optimizer1, ...)
    scheduler2 = LambdaLR(optimizer2, ...)
    return (
        {
            "optimizer": optimizer1,
            "lr_scheduler": {
                "scheduler": scheduler1,
                "monitor": "metric_to_track",
            },
        },
        {"optimizer": optimizer2, "lr_scheduler": scheduler2},
    )

Metrics can be made available to monitor by simply logging it using self.log('metric_to_track', metric_val) in your LightningModule.

Note

Some things to know:

• Lightning calls .backward() and .step() automatically in case of automatic optimization.

• If a learning rate scheduler is specified in configure_optimizers() with key "interval" (default “epoch”) in the scheduler configuration, Lightning will call the scheduler’s .step() method automatically in case of automatic optimization.

• If you use 16-bit precision (precision=16), Lightning will automatically handle the optimizer.

• If you use torch.optim.LBFGS, Lightning handles the closure function automatically for you.

• If you use multiple optimizers, you will have to switch to ‘manual optimization’ mode and step them yourself.

• If you need to control how often the optimizer steps, override the optimizer_step() hook.

training_step(batch: dict[str, Any], batch_idx: int) → Tensor

Here you compute and return the training loss and some additional metrics for e.g. the progress bar or logger.

Parameters:

• batch – The output of your data iterable, normally a DataLoader.

• batch_idx – The index of this batch.

• dataloader_idx – The index of the dataloader that produced this batch. (only if multiple dataloaders used)

Returns:

• Tensor - The loss tensor

• dict - A dictionary which can include any keys, but must include the key 'loss' in the case of automatic optimization.

• None - In automatic optimization, this will skip to the next batch (but is not supported for multi-GPU, TPU, or DeepSpeed). For manual optimization, this has no special meaning, as returning the loss is not required.

In this step you’d normally do the forward pass and calculate the loss for a batch. You can also do fancier things like multiple forward passes or something model specific.

Example:

def training_step(self, batch, batch_idx):
    x, y, z = batch
    out = self.encoder(x)
    loss = self.loss(out, x)
    return loss

To use multiple optimizers, you can switch to ‘manual optimization’ and control their stepping:

def __init__(self):
    super().__init__()
    self.automatic_optimization = False


# Multiple optimizers (e.g.: GANs)
def training_step(self, batch, batch_idx):
    opt1, opt2 = self.optimizers()

    # do training_step with encoder
    ...
    opt1.step()
    # do training_step with decoder
    ...
    opt2.step()

Note

When accumulate_grad_batches > 1, the loss returned here will be automatically normalized by accumulate_grad_batches internally.

validation_step(batch: dict[str, Any], batch_idx: int) → Tensor

Operates on a single batch of data from the validation set. In this step you’d might generate examples or calculate anything of interest like accuracy.

Parameters:

• batch – The output of your data iterable, normally a DataLoader.

• batch_idx – The index of this batch.

• dataloader_idx – The index of the dataloader that produced this batch. (only if multiple dataloaders used)

Returns:

• Tensor - The loss tensor

• dict - A dictionary. Can include any keys, but must include the key 'loss'.

• None - Skip to the next batch.

# if you have one val dataloader:
def validation_step(self, batch, batch_idx): ...


# if you have multiple val dataloaders:
def validation_step(self, batch, batch_idx, dataloader_idx=0): ...

Examples:

# CASE 1: A single validation dataset
def validation_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    val_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'val_loss': loss, 'val_acc': val_acc})

If you pass in multiple val dataloaders, validation_step() will have an additional argument. We recommend setting the default value of 0 so that you can quickly switch between single and multiple dataloaders.

# CASE 2: multiple validation dataloaders
def validation_step(self, batch, batch_idx, dataloader_idx=0):
    # dataloader_idx tells you which dataset this is.
    x, y = batch

    # implement your own
    out = self(x)

    if dataloader_idx == 0:
        loss = self.loss0(out, y)
    else:
        loss = self.loss1(out, y)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs separately for each dataloader
    self.log_dict({f"val_loss_{dataloader_idx}": loss, f"val_acc_{dataloader_idx}": acc})

Note

If you don’t need to validate you don’t need to implement this method.

Note

When the validation_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of validation, the model goes back to training mode and gradients are enabled.

test_step(batch: dict[str, Any], batch_idx: int) → Tensor

Operates on a single batch of data from the test set. In this step you’d normally generate examples or calculate anything of interest such as accuracy.

Parameters:

• batch – The output of your data iterable, normally a DataLoader.

• batch_idx – The index of this batch.

• dataloader_idx – The index of the dataloader that produced this batch. (only if multiple dataloaders used)

Returns:

• Tensor - The loss tensor

• dict - A dictionary. Can include any keys, but must include the key 'loss'.

• None - Skip to the next batch.

# if you have one test dataloader:
def test_step(self, batch, batch_idx): ...


# if you have multiple test dataloaders:
def test_step(self, batch, batch_idx, dataloader_idx=0): ...

Examples:

# CASE 1: A single test dataset
def test_step(self, batch, batch_idx):
    x, y = batch

    # implement your own
    out = self(x)
    loss = self.loss(out, y)

    # log 6 example images
    # or generated text... or whatever
    sample_imgs = x[:6]
    grid = torchvision.utils.make_grid(sample_imgs)
    self.logger.experiment.add_image('example_images', grid, 0)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    test_acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs!
    self.log_dict({'test_loss': loss, 'test_acc': test_acc})

If you pass in multiple test dataloaders, test_step() will have an additional argument. We recommend setting the default value of 0 so that you can quickly switch between single and multiple dataloaders.

# CASE 2: multiple test dataloaders
def test_step(self, batch, batch_idx, dataloader_idx=0):
    # dataloader_idx tells you which dataset this is.
    x, y = batch

    # implement your own
    out = self(x)

    if dataloader_idx == 0:
        loss = self.loss0(out, y)
    else:
        loss = self.loss1(out, y)

    # calculate acc
    labels_hat = torch.argmax(out, dim=1)
    acc = torch.sum(y == labels_hat).item() / (len(y) * 1.0)

    # log the outputs separately for each dataloader
    self.log_dict({f"test_loss_{dataloader_idx}": loss, f"test_acc_{dataloader_idx}": acc})

Note

If you don’t need to test you don’t need to implement this method.

Note

When the test_step() is called, the model has been put in eval mode and PyTorch gradients have been disabled. At the end of the test epoch, the model goes back to training mode and gradients are enabled.

predict_step(batch: dict[str, Any], batch_idx: int) → None

Step function called during predict(). By default, it calls forward(). Override to add any processing logic.

The predict_step() is used to scale inference on multi-devices.

To prevent an OOM error, it is possible to use BasePredictionWriter callback to write the predictions to disk or database after each batch or on epoch end.

The BasePredictionWriter should be used while using a spawn based accelerator. This happens for Trainer(strategy="ddp_spawn") or training on 8 TPU cores with Trainer(accelerator="tpu", devices=8) as predictions won’t be returned.

Parameters:

• batch – The output of your data iterable, normally a DataLoader.

• batch_idx – The index of this batch.

• dataloader_idx – The index of the dataloader that produced this batch. (only if multiple dataloaders used)

Returns:

Predicted output (optional).

Example

class MyModel(LightningModule):

    def predict_step(self, batch, batch_idx, dataloader_idx=0):
        return self(batch)

dm = ...
model = MyModel()
trainer = Trainer(accelerator="gpu", devices=2)
predictions = trainer.predict(model, dm)

get_preds_vs_labels(batch_idx: int, output: Tensor, labels: Tensor, dataset_path: Tensor) → Tensor

get_test_preds_vs_labels(output: Tensor, labels: Tensor, dataset_path: Tensor, confidence_scores: Tensor) → Tensor

get_neural_network() → Module

save_model(output_folder: str) → None

Helpers

class LipidLibrary

get_regression_label(lipid_species: str) → tuple[Tensor, dict[str, Any]]

get_transformer_label(lipid_species: str, adduct: str, output_seq_length: int) → tuple[Tensor, dict[str, Any]]

parse_lipid_species_components(lipid_species: str) → list[str]

get_lipid_species_components(lipid_species: str) → LipidComponents

normalize_precursor_mass(precursor_mass: float) → float

translate_tokens_to_name(tokens: Tensor) → list[str]

custom_accuracy_scoring(prediction: Any, label: Any) → float

read_yaml(file_to_open: str) → Any

write_yaml(file_to_open: str, dict_to_write: dict[str, Any]) → None

resolve_config_paths(config: dict[str, Any]) → dict[str, Any]

Resolve all file paths in a configuration dictionary.

Paths can be: - Absolute paths: Used as-is - Relative paths: Resolved based on path type (data, model, output)

Parameters:

config – Configuration dictionary from YAML.

Returns:

Config with resolved absolute paths.

set_seeds(seed: int = 42) → None

get_project_root() → Path

Get the project root directory.

Searches upward from this file’s location for a directory containing pyproject.toml. Raises FileNotFoundError if not found (e.g. in an installed wheel layout), prompting the user to set environment variables instead.

Returns:

Path to the project root directory.

resolve_data_path(relative_path: str | Path) → Path

Resolve a data file path.

Parameters:

relative_path – Path relative to data directory.

Returns:

Absolute path to data file.

Environment variables:

LIPIDETECTIVE_DATA_DIR: Override default data directory.

resolve_model_path(relative_path: str | Path) → Path

Resolve a model file path.

Parameters:

relative_path – Path relative to models directory.

Returns:

Absolute path to model file.

Environment variables:

LIPIDETECTIVE_MODELS_DIR: Override default models directory.

resolve_output_path(relative_path: str | Path) → Path

Resolve an experiment output path.

Parameters:

relative_path – Path relative to experiments directory.

Returns:

Absolute path for experiment output.

Environment variables:

LIPIDETECTIVE_OUTPUT_DIR: Override default output directory.

resolve_config_path(relative_path: str | Path) → Path

Resolve a configuration file path.

Parameters:

relative_path – Path relative to config directory.

Returns:

Absolute path to config file.

Environment variables:

LIPIDETECTIVE_CONFIG_DIR: Override default config directory.