massdash.peakPickers.ConformerPeakPicker

class massdash.peakPickers.ConformerPeakPicker(library: SpectralLibraryLoader, pretrained_model_file: str, prediction_threshold: float = 0.5, prediction_type: Literal['logits', 'sigmoided', 'binarized'] = 'logits')

Bases: object

Class for performing peak picking using the Conformer model.

transition_group

The transition group object.

Type:

TransitionGroup

pretrained_model_file

The path to the pretrained model file.

Type:

str

window_size

The window size for peak picking. Defaults to 175.

Type:

int, optional

prediction_threshold

The prediction threshold for peak picking. Defaults to 0.5.

Type:

float, optional

prediction_type

The prediction type for peak picking. Defaults to “logits”.

Type:

str, optional

onnx_session

The onnx session.

Type:

onnxruntime.InferenceSession

_validate_model()

Validate the pretrained model is valid and an onnx model.

load_model()

Load the pretrained model.

pick()

Perform peak picking.

_convertConformerFeatureToTransitionGroupFeatures()

Convert conformer predicted feature to TransitionGroupFeatures.

load_model()

Load the pretrained model.

pick(transition_group, max_int_transition: int = 1000) → List[TransitionGroupFeature]

Perform peak picking.

Parameters:

max_int_transition (int, optional) – The maximum intensity transition. Defaults to 1000.

Returns:

The list of transition group features.

Return type:

List[TransitionGroupFeature]