Skip to content

Interfaces

Decodable

An interface for models that can be decoded by algorithms that conform to the search interface.

decode(sequence) abstractmethod

Map sequences of indices to residues using the model's residue vocabulary.

Parameters:

Name Type Description Default
sequence LongTensor

The sequence of residue indices to be mapped to the corresponding residue strings.

 required
Source code in instanovo/inference/interfaces.py 
57
58
59
60
61
62
63
64
65
66
@abstractmethod
def decode(self, sequence: torch.LongTensor) -> list[str]:
    """Map sequences of indices to residues using the model's residue vocabulary.

    Args:
        sequence (torch.LongTensor):
            The sequence of residue indices to be mapped
            to the corresponding residue strings.
    """
    pass

get_empty_index() abstractmethod

Get the empty token's index in the model's residue vocabulary.

Source code in instanovo/inference/interfaces.py 
73
74
75
76
@abstractmethod
def get_empty_index(self) -> int:
    """Get the empty token's index in the model's residue vocabulary."""
    pass

get_eos_index() abstractmethod

Get the end of sequence token's index in the model's residue vocabulary.

Source code in instanovo/inference/interfaces.py 
68
69
70
71
@abstractmethod
def get_eos_index(self) -> int:
    """Get the end of sequence token's index in the model's residue vocabulary."""
    pass

get_residue_masses(mass_scale) abstractmethod

Get residue masses for the model's residue vocabulary.

Parameters:

Name Type Description Default
mass_scale int

The scale in Daltons at which masses are calculated and rounded off. For example, a scale of 10000 would represent masses at a scale of 1e4 Da.

 required
Source code in instanovo/inference/interfaces.py 
44
45
46
47
48
49
50
51
52
53
54
55
@abstractmethod
def get_residue_masses(self, mass_scale: int) -> torch.LongTensor:
    """Get residue masses for the model's residue vocabulary.

    Args:
        mass_scale (int):
            The scale in Daltons at which masses are
            calculated and rounded off. For example,
            a scale of 10000 would represent masses
            at a scale of 1e4 Da.
    """
    pass

init(spectra, precursors, *args, **kwargs) abstractmethod

Initialize the search state.

Parameters:

Name Type Description Default
spectra FloatTensor

The spectra to be sequenced.

 required
precursors torch.FloatTensor[batch size, 3]

The precursor mass, charge and mass-to-charge ratio.

 required
Source code in instanovo/inference/interfaces.py 
13
14
15
16
17
18
19
20
21
22
23
24
25
26
@abstractmethod
def init(  # type:ignore
    self, spectra: torch.FloatTensor, precursors: torch.FloatTensor, *args, **kwargs
) -> Any:
    """Initialize the search state.

    Args:
        spectra (torch.FloatTensor):
            The spectra to be sequenced.

        precursors (torch.FloatTensor[batch size, 3]):
            The precursor mass, charge and mass-to-charge ratio.
    """
    pass

score_candidates(sequences, precursor_mass_charge, *args, **kwargs) abstractmethod

Generate and score the next set of candidates.

Parameters:

Name Type Description Default
sequences LongTensor

Partial residue sequences in generated the course of decoding.

 required
precursor_mass_charge torch.FloatTensor[batch size, 3]

The precursor mass, charge and mass-to-charge ratio.

 required
Source code in instanovo/inference/interfaces.py 
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
@abstractmethod
def score_candidates(  # type:ignore
    self, sequences: torch.LongTensor, precursor_mass_charge: torch.FloatTensor, *args, **kwargs
) -> torch.FloatTensor:
    """Generate and score the next set of candidates.

    Args:
        sequences (torch.LongTensor):
            Partial residue sequences in generated
            the course of decoding.

        precursor_mass_charge (torch.FloatTensor[batch size, 3]):
            The precursor mass, charge and mass-to-charge ratio.
    """
    pass

Decoder(model)

A class that implements some search algorithm for decoding from a model that conforms to the Decodable interface.

Parameters:

Name Type Description Default
model Decodable

The model to predict residue sequences from using the implemented search algorithm.

 required
Source code in instanovo/inference/interfaces.py 
89
90
def __init__(self, model: Decodable):
    self.model = model

decode(spectra, precursors, *args, **kwargs) abstractmethod

Generate the predicted residue sequence using the decoder's search algorithm.

Parameters:

Name Type Description Default
spectra FloatTensor

The spectra to be sequenced.

 required
precursors FloatTensor

The precursor mass, charge and mass-to-charge ratio.

 required
Source code in instanovo/inference/interfaces.py 
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
@abstractmethod
def decode(  # type:ignore
    self, spectra: torch.FloatTensor, precursors: torch.FloatTensor, *args, **kwargs
) -> list[list[str]]:
    """Generate the predicted residue sequence using the decoder's search algorithm.

    Args:
        spectra (torch.FloatTensor):
            The spectra to be sequenced.

        precursors (torch.FloatTensor):
            The precursor mass, charge and mass-to-charge ratio.

    """
    pass