import abc
from pathlib import Path
from typing import List
import pandas as pd
from sklearn.feature_extraction import DictVectorizer
from sklearn.preprocessing import StandardScaler
from immuneML.analysis.data_manipulation.NormalizationType import NormalizationType
from immuneML.caching.CacheHandler import CacheHandler
from immuneML.data_model.encoded_data.EncodedData import EncodedData
from immuneML.data_model.receptor.receptor_sequence import ReceptorSequence
from immuneML.encodings.DatasetEncoder import DatasetEncoder
from immuneML.encodings.EncoderParams import EncoderParams
from immuneML.encodings.kmer_frequency.ReadsType import ReadsType
from immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType import SequenceEncodingType
from immuneML.encodings.preprocessing.FeatureScaler import FeatureScaler
from immuneML.environment.Constants import Constants
from immuneML.environment.SequenceType import SequenceType
from immuneML.util.EncoderHelper import EncoderHelper
from immuneML.util.ParameterValidator import ParameterValidator
from immuneML.util.ReflectionHandler import ReflectionHandler
[docs]class KmerFrequencyEncoder(DatasetEncoder):
"""
The KmerFrequencyEncoder class encodes a repertoire, sequence or receptor by frequencies of k-mers it contains.
A k-mer is a sequence of letters of length k into which an immune receptor sequence can be decomposed.
K-mers can be defined in different ways, as determined by the sequence_encoding.
Arguments:
sequence_encoding (:py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType`): The type of k-mers that are used. The simplest sequence_encoding is :py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.CONTINUOUS_KMER`, which uses contiguous subsequences of length k to represent the k-mers. When gapped k-mers are used (:py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.GAPPED_KMER`, :py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.GAPPED_KMER`), the k-mers may contain gaps with a size between min_gap and max_gap, and the k-mer length is defined as a combination of k_left and k_right. When IMGT k-mers are used (:py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.IMGT_CONTINUOUS_KMER`, :py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.IMGT_GAPPED_KMER`), IMGT positional information is taken into account (i.e. the same sequence in a different position is considered to be a different k-mer). When the identity representation is used (:py:mod:`~immuneML.encodings.kmer_frequency.sequence_encoding.SequenceEncodingType.IDENTITY`), the k-mers just correspond to the original sequences.
normalization_type (:py:mod:`~immuneML.analysis.data_manipulation.NormalizationType`): The way in which the k-mer frequencies should be normalized. The default value for normalization_type is l2.
reads (:py:mod:`~immuneML.encodings.kmer_frequency.ReadsType`): Reads type signify whether the counts of the sequences in the repertoire will be taken into account. If :py:mod:`~immuneML.encodings.kmer_frequency.ReadsType.UNIQUE`, only unique sequences (clonotypes) are encoded, and if :py:mod:`~immuneML.encodings.kmer_frequency.ReadsType.ALL`, the sequence 'count' value is taken into account when determining the k-mer frequency. The default value for reads is unique.
k (int): Length of the k-mer (number of amino acids) when ungapped k-mers are used. The default value for k is 3.
k_left (int): When gapped k-mers are used, k_left indicates the length of the k-mer left of the gap. The default value for k_left is 1.
k_right (int): Same as k_left, but k_right determines the length of the k-mer right of the gap. The default value for k_right is 1.
min_gap (int): Minimum gap size when gapped k-mers are used. The default value for min_gap is 0.
max_gap: (int): Maximum gap size when gapped k-mers are used. The default value for max_gap is 0.
sequence_type (:py:mod:`~immuneML.environment.SequenceType.SequenceType`): Whether to work with nucleotide or amino acid sequences. Amino acid sequences are the default. To work with either sequence type, the sequences of the desired type should be included in the datasets, e.g., listed under 'columns_to_load' parameter. By default, both types will be included if available. Valid values are: AMINO_ACID and NUCLEOTIDE.
scale_to_unit_variance (bool): whether to scale the design matrix after normalization to have unit variance per feature. Setting this argument to True might improve the subsequent classifier's performance depending on the type of the classifier. The default value for scale_to_unit_variance is true.
scale_to_zero_mean (bool): whether to scale the design matrix after normalization to have zero mean per feature. Setting this argument to True might improve the subsequent classifier's performance depending on the type of the classifier. However, if the original design matrix was sparse, setting this argument to True will destroy the sparsity and will increase the memory consumption. The default value for scale_to_zero_mean is false.
YAML specification:
.. indent with spaces
.. code-block:: yaml
my_continuous_kmer:
KmerFrequency:
normalization_type: RELATIVE_FREQUENCY
reads: UNIQUE
sequence_encoding: CONTINUOUS_KMER
sequence_type: NUCLEOTIDE
k: 3
scale_to_unit_variance: True
scale_to_zero_mean: True
my_gapped_kmer:
KmerFrequency:
normalization_type: RELATIVE_FREQUENCY
reads: UNIQUE
sequence_encoding: GAPPED_KMER
sequence_type: AMINO_ACID
k_left: 2
k_right: 2
min_gap: 1
max_gap: 3
scale_to_unit_variance: True
scale_to_zero_mean: False
"""
STEP_ENCODED = "encoded"
STEP_VECTORIZED = "vectorized"
STEP_NORMALIZED = "normalized"
STEP_SCALED = "scaled"
dataset_mapping = {
"RepertoireDataset": "KmerFreqRepertoireEncoder",
"SequenceDataset": "KmerFreqSequenceEncoder",
"ReceptorDataset": "KmerFreqReceptorEncoder"
}
def __init__(self, normalization_type: NormalizationType, reads: ReadsType, sequence_encoding: SequenceEncodingType, k: int = 0,
k_left: int = 0, k_right: int = 0, min_gap: int = 0, max_gap: int = 0, metadata_fields_to_include: list = None,
name: str = None, scale_to_unit_variance: bool = False, scale_to_zero_mean: bool = False, sequence_type: SequenceType = None):
self.normalization_type = normalization_type
self.reads = reads
self.sequence_encoding = sequence_encoding
self.sequence_type = sequence_type
self.k = k
self.k_left = k_left
self.k_right = k_right
self.min_gap = min_gap
self.max_gap = max_gap
self.metadata_fields_to_include = metadata_fields_to_include if metadata_fields_to_include is not None else []
self.name = name
self.scale_to_unit_variance = scale_to_unit_variance
self.scale_to_zero_mean = scale_to_zero_mean
self.scaler = None
self.vectorizer = None
@staticmethod
def _prepare_parameters(normalization_type: str, reads: str, sequence_encoding: str, k: int = 0, k_left: int = 0,
k_right: int = 0, min_gap: int = 0, max_gap: int = 0, metadata_fields_to_include: list = None, name: str = None,
scale_to_unit_variance: bool = False, scale_to_zero_mean: bool = False, sequence_type: str = None):
location = KmerFrequencyEncoder.__name__
ParameterValidator.assert_in_valid_list(normalization_type.upper(), [item.name for item in NormalizationType], location, "normalization_type")
ParameterValidator.assert_in_valid_list(reads.upper(), [item.name for item in ReadsType], location, "reads")
ParameterValidator.assert_in_valid_list(sequence_encoding.upper(), [item.name for item in SequenceEncodingType], location, "sequence_encoding")
ParameterValidator.assert_type_and_value(scale_to_zero_mean, bool, location, "scale_to_zero_mean")
ParameterValidator.assert_type_and_value(scale_to_unit_variance, bool, location, "scale_to_unit_variance")
ParameterValidator.assert_type_and_value(sequence_type, str, location, 'sequence_type')
ParameterValidator.assert_in_valid_list(sequence_type.upper(), [st.name for st in SequenceType], location, 'sequence_type')
if "IMGT" in sequence_encoding.upper():
assert sequence_type.upper() == SequenceType.AMINO_ACID.name, f"{location}: for IMGT-based k-mer frequency encoding (here: " \
f"{sequence_encoding.upper()}), sequence type has to be 'amino_acid'."
vars_to_check = {"k": k, "k_left": k_left, "k_right": k_right, "min_gap": min_gap, "max_gap": max_gap}
for param in vars_to_check.keys():
ParameterValidator.assert_type_and_value(vars_to_check[param], int, location, param, min_inclusive=0)
if "gap" in sequence_encoding.lower():
assert k_left != 0 and k_right != 0, f"KmerFrequencyEncoder: sequence encoding {sequence_encoding} was chosen, but k_left " \
f"({k_left}) or k_right ({k_right}) have to be set and larger than 0."
return {
"normalization_type": NormalizationType[normalization_type.upper()],
"reads": ReadsType[reads.upper()],
"sequence_encoding": SequenceEncodingType[sequence_encoding.upper()],
"name": name,
"scale_to_zero_mean": scale_to_zero_mean, "scale_to_unit_variance": scale_to_unit_variance,
'sequence_type': SequenceType[sequence_type.upper()],
**vars_to_check
}
[docs] @staticmethod
def build_object(dataset=None, **params):
try:
prepared_params = KmerFrequencyEncoder._prepare_parameters(**params)
encoder = ReflectionHandler.get_class_by_name(KmerFrequencyEncoder.dataset_mapping[dataset.__class__.__name__],
"kmer_frequency/")(**prepared_params)
except ValueError:
raise ValueError("{} is not defined for dataset of type {}.".format(KmerFrequencyEncoder.__name__, dataset.__class__.__name__))
return encoder
[docs] def encode(self, dataset, params: EncoderParams):
cache_params = self._prepare_caching_params(dataset, params)
encoded_dataset = CacheHandler.memo_by_params(cache_params, lambda: self._encode_new_dataset(dataset, params))
EncoderHelper.sync_encoder_with_cache(cache_params, lambda: {'vectorizer': self.vectorizer, 'scaler': self.scaler}, self,
['vectorizer', 'scaler'])
return encoded_dataset
def _prepare_caching_params(self, dataset, params: EncoderParams, step: str = ""):
return (("dataset_identifier", dataset.identifier),
("labels", tuple(params.label_config.get_labels_by_name())),
("encoding", KmerFrequencyEncoder.__name__),
("learn_model", params.learn_model),
("step", step),
("encoding_params", tuple(vars(self).items())))
def _encode_data(self, dataset, params: EncoderParams) -> EncodedData:
encoded_example_list, example_ids, encoded_labels, feature_annotation_names = CacheHandler.memo_by_params(
self._prepare_caching_params(dataset, params, KmerFrequencyEncoder.STEP_ENCODED),
lambda: self._encode_examples(dataset, params))
self._initialize_vectorizer(params)
vectorized_examples = self._vectorize_encoded(examples=encoded_example_list, params=params)
feature_names = self.vectorizer.feature_names_
normalized_examples = FeatureScaler.normalize(vectorized_examples, self.normalization_type)
if self.scale_to_unit_variance:
examples = self.scale_normalized(params, dataset, normalized_examples)
else:
examples = normalized_examples
feature_annotations = self._get_feature_annotations(feature_names, feature_annotation_names)
encoded_data = EncodedData(examples=examples,
labels=encoded_labels,
feature_names=feature_names,
example_ids=example_ids,
feature_annotations=feature_annotations,
encoding=KmerFrequencyEncoder.__name__)
return encoded_data
[docs] def scale_normalized(self, params, dataset, normalized_examples):
if params.learn_model:
self.scaler = StandardScaler(with_mean=self.scale_to_zero_mean)
examples = CacheHandler.memo_by_params(
self._prepare_caching_params(dataset, params, step=KmerFrequencyEncoder.STEP_SCALED),
lambda: FeatureScaler.standard_scale_fit(self.scaler, normalized_examples, with_mean=self.scale_to_zero_mean))
else:
examples = CacheHandler.memo_by_params(
self._prepare_caching_params(dataset, params, step=KmerFrequencyEncoder.STEP_SCALED),
lambda: FeatureScaler.standard_scale(self.scaler, normalized_examples, with_mean=self.scale_to_zero_mean))
return examples
@abc.abstractmethod
def _encode_new_dataset(self, dataset, params: EncoderParams):
pass
@abc.abstractmethod
def _encode_examples(self, dataset, params: EncoderParams):
pass
def _initialize_vectorizer(self, params: EncoderParams):
if self.vectorizer is None or params.learn_model:
self.vectorizer = DictVectorizer(sparse=True, dtype=float)
def _vectorize_encoded(self, examples: list, params: EncoderParams):
if params.learn_model:
vectorized_examples = self.vectorizer.fit_transform(examples)
else:
vectorized_examples = self.vectorizer.transform(examples)
return vectorized_examples
def _get_feature_annotations(self, feature_names, feature_annotation_names):
feature_annotations = pd.DataFrame({"feature": feature_names})
feature_annotations[feature_annotation_names] = feature_annotations['feature'].str.split(Constants.FEATURE_DELIMITER, expand=True)
return feature_annotations
def _prepare_sequence_encoder(self):
class_name = self.sequence_encoding.value
sequence_encoder = ReflectionHandler.get_class_by_name(class_name, "encodings/")
return sequence_encoder
def _encode_sequence(self, sequence: ReceptorSequence, params: EncoderParams, sequence_encoder, counts):
params.model = vars(self)
features = sequence_encoder.encode_sequence(sequence, params)
if features is not None:
for i in features:
if self.reads == ReadsType.UNIQUE:
counts[i] += 1
elif self.reads == ReadsType.ALL:
counts[i] += sequence.metadata.count
return counts
[docs] def get_additional_files(self) -> List[str]:
return []
[docs] @staticmethod
def export_encoder(path: Path, encoder) -> Path:
encoder_file = DatasetEncoder.store_encoder(encoder, path / "encoder.pickle")
return encoder_file
[docs] @staticmethod
def load_encoder(encoder_file: Path):
encoder = DatasetEncoder.load_encoder(encoder_file)
return encoder