yolksac_human.py

# coding=utf-8
# Copyright 2020 The TensorFlow Datasets Authors and the HuggingFace Datasets Authors.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
#     http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.

# Lint as: python3
"""The RNA Expression Baseclass."""


import json
import os
import anndata as ad
import pyarrow as pa
import pandas as pd
import numpy as np

import datasets
CITATION = """
Test
"""

DESCRIPTION = """
Test
"""

class RNAExpConfig(datasets.BuilderConfig):
    """BuilderConfig for RNAExpConfig."""

    def __init__(self, features, data_url, citation, url, raw_counts="X", **kwargs):
        """BuilderConfig for RNAExpConfig.
        Args:
          features: `list[string]`, list of the features that will appear in the
            feature dict. Should not include "label".
          data_url: `string`, url to download the zip file from.
          citation: `string`, citation for the data set.
          url: `string`, url for information about the data set.

          **kwargs: keyword arguments forwarded to super.
        """
        # Version history:
        # 0.0.1: Initial version.
        super(RNAExpConfig, self).__init__(version=datasets.Version("0.0.1"), **kwargs)
        self.features = features
        self.data_url = data_url
        self.citation = citation
        self.url = url
        self.raw_counts = raw_counts # Could be raw.X
        self.batch = 1000
        self.species = None




# class RNAExp(datasets.GeneratorBasedBuilder):
class RNAExp(datasets.ArrowBasedBuilder):
    """RNA Expression Baseclass."""

    def _info(self):
        self.config = RNAExpConfig(
            name="human_yolk_sac",
            description = DESCRIPTION,
            features=["raw_counts",'LVL1', 'LVL2', 'LVL3'],
            raw_counts = "X",
            data_url="./data/17_04_24_YolkSacRaw_F158_WE_annots.h5ad",
            citation=CITATION,
            url="https://www.ebi.ac.uk/biostudies/arrayexpress/studies/E-MTAB-11673")

        features = {"raw_counts": datasets.features.Sequence(feature=datasets.Value("int32"))}

        # features = {"raw_counts": datasets.features.Sequence(feature={"gene":datasets.Value("string"),"count":datasets.Value("int32")})}
        # features = {"raw_counts": datasets.Value("int32") for gene in adata.var.index.str.lower().tolist()}
        for feature in self.config.features:
            if features.get(feature,None) is None:
                features[feature] = datasets.Value("string")

        # features["gene_names"] = datasets.Sequence(datasets.Value("string"))

        return datasets.DatasetInfo(
            description= self.config.description,
            features=None, #datasets.Features(features),
            homepage=self.config.url,
            citation=self.config.citation,
        )

    def _split_generators(self, dl_manager):
        self.anndata_file = dl_manager.download_and_extract(self.config.data_url)
        

        return [
            datasets.SplitGenerator(
                name=datasets.Split.TRAIN,
                gen_kwargs={"split": "train","expression_file": self.anndata_file,"batch_size":self.config.batch},#,"gene_names_file": self.gene_names_file},    
            )
        ]

    def _generate_examples(self, expression_file, split):

        # genes = pd.read_csv(gene_names_file)
        adata = ad.read_h5ad(expression_file)
        self.genes_list = adata.var.index.str.lower().tolist()


        if self.config.raw_counts =="X":
            X = adata.X
        else:
            X = adata.var[raw_counts]
        num_cells = X.shape[0]
        for _id,cell in enumerate(X):
            example = {"raw_counts": cell.toarray().flatten()}

            for feature in self.config.features:
                if example.get(feature,None) is None:
                    example[feature] = adata.obs[feature][_id]

            yield _id,example

    def _generate_tables(self, expression_file,batch_size,split):
        idx = 0
        adata = ad.read_h5ad(expression_file,backed='r')
        genes = adata.var_names.str.lower().to_list()

        features = {"raw_counts": datasets.features.Sequence(datasets.features.Value("int32"),id = ",".join(adata.var.index.str.lower().tolist()))}
        for feature in self.config.features:
            if features.get(feature,None) is None:
                features[feature] = datasets.Value("string")

        self.info.features = datasets.Features(features)
        

        # self.info.features['gene_names'] = datasets.features.ClassLabel(names = genes)
        
        # self.info.description   = adata.var.index.str.lower().tolist() #"+".join(adata.var.index.str.lower().tolist())
        for batch in range(0,adata.shape[0],batch_size):
            chunk = adata.X[batch:batch+batch_size].todense().astype('int32')
            df = pd.DataFrame(chunk,columns=adata.var.index.str.lower())
            df["raw_counts"] = [x for x in df.to_numpy()]
            df = df[["raw_counts"]]
            ## We create a dummy column with all the names of the genes as list. We don't use this as value since this would unnecessarily increase the size of the dataset
            ## Another option would be to replace the description with the list of genes 
            # df[",".join(adata.var.index.str.lower().tolist())] = True
            # df['gene_names'] = True
            
            for feature in self.config.features:
                if feature != "raw_counts":
                    df[feature] = adata.obs[feature][batch:batch+batch_size].tolist()
            
            # df['gene_names'] = [adata.var.index.str.lower().tolist()]*batch_size
            # print(df)
            pa_table = pa.Table.from_pandas(df)
            yield idx, pa_table
            idx += 1