diff options
Diffstat (limited to 'gn3/db')
-rw-r--r-- | gn3/db/traits.py | 93 |
1 files changed, 93 insertions, 0 deletions
diff --git a/gn3/db/traits.py b/gn3/db/traits.py index f2673c8..1c6aaa7 100644 --- a/gn3/db/traits.py +++ b/gn3/db/traits.py @@ -1,12 +1,81 @@ """This class contains functions relating to trait data manipulation""" import os +from functools import reduce from typing import Any, Dict, Union, Sequence + from gn3.settings import TMPDIR from gn3.random import random_string from gn3.function_helpers import compose from gn3.db.datasets import retrieve_trait_dataset +def export_trait_data( + trait_data: dict, samplelist: Sequence[str], dtype: str = "val", + var_exists: bool = False, n_exists: bool = False): + """ + Export data according to `samplelist`. Mostly used in calculating + correlations. + + DESCRIPTION: + Migrated from + https://github.com/genenetwork/genenetwork1/blob/master/web/webqtl/base/webqtlTrait.py#L166-L211 + + PARAMETERS + trait: (dict) + The dictionary of key-value pairs representing a trait + samplelist: (list) + A list of sample names + dtype: (str) + ... verify what this is ... + var_exists: (bool) + A flag indicating existence of variance + n_exists: (bool) + A flag indicating existence of ndata + """ + def __export_all_types(tdata, sample): + sample_data = [] + if tdata[sample]["value"]: + sample_data.append(tdata[sample]["value"]) + if var_exists: + if tdata[sample]["variance"]: + sample_data.append(tdata[sample]["variance"]) + else: + sample_data.append(None) + if n_exists: + if tdata[sample]["ndata"]: + sample_data.append(tdata[sample]["ndata"]) + else: + sample_data.append(None) + else: + if var_exists and n_exists: + sample_data += [None, None, None] + elif var_exists or n_exists: + sample_data += [None, None] + else: + sample_data.append(None) + + return tuple(sample_data) + + def __exporter(accumulator, sample): + # pylint: disable=[R0911] + if sample in trait_data["data"]: + if dtype == "val": + return accumulator + (trait_data["data"][sample]["value"], ) + if dtype == "var": + return accumulator + (trait_data["data"][sample]["variance"], ) + if dtype == "N": + return accumulator + (trait_data["data"][sample]["ndata"], ) + if dtype == "all": + return accumulator + __export_all_types(trait_data["data"], sample) + raise KeyError("Type `%s` is incorrect" % dtype) + if var_exists and n_exists: + return accumulator + (None, None, None) + if var_exists or n_exists: + return accumulator + (None, None) + return accumulator + (None,) + + return reduce(__exporter, samplelist, tuple()) + def get_trait_csv_sample_data(conn: Any, trait_name: int, phenotype_id: int): """Fetch a trait and return it as a csv string""" @@ -674,3 +743,27 @@ def generate_traits_filename(base_path: str = TMPDIR): """Generate a unique filename for use with generated traits files.""" return "{}/traits_test_file_{}.txt".format( os.path.abspath(base_path), random_string(10)) + +def export_informative(trait_data: dict, inc_var: bool = False) -> tuple: + """ + Export informative strain + + This is a migration of the `exportInformative` function in + web/webqtl/base/webqtlTrait.py module in GeneNetwork1. + + There is a chance that the original implementation has a bug, especially + dealing with the `inc_var` value. It the `inc_var` value is meant to control + the inclusion of the `variance` value, then the current implementation, and + that one in GN1 have a bug. + """ + def __exporter__(acc, data_item): + if not inc_var or data_item["variance"] is not None: + return ( + acc[0] + (data_item["sample_name"],), + acc[1] + (data_item["value"],), + acc[2] + (data_item["variance"],)) + return acc + return reduce( + __exporter__, + filter(lambda td: td["value"] is not None, trait_data["data"].values()), + (tuple(), tuple(), tuple())) |