From 5b33832cbb6a7eb40cd87128efc3bcf2b5a54726 Mon Sep 17 00:00:00 2001
From: Frederick Muriuki Muriithi
Date: Thu, 17 Oct 2024 14:32:22 -0500
Subject: Fetch samples from database

Fetch the samples from the database. These will be used to verify
that the samples in the phenotype files already exist in the database
and are valid.
---
 scripts/rqtl2/phenotypes_qc.py | 12 ++++++++++++
 1 file changed, 12 insertions(+)

diff --git a/scripts/rqtl2/phenotypes_qc.py b/scripts/rqtl2/phenotypes_qc.py
index e495a97..d565be2 100644
--- a/scripts/rqtl2/phenotypes_qc.py
+++ b/scripts/rqtl2/phenotypes_qc.py
@@ -3,6 +3,7 @@ import sys
 import tempfile
 from pathlib import Path
 from zipfile import ZipFile
+from functools import reduce
 import multiprocessing as mproc
 from logging import Logger, getLogger, StreamHandler
 
@@ -13,6 +14,8 @@ from r_qtl import r_qtl2_qc as rqc
 from r_qtl import exceptions as rqe
 
 from uploader.files import sha256_digest_over_file
+from uploader.samples.models import samples_by_species_and_population
+
 from scripts.rqtl2.entry import build_main
 from scripts.rqtl2.cli_parser import add_bundle_argument
 from scripts.cli_parser import init_cli_parser, add_global_data_arguments
@@ -144,6 +147,15 @@ def run_qc(# pylint: disable=[too-many-arguments]
             ((ftype, cdata, extractiondir)
              for ftype in ("pheno", "phenocovar", "phenose", "phenonum")))
 
+    #       - Fetch samples/individuals from database.
+    samples = tuple(
+        item for item in set(reduce(
+            lambda acc, item: acc + (
+                item["Name"], item["Name2"], item["Symbol"], item["Alias"]),
+            samples_by_species_and_population(dbconn, speciesid, populationid),
+            tuple()))
+        if bool(item))
+
     #       - Check that `description` and `units` is present in phenocovar for
     #         all phenotypes
     #       - Check all phenotypes in pheno files exist in phenocovar files
-- 
cgit v1.2.3