Merge pull request #1240 from Yashsingh045/GH1237-seed-random

Standardize test suite randomization using np.random

Author: jonbrenas

tests/anoph/test_aim_data.py

@@ -1,7 +1,7 @@
 import itertools
-import random
 
 import plotly.graph_objects as go
+import numpy as np
 import pytest
 import xarray as xr
 from numpy.testing import assert_array_equal
@@ -88,9 +88,9 @@ def test_aim_calls(aims, ag3_sim_api):
     all_releases = api.releases
     parametrize_sample_sets = [
         None,
-        random.choice(all_sample_sets),
-        random.sample(all_sample_sets, 2),
-        random.choice(all_releases),
+        str(np.random.choice(all_sample_sets)),
+        np.random.choice(all_sample_sets, size=2, replace=False).tolist(),
+        np.random.choice(all_releases),
     ]
 
     # Parametrize sample_query.
@@ -179,9 +179,9 @@ def test_plot_aim_heatmap(aims, ag3_sim_api):
     all_releases = api.releases
     parametrize_sample_sets = [
         None,
-        random.choice(all_sample_sets),
-        random.sample(all_sample_sets, 2),
-        random.choice(all_releases),
+        str(np.random.choice(all_sample_sets)),
+        np.random.choice(all_sample_sets, size=2, replace=False).tolist(),
+        np.random.choice(all_releases),
     ]
 
     # Parametrize sample_query.

tests/anoph/test_cnv_data.py

@@ -1,5 +1,3 @@
-import random
-
 import bokeh.models
 import dask.array as da
 import numpy as np
@@ -136,14 +134,15 @@ def test_open_cnv_coverage_calls(fixture, api: AnophelesCnvData):
     # Check with a sample set that should not exist
     with pytest.raises(ValueError):
         root = api.open_cnv_coverage_calls(
-            sample_set="foobar", analysis=random.choice(api.coverage_calls_analysis_ids)
+            sample_set="foobar",
+            analysis=str(np.random.choice(api.coverage_calls_analysis_ids)),
         )
 
     # Check with an analysis that should not exist
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     with pytest.raises(ValueError):
         root = api.open_cnv_coverage_calls(
-            sample_set=random.choice(all_sample_sets), analysis="foobar"
+            sample_set=str(np.random.choice(all_sample_sets)), analysis="foobar"
         )
 
     # Check with a sample set and analysis that should not exist
@@ -343,15 +342,15 @@ def test_cnv_hmm(fixture, api: AnophelesCnvData):
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     parametrize_sample_sets = [
         None,
-        random.choice(all_sample_sets),
-        random.sample(all_sample_sets, 2),
-        random.choice(all_releases),
+        str(np.random.choice(all_sample_sets)),
+        np.random.choice(all_sample_sets, size=2, replace=False).tolist(),
+        np.random.choice(all_releases),
     ]
 
     # Parametrize region.
     parametrize_region = [
         fixture.random_contig(),
-        random.sample(api.contigs, 2),
+        np.random.choice(api.contigs, size=2, replace=False).tolist(),
         fixture.random_region_str(),
     ]
 
@@ -421,7 +420,7 @@ def test_cnv_hmm(fixture, api: AnophelesCnvData):
 def test_cnv_hmm__max_coverage_variance(fixture, api: AnophelesCnvData):
     # Set up test.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_set = random.choice(all_sample_sets)
+    sample_set = str(np.random.choice(all_sample_sets))
     region = fixture.random_contig()
 
     # Parametrize max_coverage_variance.
@@ -465,15 +464,17 @@ def test_cnv_hmm__max_coverage_variance(fixture, api: AnophelesCnvData):
 def test_cnv_coverage_calls(fixture, api: AnophelesCnvData):
     # Parametrize sample_sets.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    parametrize_sample_sets = random.sample(all_sample_sets, 3)
+    parametrize_sample_sets = np.random.choice(
+        all_sample_sets, size=3, replace=False
+    ).tolist()
 
     # Parametrize analysis.
     parametrize_analysis = api.coverage_calls_analysis_ids
 
     # Parametrize region.
     parametrize_region = [
         fixture.random_contig(),
-        random.sample(api.contigs, 2),
+        np.random.choice(api.contigs, size=2, replace=False).tolist(),
         fixture.random_region_str(),
     ]
 
@@ -551,15 +552,15 @@ def test_cnv_discordant_read_calls(fixture, api: AnophelesCnvData):
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     parametrize_sample_sets = [
         None,
-        random.choice(all_sample_sets),
-        random.sample(all_sample_sets, 2),
-        random.choice(all_releases),
+        str(np.random.choice(all_sample_sets)),
+        np.random.choice(all_sample_sets, size=2, replace=False).tolist(),
+        np.random.choice(all_releases),
     ]
 
     # Parametrize contig.
     parametrize_contig = [
-        random.choice(api.contigs),
-        random.sample(api.contigs, 2),
+        str(np.random.choice(api.contigs)),
+        np.random.choice(api.contigs, size=2, replace=False).tolist(),
     ]
 
     for sample_sets in parametrize_sample_sets:
@@ -631,13 +632,13 @@ def test_cnv_discordant_read_calls(fixture, api: AnophelesCnvData):
         match="No CNV discordant read calls data found|no CNVs available for contig",
     ):
         api.cnv_discordant_read_calls(
-            contigs="foobar", sample_sets=random.choice(all_sample_sets)
+            contigs="foobar", sample_sets=str(np.random.choice(all_sample_sets))
         )
 
     # Check with a sample set that should not exist
     with pytest.raises(ValueError):
         api.cnv_discordant_read_calls(
-            contigs=random.choice(api.contigs), sample_sets="foobar"
+            contigs=np.random.choice(api.contigs), sample_sets="foobar"
         )
 
     # Check with a contig and sample set that should not exist
@@ -649,8 +650,8 @@ def test_cnv_discordant_read_calls(fixture, api: AnophelesCnvData):
 def test_cnv_discordant_read_calls_deprecated_contig_alias(
     fixture, api: AnophelesCnvData
 ):
-    sample_set = random.choice(api.sample_sets()["sample_set"].to_list())
-    contig = random.choice(api.contigs)
+    sample_set = str(np.random.choice(api.sample_sets()["sample_set"].to_list()))
+    contig = str(np.random.choice(api.contigs))
     ds_contigs = api.cnv_discordant_read_calls(contigs=contig, sample_sets=sample_set)
     with pytest.warns(DeprecationWarning, match="deprecated"):
         ds_contig = api.cnv_discordant_read_calls(contig=contig, sample_sets=sample_set)
@@ -821,7 +822,7 @@ def test_cnv_discordant_read_calls__sample_query_options(
 def test_plot_cnv_hmm_coverage_track(fixture, api: AnophelesCnvData):
     # Set up test.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_set = random.choice(all_sample_sets)
+    sample_set = str(np.random.choice(all_sample_sets))
     region = fixture.random_contig()
     df_samples = api.sample_metadata(sample_sets=sample_set)
     all_sample_ids = df_samples["sample_id"].values
@@ -874,7 +875,7 @@ def test_plot_cnv_hmm_coverage_track(fixture, api: AnophelesCnvData):
 def test_plot_cnv_hmm_coverage(fixture, api: AnophelesCnvData):
     # Set up test.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_set = random.choice(all_sample_sets)
+    sample_set = str(np.random.choice(all_sample_sets))
     region = fixture.random_contig()
     df_samples = api.sample_metadata(sample_sets=sample_set)
     all_sample_ids = df_samples["sample_id"].values
@@ -928,9 +929,9 @@ def test_plot_cnv_hmm_heatmap_track(fixture, api: AnophelesCnvData):
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     parametrize_sample_sets = [
         None,
-        random.choice(all_sample_sets),
-        random.sample(all_sample_sets, 2),
-        random.choice(all_releases),
+        str(np.random.choice(all_sample_sets)),
+        np.random.choice(all_sample_sets, size=2, replace=False).tolist(),
+        np.random.choice(all_releases),
     ]
 
     for region in parametrize_region:

tests/anoph/test_cnv_frq.py

@@ -1,5 +1,3 @@
-import random
-
 import numpy as np
 import pandas as pd
 import xarray as xr
@@ -95,10 +93,10 @@ def test_gene_cnv_frequencies_with_str_cohorts(
     api: AnophelesCnvFrequencyAnalysis,
     cohorts,
 ):
-    region = random.choice(api.contigs)
+    region = str(np.random.choice(api.contigs))
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_sets = random.choice(all_sample_sets)
-    min_cohort_size = random.randint(0, 2)
+    sample_sets = str(np.random.choice(all_sample_sets))
+    min_cohort_size = int(np.random.randint(0, 3))
 
     # Set up call params.
     params = dict(
@@ -148,8 +146,8 @@ def test_gene_cnv_frequencies_with_min_cohort_size(
 ):
     # Pick test parameters at random.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_sets = random.choice(all_sample_sets)
-    region = random.choice(api.contigs)
+    sample_sets = str(np.random.choice(all_sample_sets))
+    region = str(np.random.choice(api.contigs))
     cohorts = "admin1_year"
 
     # Set up call params.
@@ -199,13 +197,13 @@ def test_gene_cnv_frequencies_with_str_cohorts_and_sample_query(
     # Pick test parameters at random.
     sample_sets = None
     min_cohort_size = 0
-    region = random.choice(api.contigs)
-    cohorts = random.choice(
-        ["admin1_year", "admin1_month", "admin2_year", "admin2_month"]
+    region = str(np.random.choice(api.contigs))
+    cohorts = str(
+        np.random.choice(["admin1_year", "admin1_month", "admin2_year", "admin2_month"])
     )
     df_samples = api.sample_metadata(sample_sets=sample_sets)
     countries = df_samples["country"].unique()
-    country = random.choice(countries)
+    country = str(np.random.choice(countries))
     sample_query = f"country == '{country}'"
 
     # Figure out expected cohort labels.
@@ -247,13 +245,13 @@ def test_gene_cnv_frequencies_with_str_cohorts_and_sample_query_options(
     # Pick test parameters at random.
     sample_sets = None
     min_cohort_size = 0
-    region = random.choice(api.contigs)
-    cohorts = random.choice(
-        ["admin1_year", "admin1_month", "admin2_year", "admin2_month"]
+    region = str(np.random.choice(api.contigs))
+    cohorts = str(
+        np.random.choice(["admin1_year", "admin1_month", "admin2_year", "admin2_month"])
     )
     df_samples = api.sample_metadata(sample_sets=sample_sets)
     countries = df_samples["country"].unique().tolist()
-    countries_list = random.sample(countries, 2)
+    countries_list = np.random.choice(countries, size=2, replace=False).tolist()
     sample_query_options = {
         "local_dict": {
             "countries_list": countries_list,
@@ -303,8 +301,8 @@ def test_gene_cnv_frequencies_with_dict_cohorts(
 ):
     # Pick test parameters at random.
     sample_sets = None  # all sample sets
-    min_cohort_size = random.randint(0, 2)
-    region = random.choice(api.contigs)
+    min_cohort_size = int(np.random.randint(0, 3))
+    region = str(np.random.choice(api.contigs))
 
     # Create cohorts by country.
     df_samples = api.sample_metadata(sample_sets=sample_sets)
@@ -343,10 +341,10 @@ def test_gene_cnv_frequencies_without_drop_invariant(
 ):
     # Pick test parameters at random.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_sets = random.choice(all_sample_sets)
-    min_cohort_size = random.randint(0, 2)
-    region = random.choice(api.contigs)
-    cohorts = random.choice(["admin1_year", "admin2_month", "country"])
+    sample_sets = str(np.random.choice(all_sample_sets))
+    min_cohort_size = int(np.random.randint(0, 3))
+    region = str(np.random.choice(api.contigs))
+    cohorts = str(np.random.choice(["admin1_year", "admin2_month", "country"]))
 
     # Figure out expected cohort labels.
     df_samples = api.sample_metadata(sample_sets=sample_sets)
@@ -398,9 +396,9 @@ def test_gene_cnv_frequencies_with_bad_region(
 ):
     # Pick test parameters at random.
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_sets = random.choice(all_sample_sets)
-    min_cohort_size = random.randint(0, 2)
-    cohorts = random.choice(["admin1_year", "admin2_month", "country"])
+    sample_sets = str(np.random.choice(all_sample_sets))
+    min_cohort_size = int(np.random.randint(0, 3))
+    cohorts = str(np.random.choice(["admin1_year", "admin2_month", "country"]))
 
     # Set up call params.
     params = dict(
@@ -424,9 +422,9 @@ def test_gene_cnv_frequencies_with_max_coverage_variance(
     max_coverage_variance,
 ):
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
-    sample_sets = random.choice(all_sample_sets)
-    cohorts = random.choice(["admin1_year", "admin2_month", "country"])
-    region = random.choice(api.contigs)
+    sample_sets = str(np.random.choice(all_sample_sets))
+    cohorts = str(np.random.choice(["admin1_year", "admin2_month", "country"]))
+    region = str(np.random.choice(api.contigs))
 
     params = dict(
         region=region,
@@ -503,7 +501,7 @@ def test_gene_cnv_frequencies_advanced_with_sample_query(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     df_samples = api.sample_metadata(sample_sets=all_sample_sets)
     countries = df_samples["country"].unique()
-    country = random.choice(countries)
+    country = str(np.random.choice(countries))
     sample_query = f"country == '{country}'"
 
     check_gene_cnv_frequencies_advanced(
@@ -522,7 +520,7 @@ def test_gene_cnv_frequencies_advanced_with_sample_query_options(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     df_samples = api.sample_metadata(sample_sets=all_sample_sets)
     countries = df_samples["country"].unique().tolist()
-    countries_list = random.sample(countries, 2)
+    countries_list = np.random.choice(countries, size=2, replace=False).tolist()
     sample_query_options = {
         "local_dict": {
             "countries_list": countries_list,
@@ -549,7 +547,7 @@ def test_gene_cnv_frequencies_advanced_with_min_cohort_size(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     area_by = "admin1_iso"
     period_by = "year"
-    region = random.choice(api.contigs)
+    region = str(np.random.choice(api.contigs))
 
     if min_cohort_size <= 10:
         # Expect this to find at least one cohort, so go ahead with full
@@ -585,7 +583,7 @@ def test_gene_cnv_frequencies_advanced_with_max_coverage_variance(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     area_by = "admin1_iso"
     period_by = "year"
-    region = random.choice(api.contigs)
+    region = str(np.random.choice(api.contigs))
 
     if max_coverage_variance >= 0.4:
         # Expect this to find at least one cohort, so go ahead with full
@@ -620,7 +618,7 @@ def test_gene_cnv_frequencies_advanced_with_nobs_mode(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     area_by = "admin1_iso"
     period_by = "year"
-    region = random.choice(api.contigs)
+    region = str(np.random.choice(api.contigs))
 
     check_gene_cnv_frequencies_advanced(
         api=api,
@@ -642,7 +640,7 @@ def test_gene_cnv_frequencies_advanced_with_variant_query(
     all_sample_sets = api.sample_sets()["sample_set"].to_list()
     area_by = "admin1_iso"
     period_by = "year"
-    region = random.choice(api.contigs)
+    region = str(np.random.choice(api.contigs))
     variant_query = f"cnv_type == '{variant_query_option}'"
 
     check_gene_cnv_frequencies_advanced(
@@ -710,16 +708,16 @@ def check_gene_cnv_frequencies_advanced(
 ):
     # Pick test parameters at random.
     if region is None:
-        region = random.choice(api.contigs)
+        region = str(np.random.choice(api.contigs))
     if area_by is None:
-        area_by = random.choice(["country", "admin1_iso", "admin2_name"])
+        area_by = str(np.random.choice(["country", "admin1_iso", "admin2_name"]))
     if period_by is None:
-        period_by = random.choice(["year", "quarter", "month", "random_year"])
+        period_by = str(np.random.choice(["year", "quarter", "month", "random_year"]))
     if sample_sets is None:
         all_sample_sets = api.sample_sets()["sample_set"].to_list()
-        sample_sets = random.choice(all_sample_sets)
+        sample_sets = str(np.random.choice(all_sample_sets))
     if min_cohort_size is None:
-        min_cohort_size = random.randint(0, 2)
+        min_cohort_size = int(np.random.randint(0, 3))
 
     if period_by == "random_year":
         # Add a random_year column to the sample metadata, if there isn't already.