def test_dropping_all_samples_and_features():
"""Tests Qurro's behavior when > 0 feature(s) and all samples are unsupported.
In particular, qurro should just throw an error about the
feature(s) being unsupported -- the feature check should come before the
sample check (not for any particular reason, that's just how I wrote the
code).
"""
run_integration_test(
"matching_test",
"matching_test/all_samples_and_one_feature_dropped",
"differentials.tsv",
"all_samples_and_one_feature_dropped.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expect_all_unsupported_samples=True,
expected_unsupported_features=1,
)
run_integration_test(
"matching_test",
"matching_test/all_samples_and_two_features_dropped",
"differentials.tsv",
"all_samples_and_two_features_dropped.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expect_all_unsupported_samples=True,
expected_unsupported_features=2,
)
def test_dropping_all_samples():
"""Tests that Qurro raises an error when all samples are unsupported."""
run_integration_test(
"matching_test",
"matching_test/all_samples_dropped",
"differentials.tsv",
"all_samples_dropped.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expect_all_unsupported_samples=True,
)
def test_integration_q2_byrd():
"""Tests qurro on songbird output in the context of QIIME 2."""
run_integration_test(
"byrd",
"q2_byrd",
"byrd_differentials.tsv",
"byrd_skin_table.biom",
"byrd_metadata.txt",
use_q2=True,
q2_ranking_tool="songbird",
)
def test_integration_q2_sleep_apnea():
"""Tests qurro on DEICODE output in the context of QIIME 2."""
run_integration_test(
"sleep_apnea",
"q2_sleep_apnea",
"ordination.txt",
"qiita_10422_table.biom",
"qiita_10422_metadata.tsv",
feature_metadata_name="taxonomy.tsv",
use_q2=True,
q2_ranking_tool="DEICODE",
)
def test_byrd():
"""Tests Qurro's JSON generation on the Byrd et al. 2017 dataset.
This is really a test to make sure that qurro can properly handle
songbird output.
"""
run_integration_test(
"byrd",
"byrd",
"byrd_differentials.tsv",
"byrd_skin_table.biom",
"byrd_metadata.txt",
)
def test_moving_pictures():
"""Tests qurro' JSON generation on the "Moving Pictures of the Human
Microbiome" dataset.
"""
run_integration_test(
"moving_pictures",
"q2_moving_pictures",
"ordination.txt",
"feature-table.biom",
"sample-metadata.tsv",
feature_metadata_name="taxonomy.tsv",
use_q2=True,
q2_ranking_tool="DEICODE",
)
def test_sleep_apnea():
"""Tests Qurro's JSON generation on a "sleep apnea" dataset.
This is really a test to make sure that qurro can properly handle
DEICODE output.
"""
run_integration_test(
"sleep_apnea",
"sleep_apnea",
"ordination.txt",
"qiita_10422_table.biom",
"qiita_10422_metadata.tsv",
feature_metadata_name="taxonomy.tsv",
)
def test_red_sea():
"""Tests Qurro's JSON generation on a dataset from a study of the Red
Sea.
This is really a test to make sure that qurro can properly handle
this sort of unconventionally-named-feature data.
"""
run_integration_test(
"red_sea",
"red_sea",
"differentials.tsv",
"redsea.biom",
"redsea_metadata.txt",
feature_metadata_name="feature_metadata.txt",
)
def test_mackerel():
"""Tests Qurro's JSON generation on the mackerel dataset (Qiita Study ID
11721).
This is less of a test and more of just a way to ensure that we
continuously update the mackerel demo whenever we update Qurro.
"""
run_integration_test(
"mackerel",
"mackerel",
"differentials.tsv",
"feature-table.biom",
"sample-metadata.tsv",
feature_metadata_name="taxonomy.tsv",
expected_unsupported_samples=1067,
)
def test_integration_q2_red_sea():
"""Tests qurro on songbird output with non-strictly-16S data in the
context of QIIME 2.
(Technically the Byrd data is also from a metagenomic study, but the
features here are named pretty differently.)
"""
run_integration_test(
"red_sea",
"q2_red_sea",
"differentials.tsv",
"redsea.biom",
"redsea_metadata.txt",
feature_metadata_name="feature_metadata.txt",
use_q2=True,
q2_ranking_tool="songbird",
)
def test_empty_sample_integration_basic():
rpj, spj, cj = run_integration_test(
*params, feature_metadata_name="feature_metadata.txt")
# look at sample plot json -- verify that just Sample2 dropped (due to
# being empty: having a count of 0 for every feature)
sample_data = get_data_from_plot_json(spj)
for s in ("Sample1", "Sample3", "Sample5", "Sample6", "Sample7"):
assert s in sample_data
assert "Sample4" not in sample_data
assert "Sample2" not in sample_data
def test_dropping_features():
"""Tests that Qurro raises an error when > 0 feature(s) are unsupported."""
# Test that dropping 1 feature produces an error
run_integration_test(
"matching_test",
"matching_test/dropped_feature",
"differentials.tsv",
"dropped_feature.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expected_unsupported_features=1,
)
# Test that dropping 2 features produces an error
run_integration_test(
"matching_test",
"matching_test/dropped_features",
"differentials.tsv",
"dropped_features.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expected_unsupported_features=2,
)
def test_sample_dropping_stats():
"""Tests Qurro's JSON generation on a dataset with weird sample
metadata.
"""
rank_json, sample_json, count_json = run_integration_test(
"sample_stats_test",
"q2_sample_stats_test",
"differentials.tsv",
"sst.biom",
"sample_metadata.txt",
expected_unsupported_samples=1,
use_q2=True,
q2_ranking_tool="songbird",
q2_table_biom_format="BIOMV100Format",
)
validate_sample_stats_test_sample_plot_json(sample_json)
def test_sample_dropping_stats():
"""Tests Qurro's JSON generation on a dataset with weird sample
metadata.
The output from this test will be used in JS tests -- this verifies that
both the python and JS parts of Qurro can handle weird metadata
appropriately.
"""
rank_json, sample_json, count_json = run_integration_test(
"sample_stats_test",
"sample_stats_test",
"differentials.tsv",
"sst.biom",
"sample_metadata.txt",
expected_unsupported_samples=1,
)
validate_sample_stats_test_sample_plot_json(sample_json)
def test_empty_sample_integration_extreme_feature_count():
params_copy = params[:]
params_copy[1] = "empty_samples_yes_efc"
rpj2, spj2, cj2 = run_integration_test(
*params_copy,
feature_metadata_name="feature_metadata.txt",
extreme_feature_count=1,
)
# look at sample plot json -- verify that Sample2 + 3 dropped
# (Sample2 due to being empty, and Sample3 due to being empty after the one
# feature it had a count for, Taxon3, was removed due to -x.)
# (Also, verify that most of the features were dropped due to -x.)
sample_data_2 = get_data_from_plot_json(spj2)
for s2 in ("Sample1", "Sample5", "Sample6", "Sample7"):
assert s2 in sample_data_2
assert "Sample4" not in sample_data_2
assert "Sample2" not in sample_data_2
assert "Sample3" not in sample_data_2
def test_feature_metadata_and_dropped_sample():
"""Tests the behavior of Qurro in matching sample metadata, feature
metadata, ranks, and the BIOM table together.
"""
rank_json, sample_json, count_json = run_integration_test(
"matching_test",
"matching_test",
"differentials.tsv",
"mt.biom",
"sample_metadata.txt",
feature_metadata_name="feature_metadata.txt",
expected_unsupported_samples=1,
)
data_name = rank_json["data"]["name"]
# Check that feature metadata annotations were done correctly
for feature in rank_json["datasets"][data_name]:
txid = feature["Feature ID"]
if feature["Feature ID"] == "Taxon3":
assert feature["FeatureMetadata1"] == "Yeet"
# this should be interpreted this as a string, since "lol" is in
# the same column
assert feature["FeatureMetadata2"] == "100"
elif feature["Feature ID"] == "Taxon5":
# This should be interpeted as a string, also
assert feature["FeatureMetadata1"] == "null"
assert feature["FeatureMetadata2"] == "lol"
else:
# Check that the other taxa haven't been annotated with any
# metadata.
assert feature["FeatureMetadata1"] is None
assert feature["FeatureMetadata2"] is None
# Assert that Sample4 was dropped from the "main" dataset of the plot
data_name = sample_json["data"]["name"]
for sample in sample_json["datasets"][data_name]:
assert sample["Sample ID"] != "Sample4"
for txid in count_json:
# Assert that Sample4 was also dropped from the counts data in the JSON
assert "Sample4" not in count_json[txid]
