From 720d06900ebf5a57b0dcfe221701fcc64bb9787d Mon Sep 17 00:00:00 2001
From: Micha Birklbauer <micha.birklbauer@gmail.com>
Date: Thu, 11 Jan 2024 13:44:04 +0100
Subject: [PATCH] finalize tests

---
 tests/tests.py | 102 +++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 102 insertions(+)

diff --git a/tests/tests.py b/tests/tests.py
index 099655f..a785db1 100644
--- a/tests/tests.py
+++ b/tests/tests.py
@@ -74,3 +74,105 @@ def test5_msannika_fdr():
 
     # check nr of grouped crosslinks
     assert result[0].shape[0] == 3192
+
+# csms file with fdr should validate
+def test6_msannika_fdr():
+
+    import pandas as pd
+    from msannika_fdr import main
+    from msannika_fdr import MSAnnika_CSM_Validator as csm_val
+
+    result = main(["DSSO_CSMs.xlsx", "-fdr", "0.01"])
+
+    # check nr of results
+    assert len(result) == 3
+
+    #### crosslinks ####
+
+    # check nr of grouped crosslinks
+    assert result[0].shape[0] == 3192
+
+    #### validated crosslinks ####
+
+    # check nr of validated crosslinks
+    assert result[2].shape[0] == 714
+
+    # check fdr of validated crosslinks
+    assert result[2][result[2]["Decoy"] == True].shape[0] / result[2][result[2]["Decoy"] == False].shape[0] < 0.01
+
+    #### validated csms ####
+
+    # check nr of validated csms
+    assert result[1].shape[0] == 3021
+
+    # check fdr of validated csms
+    results[1]["Class"] = pd.apply(lambda row: csm_val.get_class(row), axis = 1)
+    assert result[1][result[1]["Class"] == "Decoy"].shape[0] / result[1][result[1]["Class"] == "Target"].shape[0] < 0.01
+
+# csms file with fdr should validate
+def test7_msannika_fdr():
+
+    import pandas as pd
+    from msannika_fdr import main
+    from msannika_fdr import MSAnnika_CSM_Validator as csm_val
+
+    result = main(["DSSO_CSMs.xlsx", "-fdr", "1"])
+
+    # check nr of results
+    assert len(result) == 3
+
+    #### crosslinks ####
+
+    # check nr of grouped crosslinks
+    assert result[0].shape[0] == 3192
+
+    #### validated crosslinks ####
+
+    # check nr of validated crosslinks
+    assert result[2].shape[0] == 714
+
+    # check fdr of validated crosslinks
+    assert result[2][result[2]["Decoy"] == True].shape[0] / result[2][result[2]["Decoy"] == False].shape[0] < 0.01
+
+    #### validated csms ####
+
+    # check nr of validated csms
+    assert result[1].shape[0] == 3021
+
+    # check fdr of validated csms
+    results[1]["Class"] = pd.apply(lambda row: csm_val.get_class(row), axis = 1)
+    assert result[1][result[1]["Class"] == "Decoy"].shape[0] / result[1][result[1]["Class"] == "Target"].shape[0] < 0.01
+
+# csms file with fdr should validate
+def test8_msannika_fdr():
+
+    import pandas as pd
+    from msannika_fdr import main
+    from msannika_fdr import MSAnnika_CSM_Validator as csm_val
+
+    result = main(["DSSO_CSMs.xlsx", "-fdr", "0.05"])
+
+    # check nr of results
+    assert len(result) == 3
+
+    #### crosslinks ####
+
+    # check nr of grouped crosslinks
+    assert result[0].shape[0] == 3192
+
+    #### validated crosslinks ####
+
+    # check nr of validated crosslinks
+    assert result[2].shape[0] == 807
+
+    # check fdr of validated crosslinks
+    assert result[2][result[2]["Decoy"] == True].shape[0] / result[2][result[2]["Decoy"] == False].shape[0] < 0.01
+
+    #### validated csms ####
+
+    # check nr of validated csms
+    assert result[1].shape[0] == 3331
+
+    # check fdr of validated csms
+    results[1]["Class"] = pd.apply(lambda row: csm_val.get_class(row), axis = 1)
+    assert result[1][result[1]["Class"] == "Decoy"].shape[0] / result[1][result[1]["Class"] == "Target"].shape[0] < 0.01