add unittests, fix absolute links

Author: peitek

analysis/BehavioralBrain.py

@@ -1,17 +1,16 @@
 import numpy
 import pandas as pd
 import matplotlib.pyplot as plt
-
-import BrainActivationAnalysis
-from analysis import BehavioralSubjective
-
-plt.rcParams.update({'font.size': 26})
 import seaborn as sns
 import scipy
 
-# TODO this file is supposed to analyze complexity metrics versus behavioral data
+from analysis import BehavioralSubjective, BrainActivationAnalysis
+from config import ROOT_DIR
+
+plt.rcParams.update({'font.size': 26})
 graph_label = dict(color='#202020', alpha=0.9)
 
+
 def plot_correlation(df, computeResponseTime, ba, activation=True):
     if computeResponseTime:
         variable = 'ResponseTime'
@@ -53,11 +52,10 @@ def plot_correlation(df, computeResponseTime, ba, activation=True):
     else:
         pre = 'Deactivation_'
 
-    plt.savefig('output/' + pre + ba + '_' + variable + '.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/' + pre + ba + '_' + variable + '.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def compute_behavioral_brain(df_ba_cond, behavioral_data, activation):
-
     behavioral_ba = pd.merge(df_ba_cond, behavioral_data, how='left', left_on=['participant', 'Snippet'], right_on=['Participant', 'Snippet'])
 
     # check whether there is response times < 5s and exclude them since they are accidental clicks
@@ -76,8 +74,8 @@ def compute_behavioral_brain(df_ba_cond, behavioral_data, activation):
 
 
 def main():
-    df_ba_part_cond_act = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Activation.csv')
-    df_ba_part_cond_deact = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Deactivation.csv')
+    df_ba_part_cond_act = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Activation.csv')
+    df_ba_part_cond_deact = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Deactivation.csv')
 
     behavioral_data = BehavioralSubjective.load_behavioral_data()
 

analysis/BehavioralSubjective.py

@@ -7,14 +7,16 @@
 import seaborn as sns
 import scipy
 
+from config import ROOT_DIR
+
 plt.rcParams.update({'font.size': 26})
 
 graph_label = dict(color='#202020', alpha=0.9)
 
 
 def load_snippet_metrics():
     print('\n##### \nread complexity metric data')
-    snippet_metrics = pd.read_csv("../data/metrics/SnippetComplexityMetricsValues.csv", delimiter=",")
+    snippet_metrics = pd.read_csv(ROOT_DIR + "/data/metrics/SnippetComplexityMetricsValues.csv", delimiter=",")
     snippet_metrics = snippet_metrics.astype('float', errors='ignore')
     print(snippet_metrics.head(5))
     return snippet_metrics
@@ -23,7 +25,7 @@ def load_snippet_metrics():
 def load_behavioral_data():
     print('\n##### \nread behavioral data')
 
-    behavioral_data = pd.read_csv("../data/behavioral/ParticipantBehavior.csv", delimiter=",")
+    behavioral_data = pd.read_csv(ROOT_DIR + "/data/behavioral/ParticipantBehavior.csv", delimiter=",")
     behavioral_data = behavioral_data.astype('float', errors='ignore')
 
     behavioral_data = behavioral_data[behavioral_data['Condition'] == "Comprehension"]
@@ -64,7 +66,7 @@ def plot_correlation_correctness(df, metric):
     sns.despine()
     plt.tight_layout()
 
-    plt.savefig('output/' + metric + '_Correctness.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/' + metric + '_Correctness.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def plot_correlation_responsetime(df, metric):
@@ -97,7 +99,7 @@ def plot_correlation_responsetime(df, metric):
     sns.despine()
     plt.tight_layout()
 
-    plt.savefig('output/' + metric + '_ResponseTime.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/' + metric + '_ResponseTime.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def plot_correlation_subjcomplexity_metrics(df, metric):
@@ -130,7 +132,7 @@ def plot_correlation_subjcomplexity_metrics(df, metric):
     sns.despine()
     plt.tight_layout()
 
-    plt.savefig('output/SubjComplexity_' + metric + '.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/SubjComplexity_' + metric + '.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def plot_correlation_subjcomplexity_responsetime(df):
@@ -158,7 +160,7 @@ def plot_correlation_subjcomplexity_responsetime(df):
     sns.despine()
     plt.tight_layout()
 
-    plt.savefig('output/SubjComplexity_ResponseTime.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/SubjComplexity_ResponseTime.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def plot_correlation_subjcomplexity_correctness(df):
@@ -186,7 +188,7 @@ def plot_correlation_subjcomplexity_correctness(df):
     sns.despine()
     plt.tight_layout()
 
-    plt.savefig('output/SubjComplexity_Correctness.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
+    plt.savefig(ROOT_DIR + '/analysis/output/SubjComplexity_Correctness.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
 
 def select_color_for_metric(metric):
@@ -242,7 +244,7 @@ def main():
 
     # correlate with behavioral data
     print('\n##### \n correlating subjective complexity with behavioral data')
-    snippet_subjective_complexity = pd.read_csv('../data/subjective/SnippetSubjectiveComplexityRatings.csv')
+    snippet_subjective_complexity = pd.read_csv(ROOT_DIR + '/data/subjective/SnippetSubjectiveComplexityRatings.csv')
 
     print('mean:' + str(snippet_subjective_complexity['subj_complexity'].mean()))
     print('std:' + str(snippet_subjective_complexity['subj_complexity'].std(ddof=1)))

analysis/BrainActivationAnalysis.py

@@ -8,11 +8,12 @@
 import statsmodels.api as sm
 from statsmodels.formula.api import ols
 
+from analysis import BehavioralSubjective
+from config import ROOT_DIR
+
 plt.rcParams.update({'font.size': 24})
 plt.rcParams['font.family'] = 'Calibri'
 
-import BehavioralSubjective
-
 graph_label = dict(color='#101010', alpha=0.95)
 
 
@@ -47,9 +48,9 @@ def plot_ba_subj_rating(df, ba, activation=True, participant=None):
     plt.tight_layout()
 
     if activation:
-        prefix = 'output/act_subj_'
+        prefix = ROOT_DIR + '/analysis/output/act_subj_'
     else:
-        prefix = 'output/deact_subj_'
+        prefix = ROOT_DIR + '/analysis/output/deact_subj_'
 
     if participant:
         prefix += participant + '_'
@@ -112,9 +113,9 @@ def plot_ba_for_metric(df, metric, ba, activation=True):
     plt.tight_layout()
 
     if activation:
-        prefix = 'output/activation_'
+        prefix = ROOT_DIR + '/analysis/output/activation_'
     else:
-        prefix = 'output/deactivation_'
+        prefix = ROOT_DIR + '/analysis/output/deactivation_'
 
     plt.savefig(prefix + metric + '_' + ba + '.pdf', dpi=300, bbox_inches='tight', pad_inches=0)
 
@@ -154,11 +155,11 @@ def compute_statistics(df, activation=True):
 
 
 def compute_ba_subj_rating(activation=True):
-    df_subj_complexity = pd.read_csv('../data/subjective/SnippetSubjectiveComplexityRatings.csv')
+    df_subj_complexity = pd.read_csv(ROOT_DIR + '/data/subjective/SnippetSubjectiveComplexityRatings.csv')
     if activation:
-        df_ba_part_cond = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Activation.csv')
+        df_ba_part_cond = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Activation.csv')
     else:
-        df_ba_part_cond = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Deactivation.csv')
+        df_ba_part_cond = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Participant_Deactivation.csv')
 
     df_ba_part_cond_subj = pd.merge(df_ba_part_cond, df_subj_complexity, left_on=['participant', 'Snippet'], right_on=['participant', 'snippet'])
     df_ba_part_cond_subj.sort_values(by='participant', inplace=True)
@@ -174,8 +175,8 @@ def compute_ba_subj_rating(activation=True):
 
 def main():
     print('\n# Analyzing fMRI and complexity metrics data')
-    df_ba_cond_act = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Activation.csv')
-    df_ba_cond_deact = pd.read_csv('../data/fMRI/fMRI_Analyzed_BA_Snippet_Deactivation.csv')
+    df_ba_cond_act = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Activation.csv')
+    df_ba_cond_deact = pd.read_csv(ROOT_DIR + '/data/fMRI/fMRI_Analyzed_BA_Snippet_Deactivation.csv')
 
     snippet_metrics = BehavioralSubjective.load_snippet_metrics()
 

analysis/__init__.py

@@ -0,0 +1,3 @@
+import os
+
+ROOT_DIR = os.path.dirname(os.path.abspath(__file__))

config.py

@@ -0,0 +1,55 @@
+import unittest
+import os
+
+from analysis import BehavioralSubjective, BehavioralBrain, BrainActivationAnalysis
+from config import ROOT_DIR
+
+OUTPUT_DIR = ROOT_DIR + '/analysis/output'
+
+
+def empty_output_dir():
+    files = [f for f in os.listdir(OUTPUT_DIR) if f.endswith(".pdf")]
+    for file in files:
+        os.remove(os.path.join(OUTPUT_DIR, file))
+
+
+def get_number_of_files_in_output():
+    path, dirs, files = next(os.walk(OUTPUT_DIR))
+    file_count = len(files)
+    return file_count
+
+
+class PipelineTest(unittest.TestCase):
+    def test_behavioral_subjective(self):
+        empty_output_dir()
+
+        file_count_initial = get_number_of_files_in_output()
+        self.assertEqual(1, file_count_initial)  # only .gitkeep file
+
+        BehavioralSubjective.main()
+        file_count_end = get_number_of_files_in_output()
+        self.assertEqual(15, file_count_end)  # pipeline should create 14 files + .gitkeep
+
+    def test_behavioral_brain(self):
+        empty_output_dir()
+
+        file_count_initial = get_number_of_files_in_output()
+        self.assertEqual(1, file_count_initial)  # only .gitkeep file
+
+        BehavioralBrain.main()
+        file_count_end = get_number_of_files_in_output()
+        self.assertEqual(13, file_count_end)  # pipeline should create 12 files + .gitkeep
+
+    def test_brain_activation(self):
+        empty_output_dir()
+
+        file_count_initial = get_number_of_files_in_output()
+        self.assertEqual(1, file_count_initial)  # only .gitkeep file
+
+        BrainActivationAnalysis.main()
+        file_count_end = get_number_of_files_in_output()
+        self.assertEqual(31, file_count_end)  # pipeline should create 30 files + .gitkeep
+
+
+if __name__ == '__main__':
+    unittest.main()