plot generator

.gitignore

@@ -0,0 +1,3 @@
+.idea/
+historical_wer_mean_plot.png
+historical_srr_plot.png

Jenkinsfile

@@ -5,20 +5,15 @@ pipeline {
              steps {
                  script {
                      step ([$class: 'CopyArtifact',
-                     projectName: 'ASR-eval',
+                     projectName: 's407259-metrics',
-                     target: 'Infra']);
+                     target: 's407259-matrics-data']);
 
 
                      sh 'git --version'
                      sh 'git checkout'
-                     sh 'python3 script.py'
+                     sh 'python3 generate_plot.py'
-                     archiveArtifacts artifacts: 'num_lines.txt'
+                     archiveArtifacts artifacts: 'historical_srr_plot.png'
-                     sh 'sudo python3 calculate_metrics.py'
+                     archiveArtifacts artifacts: 'historical_wer_mean_plot.png'
-                     archiveArtifacts artifacts: 'wikiniews_results_with_wer.tsv'
-                     archiveArtifacts artifacts: 'wer_mean.txt'
-                     archiveArtifacts artifacts: 'srr.txt'
-                     archiveArtifacts artifacts: 'historical_wer_mean.txt'
-                     archiveArtifacts artifacts: 'historical_srr.txt'
 
                  }
              }

calculate_metrics.py

@@ -1,39 +0,0 @@
-import pandas as pd
-from jiwer import wer
-
-df_data = pd.read_csv('./Infra/wikiniews_results.tsv', sep='\t', index_col=False, header=None,
-                      skip_blank_lines=False, keep_default_na=False, names=[1,2,3,4,5])
-
-correct_answers = 0
-
-for index, row in df_data.iterrows():
-
-
-    wer_result = wer(row[2], row[3])
-
-    df_data.loc[index, 5] = wer_result
-
-    if wer_result == 0.0:
-
-        correct_answers += 1
-
-
-wer_mean = df_data[5].mean()
-print(wer_mean)
-
-srr = (correct_answers*1)/len(df_data)
-print(srr)
-
-df_data.to_csv('./wikiniews_results_with_wer.tsv',sep='\t', header=None)
-
-with open("wer_mean.txt", "w") as text_file:
-    text_file.write(str(df_data[5].mean()))
-
-with open("srr.txt", "w") as text_file:
-    text_file.write(str(srr))
-
-with open("historical_wer_mean.txt", "a") as text_file:
-    text_file.write(str(df_data[5].mean()) + '\n')
-
-with open("historical_srr.txt", "a") as text_file:
-    text_file.write(str(srr) + '\n')

generate_plot.py

@@ -0,0 +1,31 @@
+import pandas as pd
+from matplotlib.ticker import MaxNLocator
+
+data = pd.read_csv('./s407259-matrics-data/historical_wer_mean.txt', header=None)
+
+plot = data.tail(50).plot()
+
+plot.xaxis.set_major_locator(MaxNLocator(integer=True))
+
+plot.get_legend().remove()
+
+plot.set_xlabel('sample')
+
+plot.set_ylabel('wer')
+
+plot.get_figure().savefig('./historical_wer_mean_plot')
+
+
+data = pd.read_csv('./s407259-matrics-data/historical_srr.txt', header=None)
+
+plot = data.tail(50).plot()
+
+plot.xaxis.set_major_locator(MaxNLocator(integer=True))
+
+plot.get_legend().remove()
+
+plot.set_xlabel('sample')
+
+plot.set_ylabel('srr')
+
+plot.get_figure().savefig('./historical_srr_plot')

requirements.txt

@@ -1,2 +1,2 @@
 pandas
-jiwer
+matplotlib

script.py

@@ -1,13 +0,0 @@
-print('starting 2')
-import sys
-from importlib import reload
-import os
-
-reload(sys)
-num_lines = os.popen('wc -l ./Infra/wikiniews_results.tsv').read()
-
-print('lines number')
-print(num_lines)
-file = open('num_lines.txt', 'w')
-file.write(str(num_lines))
-file.close()