comparison scripts/output_metadata.py @ 98:5eba53437755 branch-tests

notebooks for publication

 import numpy as np
 import pandas as pd
 import pickle 
 import os
-
-%load_ext autoreload
-%autoreload 2
 
 import sys
 sys.path.append('../')
 import scripts.outliers as outliers
 import scripts.utils as utils

 X_list, Y, Yaudio = dataset
 X = np.concatenate(X_list, axis=1)
 
 cols_to_keep = ['Country', 'continent', 'REGION', 'LocDetails',
                 'Language', 'Language_iso3', 'Culture', 'Genre_Album', 'Year', 'Decade', 
-                 'songurls_Album', 'Speech', 'Melspec', 'Chroma', 'Melodia']
+                 'songurls_Album', 'Audio', 'Speech', 'Melspec', 'Chroma', 'Melodia']
 cols_rename = ['Country', 'Continent', 'Region', 'Location_details', 
                'Language', 'Language_iso3', 'Culture', 'Genre', 'Year', 'Decade',
-                'Url', 'Speech', 'Melspec', 'Chroma', 'Melodia']
+                'Url', 'Audio', 'Speech', 'Melspec', 'Chroma', 'Melodia']
 col_idx = []
 for column in cols_to_keep:
     if column not in ddf.columns:
         print column
     col_idx.append(np.where(ddf.columns==column)[0])

 ddf_new.columns = cols_rename
 
 for i, yy in enumerate(ddf_new['Audio']):
     new_name = os.path.split(os.path.split(yy)[0])[-1] + '_' + os.path.split(yy)[-1]
     new_csv_name = new_name.split('.')[0]+'.csv'
-    ddf_new['Audio'].iloc[i] = os.path.join('Audio', new_name)
-    ddf_new['Chroma'].iloc[i] = os.path.join('Chroma', new_csv_name)
-    ddf_new['Melspec'].iloc[i] = os.path.join('Melspec', new_csv_name)
-    ddf_new['Melodia'].iloc[i] = os.path.join('Melodia', new_csv_name)
-    ddf_new['Speech'].iloc[i] = os.path.join('Speech', new_csv_name)
+    ddf_new.loc[i, 'Audio'] = new_name
+    ddf_new.loc[i, 'Chroma'] = os.path.join('Chroma', new_csv_name)
+    ddf_new.loc[i, 'Melspec'] = os.path.join('Melspec', new_csv_name)
+    ddf_new.loc[i, 'Melodia'] = os.path.join('Melodia', new_csv_name)
+    ddf_new.loc[i, 'Speech'] = os.path.join('Speech', new_csv_name)
     Yaudio[i] = new_name
+
+print ddf_new.head()
+print Yaudio[:20]
 
 ddf_new.to_csv('../data/metadata.csv', index=False)
 pickle.dump([X_list, Y, Yaudio], open('../data/lda_data_8.pickle', 'wb'))