Add program for removing entries such as duplicates

Author: hillwithsmallfields

model/database.py

@@ -49,6 +49,12 @@ def database_init(config):
 def get_collection_rows(collection_name):
     return [row for row in database[collection_names[collection_name]].find({})]
 
+def delete_by_link_id(del_link_id):
+    """Not for normal use.
+    Probably should only be used from command line programs."""
+    database[collection_names['people']].delete_one({'link_id': identification})
+    database[collection_names['profiles']].delete_one({'link_id': identification})
+
 def get_person_dict(identification):
     """Read the data for a person from the database, as a dictionary."""
     if identification is None:

utils/deleter.py

@@ -0,0 +1,41 @@
+#!/usr/bin/env python
+
+from __future__ import print_function
+
+import os
+import sys
+sys.path.append('model')
+sys.path.append(os.path.dirname(os.path.dirname(__file__)))
+import argparse
+import model.configuration as configuration
+import csv
+import model.database as database
+
+def main():
+    """Program to remove person entries completely.
+    Originally meant for removing accidental duplicates."""
+    parser = argparse.ArgumentParser()
+    parser.add_argument("-d", "--deletions",
+                        help="""File containing the link_ids to delete, one per line""")
+    parser.add_argument("-f", "--for-real",
+                        action='store_true',
+                        help="""Without this flag, only do a dummy run.""")
+    args = parser.parse_args()
+    for_real = args.for_real
+    config = configuration.get_config()
+    db_config = config['database']
+    collection_names = db_config['collections']
+    database.database_init(config)
+    count = 0
+    with open(args.deletions) as deletions_file:
+        count += 1
+        for del_link_id in deletions_file.readlines():
+        if for_real:
+            print("Deleting", del_link_id)
+            database.delete_by_link_id(del_link_id)
+        else:
+            print("Would delete", del_link_id)
+    print("Processed", count, "deletion records.")
+
+if __name__ == "__main__":
+    main()

utils/export-raw.py

@@ -27,7 +27,7 @@ def main():
     for row in rows:
         for key in row.keys():
             if key not in keys:
-                keys.push(key)
+                keys.append(key)
     print ("keys are", keys)
     with open(output_name, 'w') as csvfile:
         writer = csv.DictWriter(csvfile, keys)