How-to-port-old-bob-database-interface-to-the-new-csv-format.md

@@ -18,7 +18,7 @@ $ conda create -n bob8_py36 \
 -c https://www.idiap.ch/software/bob/conda \
 -c defaults \
 -c https://www.idiap.ch/software/bob/conda/label/archive \
-python=3.6 bob=8 bob.pad.face bob.db.replaymobile
+python=3.6 bob=8 bob.pad.face bob.db.replaymobile click tqdm
 # you would install the relevant bob.bio, bob.pad, and bob.db package
 
 # now, save the script provided below in a `convert_pad_database.py` file and run:
@@ -29,8 +29,10 @@ $ python convert_pad_database.py bob.pad.face.config.replay_mobile --output ~/te
 
 ```python
 import os
+import shutil
 
 import click
+import tqdm
 from bob.extension.scripts.click_helper import ConfigCommand
 from bob.extension.scripts.click_helper import ResourceOption
 from bob.extension.scripts.click_helper import log_parameters
@@ -49,7 +51,7 @@ from bob.extension.scripts.click_helper import verbosity_option
     default=["train", "dev", "eval"],
 )
 @verbosity_option(cls=ResourceOption)
-def convert_database(database, output, protocols, groups, verbose, **kwargs):
+def convert_database(database, output, protocols, groups, **kwargs):
     import csv
     import tarfile
     import logging
@@ -64,6 +66,7 @@ def convert_database(database, output, protocols, groups, verbose, **kwargs):
             click.echo(f"WARNING: failed to retrieve the list of protocols from the lower level db. The error was: {e}")
             protocols = [database.protocol]
 
+    shutil.rmtree(output, ignore_errors=True)
     os.makedirs(output, exist_ok=True)
 
     for proto in protocols:
@@ -74,16 +77,16 @@ def convert_database(database, output, protocols, groups, verbose, **kwargs):
             path = os.path.join(output, proto, grp + ".csv")
             os.makedirs(os.path.dirname(path), exist_ok=True)
             with open(path, "w", newline="") as f:
-                fieldnames = ["filename", "client_id", "attack_type"]
+                fieldnames = ["filename", "subject", "attack_type"]
                 writer = csv.DictWriter(f, fieldnames=fieldnames)
                 writer.writeheader()
-                for padfile in all_files:
+                for padfile in tqdm.tqdm(all_files):
                     writer.writerow(
                         {
                             "filename": padfile.make_path(
                                 "", database.original_extension
                             ),
-                            "client_id": padfile.client_id,
+                            "subject": padfile.client_id,
                             "attack_type": padfile.attack_type,
                         }
                     )