Mercurial > repos > goeckslab > ludwig_config_generator

annotate ludwig_autogenconfig.py @ 0:183adfc24076 draft default tip

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
author goeckslab
date Tue, 07 Jan 2025 22:46:36 +0000
parents
children
Ignore whitespace changes - Everywhere: Within whitespace: At end of lines:
rev   line source
0
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
1 import argparse
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
2 import logging
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
3
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
4 from ludwig import automl
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
5 from ludwig.utils import defaults
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
6
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
7 from pandas import read_csv
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
8
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
9 logging.basicConfig(level=logging.DEBUG)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
10 LOG = logging.getLogger(__name__)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
11
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
12
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
13 def main():
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
14 parser = argparse.ArgumentParser(
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
15 description='Render a Ludwig config')
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
16 parser.add_argument(
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
17 '--dataset',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
18 type=str,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
19 help='Path to the dataset file',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
20 required=True)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
21 parser.add_argument(
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
22 '--output_feature',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
23 type=int,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
24 help='Name for the output feature',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
25 required=True)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
26 parser.add_argument(
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
27 '--output',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
28 type=str,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
29 help='Path for the output file',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
30 required=True)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
31 parser.add_argument(
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
32 '--renderconfig',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
33 action='store_true',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
34 help='Render the config',
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
35 required=False,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
36 default=False)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
37 args = parser.parse_args()
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
38
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
39 # get the output feature name
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
40 df = read_csv(args.dataset, nrows=2, sep=None, engine='python')
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
41 names = df.columns.tolist()
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
42 target = names[args.output_feature-1]
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
43
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
44 args_init = ["--dataset", args.dataset,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
45 "--target", target,
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
46 "--output", args.output]
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
47 automl.cli_init_config(args_init)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
48
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
49 if args.renderconfig:
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
50 args_render = ["--config", args.output, "--output", args.output]
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
51 defaults.cli_render_config(args_render)
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
52
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
53
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
54 if __name__ == "__main__":
183adfc24076 planemo upload for repository https://github.com/goeckslab/Galaxy-Ludwig.git commit bdea9430787658783a51cc6c2ae951a01e455bb4
goeckslab
parents:
diff changeset
55 main()