python3でyamlファイルを読み込む

yamlファイルは、設定ファイルに使われることが多いようです。

config.yaml

hyperparameterA:
  name: 'ModelName_AAA'
  param1: 1
  param2: 2
hyperparameterB:
  name: 'ModelName_BBB'
  param1: 3
  param2: 4

test.py

import yaml

with open('config.yml', 'r') as yml:
    config = yaml.safe_load(yml)

print('--- A ---')
print('name: ', config['hyperparameterA']['name'])
print('param1: ', config['hyperparameterA']['param1'])
print('param2: ', config['hyperparameterA']['param2'])
print('--- B ---')
print('name: ', config['hyperparameterB']['name'])
print('param1: ', config['hyperparameterB']['param1'])
print('param2: ', config['hyperparameterB']['param2'])

実行

$ python3 test.py
--- A ---
name:  ModelName_AAA
param1:  1
param2:  2
--- B ---
name:  ModelName_BBB
param1:  3
param2:  4

YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.

yaml.load(yml)を使うとwarningが出る。

test.py

import yaml

with open('config.yml', 'r') as yml:
    config = yaml.load(yml)

print('--- A ---')
print('name: ', config['hyperparameterA']['name'])
print('param1: ', config['hyperparameterA']['param1'])
print('param2: ', config['hyperparameterA']['param2'])
print('--- B ---')
print('name: ', config['hyperparameterB']['name'])
print('param1: ', config['hyperparameterB']['param1'])
print('param2: ', config['hyperparameterB']['param2'])

$ python3 test.py
YAMLLoadWarning: calling yaml.load() without Loader=... is deprecated, as the default Loader is unsafe. Please read https://msg.pyyaml.org/load for full details.

yaml.safe_load(yml)もしくは、yaml.load(yml, Loader=yaml.SafeLoader)を使うとwarningが出なくなる。

参考

Pythonで設定ファイルを読み込んでみる（YAMLファイル）
【後編】YAMLを使ってPyTorchのOptimizerのパラメータ管理をすごく簡単にする方法
pythonでyamlの扱い