a2rl.save_metadata#

a2rl.save_metadata(metadata, yaml_file, compact=False)[source]#

Save an in-memory metadata object into a YAML file.

Parameters:

metadata (Metadata | MetadataDict) – Metadata object.
yaml_file (str | Path) – Path to the output YAML file.
compact (bool) – When set to True, do not output None entries.

Return type:

None

See also

Metadata, read_metadata

Examples

Save an in-memory metadata object.

>>> import a2rl as wi
>>> m = wi.Metadata(
...     states=["s", "t"],
...     actions=["a"],
...     rewards=["r"],
...     tags={"k": "v"}
... )
>>> wi.save_metadata(m, "/tmp/metadata.yaml")

>>> with open("/tmp/metadata.yaml") as f:
...     print(''.join(f.readlines()))
states:
- s
- t

actions:
- a

rewards:
- r

forced_categories: null

frequency: null

tags:
  k: v

Save metadata in compact mode to exclude null items in the YAML output.

>>> wi.save_metadata(m, "/tmp/metadata.yaml", compact=True)
>>> with open("/tmp/metadata.yaml") as f:
...     print(''.join(f.readlines()))
states:
- s
- t

actions:
- a

rewards:
- r

tags:
  k: v

Save a dictionary. Be aware that the dictionary must specifies all the Metadata fields, including the default ones. If you prefer the flexibility to not re-declare the default fields, please use Metadata instead.

>>> d: wi.MetadataDict = {
...     "states": ["s", "t"],
...     "actions": ["a"],
...     "rewards": ["r"],
...     "forced_categories": ["a"],
...     "frequency": None,
...     "tags": {},
... }
>>> wi.save_metadata(d, "/tmp/metadata.yaml", compact=True)

>>> with open("/tmp/metadata.yaml") as f:
...     print(''.join(f.readlines()))
states:
- s
- t

actions:
- a

rewards:
- r

forced_categories:
- a

tags: {}