ModifyAutoMLConfigDict

class hana_ai.tools.hana_ml_tools.config_dict_validator_tools.ModifyAutoMLConfigDict(connection_context: ConnectionContext, return_direct: bool = False)

Verify (and, if needed, amend) an AutoML config_dict via _SYS_AFL.PAL_AUTOML_CONFIG.

Preferred usage — explicit config_dict. Build the full config_dict yourself from the catalogue returned by GetPALPipelineInfo, pass it as config_dict=<full JSON object> with verify=True and no config_add / config_remove / config_replace / config_modify modifiers. The returned config_dict field is the PAL-normalized dict — hand that verbatim to the fit tool. This keeps the search space explicit and auditable, and any invalid operator / parameter is surfaced up-front with {"error": "invalid_config_dict", "detail": <pal message>} so the agent can fix it and re-verify.

Delta usage — modifiers on top of a template. The optional CONFIG_REMOVE / CONFIG_ADD / CONFIG_REPLACE / CONFIG_MODIFY rows exist for tweaking a starting template. PAL applies them in the order documented at the SAP Help portal: CONFIG_DICT → CONFIG_REMOVE → CONFIG_ADD → CONFIG_REPLACE → CONFIG_MODIFY. Because the layering is less predictable than an explicit dict, reach for these modifiers only when you really want a delta against 'default' / 'light' / 'empty'.

Returns:
str

A JSON string with:

  • pipeline_type : the normalised pipeline type.

  • config_dict : the resolved config_dict as a JSON object.

  • operators : per-operator INFO rows.

  • error / detail (optional): populated when PAL rejects the modification. error == 'invalid_config_dict' when verify=True and PAL raises.

Note

args_schema is used to define the schema of the inputs as follows:

Field

Description

pipeline_type

AutoML pipeline type: 'classifier', 'regressor' or 'timeseries'.

config_dict

Starting template or custom JSON. Defaults to 'default'.

config_remove

Estimator configs to remove (dict / JSON string).

config_add

Estimator configs to add.

config_replace

Estimator configs to replace.

config_modify

Estimator configs to modify in place.

verify

When True (default) PAL raises on invalid entries (VERIFY_CONFIG=1).
name: str

Name of the tool.

description: str

Description of the tool.

connection_context: ConnectionContext

Connection context to the HANA database.