Editing Mapper

Manual Edit allows you to manually make changes to the Mapper’s logic.

Use the mapper payload property in Run Workshop with Mapper to provide the mapping logic edits to Lume.

Below is an example:

[
  {
      "targetField": "internal_id", // target field transformation is being applied to
      "transformation": {
          "extract": "id", // source property to extract one-to-one mapping to
      }
  },

  // ... more edits
]

Learn more about manual edits here.

Prompt Edit allows you to edit the mapping logic by communicating to the AI with a natural language prompt, similar to as if you are communicating with ChatGPT.

Use the target_fields_to_prompt payload property in Run Workshop with Prompt to provide the prompts to Lume.

Each prompt corresponds to a target property, as seen in the example below.

{
    "target_prop_1": "For this field, include this source values concatenated together with commas",
    "target_prop_2": "For this field, always map to null"
}

Target Schema Edit allows you to edit the target model directly, and the AI will regenerate the logic based off of changes.

Make edits to the target schema, then use the target_schema payload property in Run Workshop with Target Schema to pass in the edited schema.

Learn how to edit Target Schemas here.

Sample Suggestion allows you to edit mapping logic by communicating to the AI what the correct mapped values are, and Lume will regenerate the mapping logic based off of your suggestion.

Use the sample payload property in Run Workshop with Sample to provide the correctly mapped record to Lume.

For example, the mapped data below actually mapped key1 incorrectly to incorrect_X.

{
    "key1": "incorrect_X",
    "key2": "correct_value"
}

After view, based on the source data of this job, you discern it should instead have mapped to correct_Y. Thus, pass in sample with the correct values for key1 and Lume will use this correction to regenerate the Mapper logic. Make sure to indicate the index of the source record being mapped from.

{
  "sample": {
    "index": 0,
    "mapped_record": {
      "key1": "correct__Y",
      "key2": "correct_value"
    },
  }
}