H2h2o
Extractors

Creates extractor.

post/extractors

Request

Request body

namestring required

Human-readable name

descriptionstring required

What this extractor does

llmstring required

Identifier or version of the language model the extractor uses

schemastring required

JSONSchema that the extractor outputs, serialized as a string. Must be empty or a JSON object; a string, array or other non-object value is rejected with 400. When schema is not empty, the output is always returned as a JSON object. When schema is empty the extractor is unconstrained and its results are returned as free-form text rather than an object.

promptstring

Custom prompt text for the extractor. If empty, default prompt will be used. If combined with schema, prompt can customize the output, but cannot change its format as it will always be JSON object.

pre_prompt_summarystring

(Optional) Pre-prompt summary text for the extractor

keep_intermediate_resultsboolean

(Optional) Flag indicating whether to keep intermediate results during extraction

system_promptstring

(Optional) System prompt text for the extractor

max_num_chunksinteger

(Optional) Maximum number of chunks to process

visionstring

(Optional) Vision mode setting

vision_llmstring

(Optional) Vision LLM model identifier

image_batch_image_promptstring

(Optional) Image batch processing prompt

image_batch_final_promptstring

(Optional) Final prompt for image batch processing

Example request

{
  "guardrails_settings": {
    "disallowed_regex_patterns": [
      "secret_disallowed_word",
      "(?!0{3})(?!6{3})[0-8]\\d{2}-(?!0{2})\\d{2}-(?!0{4})\\d{4}"
    ],
    "presidio_labels_to_flag": [
      "IBAN_CODE",
      "US_SSN"
    ],
    "pii_labels_to_flag": [
      "IBAN",
      "SSN"
    ],
    "prompt_guard_labels_to_flag": [
      "JAILBREAK"
    ],
    "guardrails_labels_to_flag": [
      "Violent Crimes",
      "Non-Violent Crimes",
      "Intellectual Property",
      "Code Interpreter Abuse"
    ],
    "guardrails_llm": "h2oai/h2o-danube3-4b-chat",
    "guardrails_safe_category": "Safe",
    "column_redaction_custom_entities_to_flag": [
      "Mother's Maiden Name"
    ],
    "column_redaction_custom_entities": {
      "Mother's Maiden Name": "Mother's maiden name"
    },
    "custom_pii_entities_to_flag": [
      "Mother's Maiden Name"
    ],
    "custom_pii_entities": {
      "Mother's Maiden Name": "Mother's maiden name"
    }
  }
}

Response

OK

namestring required

Human-readable name

descriptionstring required

What this extractor does

llmstring

(Optional) Identifier or version of the language model the extractor uses

schemastring

(Optional) JSONSchema (or other spec) that the extractor outputs

promptstring

(Optional) Prompt text for the extractor

pre_prompt_summarystring

(Optional) Pre-prompt summary text for the extractor

keep_intermediate_resultsboolean

(Optional) Flag indicating whether to keep intermediate results during extraction

system_promptstring

(Optional) System prompt text for the extractor

max_num_chunksinteger

(Optional) Maximum number of chunks to process

visionstring

(Optional) Vision mode setting

vision_llmstring

(Optional) Vision LLM model identifier

image_batch_image_promptstring

(Optional) Image batch processing prompt

image_batch_final_promptstring

(Optional) Final prompt for image batch processing

idstring required

Unique identifier of the extractor

is_publicboolean required

Flag indicating if the extractor is public

created_atstring date-time required

When the extractor definition was created

Example response

{
  "guardrails_settings": {
    "disallowed_regex_patterns": [
      "secret_disallowed_word",
      "(?!0{3})(?!6{3})[0-8]\\d{2}-(?!0{2})\\d{2}-(?!0{4})\\d{4}"
    ],
    "presidio_labels_to_flag": [
      "IBAN_CODE",
      "US_SSN"
    ],
    "pii_labels_to_flag": [
      "IBAN",
      "SSN"
    ],
    "prompt_guard_labels_to_flag": [
      "JAILBREAK"
    ],
    "guardrails_labels_to_flag": [
      "Violent Crimes",
      "Non-Violent Crimes",
      "Intellectual Property",
      "Code Interpreter Abuse"
    ],
    "guardrails_llm": "h2oai/h2o-danube3-4b-chat",
    "guardrails_safe_category": "Safe",
    "column_redaction_custom_entities_to_flag": [
      "Mother's Maiden Name"
    ],
    "column_redaction_custom_entities": {
      "Mother's Maiden Name": "Mother's maiden name"
    },
    "custom_pii_entities_to_flag": [
      "Mother's Maiden Name"
    ],
    "custom_pii_entities": {
      "Mother's Maiden Name": "Mother's maiden name"
    }
  }
}

Changes

No recorded changes to this operation across all 1 revision of this API.