components/optimization_guide/proto/autofill_field_classification_model_metadata.proto

// Copyright 2023 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

edition = "2023";

package optimization_guide.proto;

option optimize_for = LITE_RUNTIME;
option java_package = "org.chromium.components.optimization_guide.proto";
option java_outer_classname = "AutofillFieldClassificationModelMetadata";

message AutofillFieldClassificationEncodingParameters {
  // Per-field features that can be included in the model input.
  enum AutofillFieldClassificationFeature {
    // Don't use.
    FEATURE_UNKNOWN = 0;

    // The label of the field.
    FEATURE_LABEL = 1;

    // The placeholder in the field.
    FEATURE_PLACEHOLDER = 2;

    // The autocomplete attribute of the field.
    FEATURE_AUTOCOMPLETE = 3;

    // The id attribute of the field.
    FEATURE_ID = 4;

    // The name attribute of the field (not FormFieldData::name() which falls
    // back to the id attribute if the name attribute does not exist).
    FEATURE_NAME = 5;

    // The form control type attribute of the field.
    FEATURE_TYPE = 6;
  }

  // Per-form features that can be included in the model input.
  enum AutofillFieldClassificationFormLevelFeature {
    // Don't use.
    FORM_FEATURE_UNKNOWN = 0;

    // The titles of <button> and <submit> elements in a form containing the
    // field.
    FEATURE_FORM_BUTTON_TITLES = 1;

    // The id attribute of a form containing the field.
    FEATURE_FORM_ID = 2;

    // The name attribute of a form containing the field.
    FEATURE_FORM_NAME = 3;

    // The URL path of the frame containing the field.
    FEATURE_FRAME_URL_PATH = 4;
  }

  // Maximum number of form fields for which the model can predict types.
  // When calling the executor with a larger form, predictions are only returned
  // for the first `maximum_number_of_fields` many fields.
  int32 maximum_number_of_fields = 1 [features.field_presence = IMPLICIT];

  // Max tokens per field per feature.
  int32 max_tokens_per_feature = 2 [features.field_presence = IMPLICIT];

  // The per-field features provided to the model in the order they are
  // presented to the model.
  repeated AutofillFieldClassificationFeature features = 3;

  // Pre-processing of feature strings:
  // labels, name, id attributes, etc. are preprocessed by the following steps:

  // Split a camelCase or UpperCamelCase string into separate words with
  // spaces in between.
  bool split_on_camel_case = 4 [features.field_presence = IMPLICIT];

  // If true, map the string to lowercase.
  bool lowercase = 5 [features.field_presence = IMPLICIT];

  // Replace all characters listed in `replace_chars_with_whitespace` with a
  // single space. If empty, this step is skipped.
  string replace_chars_with_whitespace = 6 [features.field_presence = IMPLICIT];

  // Replace all characters listed in `remove_chars` with the empty string.
  // If empty, this step is skipped.
  string remove_chars = 7 [features.field_presence = IMPLICIT];

  // The per-form features provided to the model in the order they are presented
  // to the model.
  repeated AutofillFieldClassificationFormLevelFeature form_features = 8;
}

message AutofillFieldClassificationPostprocessingParameters {
  // If set, types are only assigned to fields that have predictions
  // with at least the specified minimum confidence.
  float confidence_threshold_per_field = 1;

  // If set, the model returns predictions only if *all* fields have predictions
  // with at least the specified minimum confidence.
  float confidence_threshold_to_disable_all_predictions = 2;

  // If true, the model is not allowed to return the same prediction for
  // multiple fields in the form.
  bool disallow_same_type_predictions = 3;
}

// Metadata for OPTIMIZATION_TARGET_AUTOFILL_FIELD_CLASSIFICATION.
message AutofillFieldClassificationModelMetadata {
  // The dictionary used for vectorization of the input labels. The strings
  // are mapped to integers in the order that they are present in the repeated
  // field. Index 0 of `input_token` maps to integer 2, since 0 is reserved for
  // padding and 1 is reserved for unknown tokens.
  repeated string input_token = 1;

  // The mapping from the model's outputs to the (integer representation of) the
  // corresponding `autofill::FieldType`.
  repeated fixed32 output_type = 2;

  // Deprecated in M132.
  reserved 3;

  // The version of the input, based on which the relevant model
  // version will be used from server.
  int64 input_version = 4;

  // Encoding parameters, used for input versions >= 3.
  AutofillFieldClassificationEncodingParameters encoding_parameters = 5;

  // Output processing parameters.
  AutofillFieldClassificationPostprocessingParameters
      postprocessing_parameters = 6;
}