third_party/lens_server_proto/lens_overlay_text.proto

// Copyright 2024 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

syntax = 'proto3';

option optimize_for = LITE_RUNTIME;

package lens;

import "lens_overlay_geometry.proto";

// The text reading order.
enum WritingDirection {
  DEFAULT_WRITING_DIRECTION_LEFT_TO_RIGHT = 0;
  WRITING_DIRECTION_RIGHT_TO_LEFT = 1;
  WRITING_DIRECTION_TOP_TO_BOTTOM = 2;
}

message Text {
  // Optional. Information describing the text.
  TextLayout text_layout = 1;

  // Optional. Dominant content language of the text. Language
  // code is CLDR/BCP-47.
  string content_language = 2;

  reserved 3;
}

// Nested text structure.
message TextLayout {
  message Word {
    // Required. Unique id within TextLayout.
    TextEntityIdentifier id = 1;

    // Optional. The text in a plain text.
    string plain_text = 2;

    // Optional. The text separator that should be appended after this word when
    // it is concatenated with the subsequent word in the same or next
    // line/paragraph into a single-line string.
    string text_separator = 3;

    // Optional. The geometry of the word.
    Geometry geometry = 4;

    enum Type {
      // Printed text.
      TEXT = 0;
      // Formula type, including mathematical or chemical formulas.
      FORMULA = 1;
    }

    // Optional. The type of this word.
    Type type = 5;

    message FormulaMetadata {
      // Optional. LaTeX representation of a formula. Can be the same as
      // `plain_text`. Example: "\frac{2}{x}=y". The plain text
      // representation of this is available in Word.plain_text.
      string latex = 1;
    }

    // Optional. Metadata for formulas. This is populated for entities of
    // `type=FORMULA`.
    FormulaMetadata formula_metadata = 6;
  }

  message Line {
    // Optional. List of words in natural reading order.
    repeated Word words = 1;

    // Optional. The geometry of the line.
    Geometry geometry = 2;
  }

  message Paragraph {
    // Required. Unique id within TextLayout.
    TextEntityIdentifier id = 1;

    // Optional. List of lines in natural reading order (see also
    // `writing_direction`).
    repeated Line lines = 2;

    // Optional. Geometry of the paragraph.
    Geometry geometry = 3;

    // Optional. The text writing direction (aka reading order).
    WritingDirection writing_direction = 4;

    // Optional. BCP-47 language code of the dominant language in this
    // paragraph.
    string content_language = 5;
  }

  // Optional. List of paragraphs in natural reading order.
  repeated Paragraph paragraphs = 1;
}

message TextEntityIdentifier {
  // Required. Unique entity id used to reference (and match) text entities and
  // ranges.
  int64 id = 1;
}