TextAnnotation.proto

syntax = "proto2";

// Protocol Buffer definition for TextAnnotation structure
// This file is used to generate Java files to write/read TA in
// Protocol Buffer format.
// Note: If you make changes to this file, make sure to generate Java
// files and update the serialization code accordingly.

package textannotation;

option java_outer_classname = "TextAnnotationImpl";
option java_package = "edu.illinois.cs.cogcomp.core.utilities.protobuf";
option java_multiple_files = false;

// Root TextAnnotation message
message TextAnnotationProto {

  // Corpus Identifier
  required string corpusId = 1;

  // Document Identifier
  required string id = 2;

  // Raw text corresponding to this annotation
  required string text = 3;

  // List of tokens
  repeated string tokens = 4;

  // Token Offset information
  repeated TokenOffsetsProto tokenOffsets = 5;

  // List of Sentences
  required SentenceProto sentences = 6;

  // List of Views
  repeated ViewProto views = 7;

  // Attribute information
  map<string, string> properties = 8;
}

// Token Offset message structure
message TokenOffsetsProto {

  // Surface form of the token
  required string form = 1;

  // Offset of the start character in source text
  required int32 startCharOffset = 2;

  // Offset of the last character in text + 1
  // Note: End offset is not part of the current token. [startCharOffset, endCharOffset)
  required int32 endCharOffset = 3;
}

// Sentence message structure
message SentenceProto {

  // Class name for the generator
  required string generator = 1;

  // Score of the sentence
  optional double score = 2 [ default = 0 ];

  // List of sentence end positions
  repeated int32 sentenceEndPositions = 3;
}

// View message structure
message ViewProto {

  // Name of the view
  required string viewName = 1;

  // View Data information
  repeated ViewDataProto viewData = 2;
}

// ViewData message structure
message ViewDataProto {

  // Type of the view. Corresponds to class name in core-utilities
  required string viewType = 1;

  // Name of the view
  required string viewName = 2;

  // Class name of the generator
  required string generator = 3;

  // Score of the View
  optional double score = 4 [ default = 0 ];

  // List of constituents in the view
  // Note: Order of constituents stored is important and should be preserved
  repeated ConstituentProto constituents = 5;

  // List of relations in the view
  // Note: Order of constituents stored is important and should be preserved
  repeated RelationProto relations = 6;
}

// Constituent message structure
message ConstituentProto {

  // Label of the constituent
  required string label = 1;

  // Score of the constituent
  optional double score = 2 [ default = 0 ];

  // Start index of the span
  // Span indices correspond to indices in the tokens list.
  required int32 start = 3;

  // End index of the span + 1
  // Span indices correspond to indices in the tokens list.
  // Usage: a span corresponds to [start, end) - end is not inclusive.
  required int32 end = 4;

  // Attribute information
  map<string, string> properties = 5;

  // Map containing scores to each label
  map<string, double> labelScoreMap = 6;
}

// Relation message structure
message RelationProto {

  // Name of the relation
  required string relationName = 1;

  // Score of the relation
  optional double score = 2 [ default = 0 ];

  // Identifier for the source constituent
  required int32 srcConstituent = 3;

  // Identifier for the target constituent
  required int32 targetConstituent = 4;

  // Attribute information
  map<string, string> properties = 5;

  // Map containing scores to each label
  map<string, double> labelScoreMap = 6;
}