feat: Added config options for derived fields and signature detection. Some annotation changes changes to mark fields explicitly output only (no changes in semantics just readability)

Google APIs · copybara-github · commit 6429996ce986 · 2025-08-15T15:14:10.000-07:00
docs: A comment for field `document_schema` in message `.google.cloud.documentai.v1.ProcessorVersion` is changed to describe as output only.
docs: A comment for field `create_time` in message `.google.cloud.documentai.v1.ProcessorVersion`  is changed to describe as output only.
docs: A comment for field `latest_evaluation` in message `.google.cloud.documentai.v1.ProcessorVersion`  is changed to describe as output only.
docs: A comment for field `kms_key_name` in message `.google.cloud.documentai.v1.ProcessorVersion`  is changed to describe as output only.
docs: A comment for field `kms_key_version_name` in message `.google.cloud.documentai.v1.ProcessorVersion`  is changed to describe as output only.
docs: A comment for field `deprecation_info` in message `.google.cloud.documentai.v1.ProcessorVersion`  is changed to describe as output only.
docs: A comment for field `create_time` in message `.google.cloud.documentai.v1.Processor`  is changed to describe as output only.

PiperOrigin-RevId: 795630471
diff --git a/google/cloud/documentai/v1/document.proto b/google/cloud/documentai/v1/document.proto
@@ -585,6 +585,8 @@ message Document {
 
         // Float value.
         float float_value = 8;
+
+        bool signature_value = 10;
       }
 
       // Optional. An optional field to store a normalized string.
@@ -601,6 +603,20 @@ message Document {
       string text = 1 [(google.api.field_behavior) = OPTIONAL];
     }
 
+    // Specifies how the entity's value is obtained.
+    enum Method {
+      // When the method is not specified, it should be treated as `EXTRACT`.
+      METHOD_UNSPECIFIED = 0;
+
+      // The entity's value is directly extracted as-is from the document
+      // text.
+      EXTRACT = 1;
+
+      // The entity's value is derived through inference and is not
+      // necessarily an exact text extraction from the document.
+      DERIVE = 2;
+    }
+
     // Optional. Provenance of the entity.
     // Text anchor indexing into the
     // [Document.text][google.cloud.documentai.v1.Document.text].
@@ -643,6 +659,9 @@ message Document {
     // Optional. Whether the entity will be redacted for de-identification
     // purposes.
     bool redacted = 12 [(google.api.field_behavior) = OPTIONAL];
+
+    // Optional. Specifies how the entity's value is obtained.
+    Method method = 15 [(google.api.field_behavior) = OPTIONAL];
   }
 
   // Relationship between
diff --git a/google/cloud/documentai/v1/document_schema.proto b/google/cloud/documentai/v1/document_schema.proto
@@ -66,6 +66,20 @@ message DocumentSchema {
         REQUIRED_MULTIPLE = 4;
       }
 
+      // Specifies how the entity's value is obtained from the document.
+      enum Method {
+        // Unspecified method. It defaults to `EXTRACT`.
+        METHOD_UNSPECIFIED = 0;
+
+        // The entity's value is directly extracted as-is from the document
+        // text.
+        EXTRACT = 1;
+
+        // The entity's value is derived through inference and is not
+        // necessarily an exact text extraction from the document.
+        DERIVE = 2;
+      }
+
       // The name of the property.  Follows the same guidelines as the
       // EntityType name.
       string name = 1;
@@ -80,6 +94,9 @@ message DocumentSchema {
       // Occurrence type limits the number of instances an entity type appears
       // in the document.
       OccurrenceType occurrence_type = 3;
+
+      // Specifies how the entity's value is obtained.
+      Method method = 8;
     }
 
     oneof value_source {
diff --git a/google/cloud/documentai/v1/documentai_v1.yaml b/google/cloud/documentai/v1/documentai_v1.yaml
@@ -94,6 +94,7 @@ publishing:
   library_settings:
   - version: google.cloud.documentai.v1
     dotnet_settings:
+      common: {}
       ignored_resources:
       - documentai.googleapis.com/Location
   proto_reference_documentation_uri: https://cloud.google.com/document-ai/docs/reference/rpc
diff --git a/google/cloud/documentai/v1/evaluation.proto b/google/cloud/documentai/v1/evaluation.proto
@@ -16,6 +16,7 @@ syntax = "proto3";
 
 package google.cloud.documentai.v1;
 
+import "google/api/field_behavior.proto";
 import "google/api/resource.proto";
 import "google/protobuf/timestamp.proto";
 
diff --git a/google/cloud/documentai/v1/processor.proto b/google/cloud/documentai/v1/processor.proto
@@ -148,29 +148,35 @@ message ProcessorVersion {
   // The display name of the processor version.
   string display_name = 2;
 
-  // The schema of the processor version. Describes the output.
-  DocumentSchema document_schema = 12;
+  // Output only. The schema of the processor version. Describes the output.
+  DocumentSchema document_schema = 12
+      [(google.api.field_behavior) = OUTPUT_ONLY];
 
   // Output only. The state of the processor version.
   State state = 6 [(google.api.field_behavior) = OUTPUT_ONLY];
 
-  // The time the processor version was created.
-  google.protobuf.Timestamp create_time = 7;
+  // Output only. The time the processor version was created.
+  google.protobuf.Timestamp create_time = 7
+      [(google.api.field_behavior) = OUTPUT_ONLY];
 
-  // The most recently invoked evaluation for the processor version.
-  EvaluationReference latest_evaluation = 8;
+  // Output only. The most recently invoked evaluation for the processor
+  // version.
+  EvaluationReference latest_evaluation = 8
+      [(google.api.field_behavior) = OUTPUT_ONLY];
 
-  // The KMS key name used for encryption.
-  string kms_key_name = 9;
+  // Output only. The KMS key name used for encryption.
+  string kms_key_name = 9 [(google.api.field_behavior) = OUTPUT_ONLY];
 
-  // The KMS key version with which data is encrypted.
-  string kms_key_version_name = 10;
+  // Output only. The KMS key version with which data is encrypted.
+  string kms_key_version_name = 10 [(google.api.field_behavior) = OUTPUT_ONLY];
 
   // Output only. Denotes that this `ProcessorVersion` is managed by Google.
   bool google_managed = 11 [(google.api.field_behavior) = OUTPUT_ONLY];
 
-  // If set, information about the eventual deprecation of this version.
-  DeprecationInfo deprecation_info = 13;
+  // Output only. If set, information about the eventual deprecation of this
+  // version.
+  DeprecationInfo deprecation_info = 13
+      [(google.api.field_behavior) = OUTPUT_ONLY];
 
   // Output only. The model type of this processor version.
   ModelType model_type = 15 [(google.api.field_behavior) = OUTPUT_ONLY];
@@ -275,8 +281,9 @@ message Processor {
     (google.api.field_behavior) = IMMUTABLE
   ];
 
-  // The time the processor was created.
-  google.protobuf.Timestamp create_time = 7;
+  // Output only. The time the processor was created.
+  google.protobuf.Timestamp create_time = 7
+      [(google.api.field_behavior) = OUTPUT_ONLY];
 
   // The [KMS key](https://cloud.google.com/security-key-management) used for
   // encryption and decryption in CMEK scenarios.
