feat: [dlp] Supporting Conversation Content Type

packages/google-privacy-dlp/README.md

@@ -13,7 +13,9 @@ are addressed with the highest priority
 
 Sensitive Data Protection (DLP) client for Node.js
 
-[//]: # "partials.introduction"
+The [Data Loss Prevention API](https://cloud.google.com/dlp/docs/) provides programmatic access to a
+powerful detection engine for personally identifiable information and other privacy-sensitive
+data in unstructured data streams.
 
 A comprehensive list of changes in each version may be found in
 [the CHANGELOG][homepage_changelog].

packages/google-privacy-dlp/protos/google/privacy/dlp/v2/dlp.proto

@@ -1383,6 +1383,11 @@ message ContentItem {
 
     // Content data to inspect or redact. Replaces `type` and `data`.
     ByteContentItem byte_item = 5;
+
+    // Represents a conversation (either complete or a slice).
+    // It is assumed that all included messages are contiguous and ordered in
+    // chronological order.
+    Conversation conversation = 7;
   }
 
   // User provided metadata for the content.
@@ -1395,6 +1400,48 @@ message ContentMetadata {
   repeated KeyValueMetadataProperty properties = 2;
 }
 
+// Complete conversation or slice of a conversation.
+// It is assumed that all included messages are contiguous and ordered in
+// chronological order.
+message Conversation {
+  // Messages exchanged within this conversation.
+  // The maximum number of messages allowed is 50k.
+  // The order of the messages is assumed to be chronological and will be used
+  // to index findings in the response.
+  repeated ConversationMessage messages = 1;
+}
+
+// Single message in a conversation.
+message ConversationMessage {
+  // The type of message.
+  // New values may be added in the future.
+  enum MessageType {
+    // Unused.
+    MESSAGE_TYPE_UNSPECIFIED = 0;
+
+    // Message contains content to be inspected.
+    CONTENT = 1;
+
+    // Message contains context only and will not have findings reported from
+    // it during inspection or redacted from it during de-identification.
+    CONTEXT = 2;
+  }
+
+  // The contents of this message.
+  string content = 1;
+
+  // The type of message.
+  MessageType message_type = 2;
+
+  // Optional. The identifier of the participant,
+  // for example, 'test-user' or 'gemini'.
+  // The participant ID can contain lowercase letters, numbers, and hyphens;
+  // that is, it must match the regular expression:
+  // `^[a-z]([a-z0-9-]{0,61}[a-z0-9])?$`.
+  // The maximum length is 63 characters.
+  string participant_id = 3;
+}
+
 // Structured content to inspect. Up to 50,000 `Value`s per request allowed. See
 // https://cloud.google.com/sensitive-data-protection/docs/inspecting-structured-text#inspecting_a_table
 // to learn more.
@@ -1558,6 +1605,9 @@ message ContentLocation {
 
     // Location within the metadata for inspected content.
     MetadataLocation metadata_location = 8;
+
+    // Location within a conversation.
+    ConversationLocation conversation_location = 10;
   }
 
   // Finding container modification timestamp, if applicable. For Cloud Storage,
@@ -1571,6 +1621,24 @@ message ContentLocation {
   string container_version = 7;
 }
 
+// Location within a conversation.
+message ConversationLocation {
+  // If set, indicates that the finding applies to all messages in the
+  // conversation.
+  message AllMessages {}
+
+  // The location of the finding within a conversation.
+  oneof location {
+    // Matches an index of a message in the conversation provided in the
+    // request.
+    int32 message_index = 1;
+
+    // If set, indicates that the finding applies to all messages in the
+    // conversation.
+    AllMessages all_messages = 2;
+  }
+}
+
 // Metadata Location
 message MetadataLocation {
   // Type of metadata containing the finding.
@@ -4061,21 +4129,20 @@ enum TransformationResultStatusType {
   // Unused.
   STATE_TYPE_UNSPECIFIED = 0;
 
-  // This will be set when a finding could not be transformed (i.e. outside user
+  // This is set when a finding cannot be transformed (i.e. outside user
   // set bucket range).
   INVALID_TRANSFORM = 1;
 
-  // This will be set when a BigQuery transformation was successful but could
-  // not be stored back in BigQuery because the transformed row exceeds
-  // BigQuery's max row size.
+  // This is set when a transformation is successful but cannot be stored in
+  // BigQuery because the transformed row exceeds BigQuery's max row size.
   BIGQUERY_MAX_ROW_SIZE_EXCEEDED = 2;
 
-  // This will be set when there is a finding in the custom metadata of a file,
+  // This is set when there is a finding in the custom metadata of a file,
   // but at the write time of the transformed file, this key / value pair is
   // unretrievable.
   METADATA_UNRETRIEVABLE = 3;
 
-  // This will be set when the transformation and storing of it is successful.
+  // This is set when the transformation and its storage are successful.
   SUCCESS = 4;
 }
 
@@ -5169,7 +5236,7 @@ message DataProfileAction {
     //    visible to queries by the time your topic receives the Pub/Sub
     //    notification.
     //  * The best practice is to use the same table for an entire organization
-    //    so that you can take advantage of the [provided Looker
+    //    so that you can take advantage of the [provided Data Studio
     //    reports](https://cloud.google.com/sensitive-data-protection/docs/analyze-data-profiles#use_a_premade_report).
     //    If you use VPC Service Controls to define security perimeters, then
     //    you must use a separate table for each boundary.