Synchronize new proto/yaml changes.

PiperOrigin-RevId: 197061318

google/privacy/dlp/dlp_v2.yaml

@@ -11,28 +11,9 @@ documentation:
     Provides methods for detection, risk analysis, and de-identification of
     privacy-sensitive fragments in text, images, and Google Cloud Platform
     storage repositories.
-  rules:
-  - selector: google.longrunning.Operations.CancelOperation
-    description: |-
-      Cancels an operation. Use the
-      [`inspect.operations.get`][/dlp/docs/reference/rest/v2beta1/inspect.operations/get]
-      to check whether the cancellation succeeded or the operation completed
-      despite cancellation.
-  - selector: google.longrunning.Operations.DeleteOperation
-    description: This method is not supported and the server returns `UNIMPLEMENTED`.
-  - selector: google.longrunning.Operations.ListOperations
-    description: Fetches the list of long running operations.
 
 backend:
   rules:
-  - selector: google.longrunning.Operations.ListOperations
-    deadline: 300.0
-  - selector: google.longrunning.Operations.GetOperation
-    deadline: 300.0
-  - selector: google.longrunning.Operations.DeleteOperation
-    deadline: 300.0
-  - selector: google.longrunning.Operations.CancelOperation
-    deadline: 300.0
   - selector: google.privacy.dlp.v2.DlpService.InspectContent
     deadline: 300.0
   - selector: google.privacy.dlp.v2.DlpService.RedactImage
@@ -84,40 +65,6 @@ backend:
   - selector: google.privacy.dlp.v2.DlpService.CancelDlpJob
     deadline: 300.0
 
-http:
-  rules:
-  - selector: google.longrunning.Operations.ListOperations
-    get: '/v1/{name=inspect/operations}'
-    additional_bindings:
-    - get: '/v2beta1/{name=inspect/operations}'
-
-    - get: '/v2beta1/{name=riskAnalysis/operations}'
-
-  - selector: google.longrunning.Operations.GetOperation
-    get: '/v1/{name=inspect/operations/*}'
-    additional_bindings:
-    - get: '/v2beta1/{name=inspect/operations/*}'
-
-    - get: '/v2beta1/{name=riskAnalysis/operations/*}'
-
-  - selector: google.longrunning.Operations.DeleteOperation
-    delete: '/v1/{name=inspect/operations/*}'
-    additional_bindings:
-    - delete: '/v2beta1/{name=inspect/operations/*}'
-
-    - delete: '/v2beta1/{name=riskAnalysis/operations/*}'
-
-  - selector: google.longrunning.Operations.CancelOperation
-    post: '/v1/{name=inspect/operations/*}:cancel'
-    body: '*'
-    additional_bindings:
-    - post: '/v2beta1/{name=inspect/operations/*}:cancel'
-      body: '*'
-
-    - post: '/v2beta1/{name=riskAnalysis/operations/*}:cancel'
-      body: '*'
-
-
 authentication:
   rules:
   - selector: '*'

google/privacy/dlp/v2/dlp.proto

@@ -894,8 +894,8 @@ message PrivacyMetric {
   // using publicly available data (like the US Census), or using a custom
   // statistical model (indicated as one or several BigQuery tables), or by
   // extrapolating from the distribution of values in the input dataset.
+  // A column with a semantic tag attached.
   message KMapEstimationConfig {
-    // A column with a semantic tag attached.
     message TaggedField {
       // Identifies the column. [required]
       FieldId field = 1;
@@ -1493,8 +1493,7 @@ message CryptoReplaceFfxFpeConfig {
   // such that:
   //
   // - a 64 bit integer is encoded followed by a single byte of value 1
-  // - a string is encoded in UTF-8 format followed by a single byte of value
-  //  å 2
+  // - a string is encoded in UTF-8 format followed by a single byte of value 2
   FieldId context = 2;
 
   oneof alphabet {

google/privacy/dlp/v2/storage.proto

@@ -223,6 +223,20 @@ message CloudStorageOptions {
     string url = 1;
   }
 
+  // How to sample bytes if not all bytes are scanned. Meaningful only when used
+  // in conjunction with bytes_limit_per_file. If not specified, scanning would
+  // start from the top.
+  enum SampleMethod {
+    SAMPLE_METHOD_UNSPECIFIED = 0;
+
+    // Scan from the top (default).
+    TOP = 1;
+
+    // For each file larger than bytes_limit_per_file, randomly pick the offset
+    // to start scanning. The scanned bytes are contiguous.
+    RANDOM_START = 2;
+  }
+
   FileSet file_set = 1;
 
   // Max number of bytes to scan from a file. If a scanned file's size is bigger
@@ -233,9 +247,16 @@ message CloudStorageOptions {
   // If empty, all files are scanned and available data format processors
   // are applied.
   repeated FileType file_types = 5;
+
+  SampleMethod sample_method = 6;
+
+  // Limits the number of files to scan to this percentage of the input FileSet.
+  // Number of files scanned is rounded down. Must be between 0 and 100,
+  // inclusively. Both 0 and 100 means no limit. Defaults to 0.
+  int32 files_limit_percent = 7;
 }
 
-// Message representing a path in Cloud Storage.
+// Message representing a single file or path in Cloud Storage.
 message CloudStoragePath {
   // A url representing a file or path (no wildcards) in Cloud Storage.
   // Example: gs://[BUCKET_NAME]/dictionary.txt
@@ -244,6 +265,19 @@ message CloudStoragePath {
 
 // Options defining BigQuery table and row identifiers.
 message BigQueryOptions {
+  // How to sample rows if not all rows are scanned. Meaningful only when used
+  // in conjunction with rows_limit. If not specified, scanning would start
+  // from the top.
+  enum SampleMethod {
+    SAMPLE_METHOD_UNSPECIFIED = 0;
+
+    // Scan from the top (default).
+    TOP = 1;
+
+    // Randomly pick the row to start scanning. The scanned rows are contiguous.
+    RANDOM_START = 2;
+  }
+
   // Complete BigQuery table reference.
   BigQueryTable table_reference = 1;
 
@@ -255,6 +289,8 @@ message BigQueryOptions {
   // rest of the rows are omitted. If not set, or if set to 0, all rows will be
   // scanned. Cannot be used in conjunction with TimespanConfig.
   int64 rows_limit = 3;
+
+  SampleMethod sample_method = 4;
 }
 
 // Shared message indicating Cloud storage type.