Type Members

1.  sealed abstract class CDCOperation extends GeneratedEnum

   Type of operation performed on vertices in this state entry.

   Type of operation performed on vertices in this state entry. Ordered by application precedence: DELETE → SPLIT → UNION → APPEND → ALIVE.

2.  final case class CDCSequence(state: Seq[CDCState] = _root_.scala.Seq.empty, commitId: ByteString = _root_.com.google.protobuf.ByteString.EMPTY, commitTimestamp: Long = 0L, unknownFields: UnknownFieldSet = _root_.scalapb.UnknownFieldSet.empty) extends GeneratedMessage with Updatable[CDCSequence] with Product with Serializable

   A grouped set of all operations that affect a cluster of related vector_ids in a single processing run.

   A grouped set of all operations that affect a cluster of related vector_ids in a single processing run.

   Operations within state are ordered by type: DELETE → SPLIT → UNION → APPEND → ALIVE. All entries share connected vector_ids — if operation A references vector_id X and operation B references vector_id X (as either vector_id or prev_vector_id), they belong to the same CDCSequence.

   Example — Ultimate scenario (one Kafka message): state = [ { operation: DELETE, prev_vector_id: a, user_id: [ifv_dead] }, { operation: SPLIT, vector_id: d, prev_vector_id: a, user_id: [ifv1] }, { operation: SPLIT, vector_id: e, prev_vector_id: b, user_id: [ifv3] }, { operation: UNION, vector_id: c, prev_vector_id: a, user_id: [ifv2] }, { operation: UNION, vector_id: c, prev_vector_id: b, user_id: [ifv4] }, { operation: APPEND, vector_id: c, user_id: [ifv_new] }, ]

   Example — ALIVE heartbeat (separate message, no structural changes): state = [ { operation: ALIVE, vector_id: a }, ]

   state

   Chronologically ordered state changes (DELETE → SPLIT → UNION → APPEND → ALIVE)

   commitId

   Meta information for debugging and metrics, shows which run the changes were pulled from

   commitTimestamp

   Run batch processing completion time

   Annotations

   @SerialVersionUID()

3.  final case class CDCState(vectorId: Option[ByteString] = _root_.scala.None, prevVectorId: Option[ByteString] = _root_.scala.None, userId: Seq[ByteString] = _root_.scala.Seq.empty, operation: CDCOperation = io.bidmachine.protobuf.vectorid.cdc.v1.CDCOperation.CDC_OPERATION_UNSPECIFIED, unknownFields: UnknownFieldSet = _root_.scalapb.UnknownFieldSet.empty) extends GeneratedMessage with Updatable[CDCState] with Product with Serializable

   A single state change entry within a CDC message.

   A single state change entry within a CDC message. Represents one operation affecting a set of IFA/IFV vertices.

   Field presence by operation: DELETE: prev_vector_id + user_id SPLIT: vector_id + prev_vector_id + user_id UNION: vector_id + prev_vector_id + user_id APPEND: vector_id + user_id ALIVE: vector_id only (TTL heartbeat, no user_id changes)

   vectorId

   Vector UUID after the operation. Present for APPEND, UNION, SPLIT, ALIVE. Absent for DELETE.

   prevVectorId

   Previous vector UUID before the operation. Present for DELETE, UNION, SPLIT. Absent for APPEND, ALIVE.

   userId

   List of IFA/IFV source IDs affected by this state change. Present for DELETE, SPLIT, UNION, APPEND. Empty for ALIVE (no mapping changes).

   operation

   The operation that produced this state change

   Annotations

   @SerialVersionUID()

4.  final case class VectorIdStateSnapshot(vectorId: ByteString = _root_.com.google.protobuf.ByteString.EMPTY, prevVectorId: Seq[ByteString] = _root_.scala.Seq.empty, userId: Seq[ByteString] = _root_.scala.Seq.empty, unknownFields: UnknownFieldSet = _root_.scalapb.UnknownFieldSet.empty) extends GeneratedMessage with Updatable[VectorIdStateSnapshot] with Product with Serializable

   The state we broadcast for streaming API Examples: Split: prev run: vector_id = 1; prev_vector_id = [ ]; user_id = [a, b, c, d, e] - original entity curr run: vector_id = 11; prev_vector_id = [1]; user_id = [a, b, c] - a message of split curr run: vector_id = 12; prev_vector_id = [1]; user_id = [d, e] - another message of split

   The state we broadcast for streaming API Examples: Split: prev run: vector_id = 1; prev_vector_id = [ ]; user_id = [a, b, c, d, e] - original entity curr run: vector_id = 11; prev_vector_id = [1]; user_id = [a, b, c] - a message of split curr run: vector_id = 12; prev_vector_id = [1]; user_id = [d, e] - another message of split

   Merge: prev run: vector_id = 11; prev_vector_id = [1]; user_id = [a, b] - original entity prev run: vector_id = 39; prev_vector_id = [3]; user_id = [e, f] - original entity curr run: vector_id = 45; prev_vector_id = [11, 39]; user_id = [a, b, e, f] - a message of merge

   Append: prev run: vector_id = 15; prev_vector_id = [1]; user_id = [a, b] - original entity curr run: vector_id = 15; prev_vector_id = [15]; user_id = [a, b, h] - a message of append

   Create: curr run: vector_id = 99; prev_vector_id = [ ]; user_id = [e, i, o] - a message of create

   vectorId

   An actual vector id

   prevVectorId

   List of previous vector ids, can be empty

   userId

   List of source ids (ifa/ifv/...), can't be empty

   Annotations

   @SerialVersionUID()