Added Seed in ModelParams for Summarization Requests (#39)

* Auto-generated. Updating Vectara public protos. (a35abced31a122305cf72002a42f58a3bd8a8e20).

* Auto-generated. Updating Vectara public protos. (e20273e10ba98df4fb29251ea09ca6f91820f5a4).

* Auto-generated. Updating Vectara public protos. (a90130634b50a73e75b8c8b09fb4cbef80275fd0).

* Auto-generated. Updating Vectara public protos. (43bd8df244bc9eaf138181a5566f182fa331da13).

* Update serving.proto

* Update serving.proto

* Auto-generated. Updating Vectara public protos. (3d3b6e8495472a032b1d288bf119f202e14786a0).

* Auto-generated. Updating Vectara public protos. (cd64312c808e9e859bfd0f5c156fc292ee99914f).

---------

Co-authored-by: bitbucket-pipelines <commits-noreply@bitbucket.org>
Co-authored-by: Paul Wozniczka <25128922+pwoznic@users.noreply.github.com>

admin.proto

@@ -57,8 +57,10 @@ message Corpus {
 
   repeated Dimension custom_dimensions = 13;
   repeated FilterAttribute filter_attributes = 14;
+
 }
 
+
 // A custom dimension is additional numeric metadata that you want to affect
 // Vectara's scoring.  For example, these could be "number of stars" ratings,
 // or other business metrics like a product's margins that you want to use
@@ -120,6 +122,20 @@ message CreateCorpusResponse {
   Status status = 2;
 }
 
+message UpdateCorpusRequest {
+  // Corpus id to update.
+  uint32 id = 1;
+  // Name to be updated.
+  optional string name = 2;
+  // Description to be updated.
+  optional string description = 3;
+}
+
+message UpdateCorpusResponse {
+  // The Status of the update.
+  Status status = 1;
+}
+
 message DeleteCorpusRequest {
   option (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_schema) = {
     json_schema: { required: [ "customerId", "corpusId" ] }
@@ -205,6 +221,8 @@ message CorpusSize {
   int64 epoch_secs = 1;
 
   // The size of the corpus.
+  // This is the sum of the number of characters in the
+  // text and metadata of all documents in the corpus.
   uint64 size = 2;
 }
 
@@ -224,6 +242,7 @@ message ReadCorpusRequest {
   bool read_custom_dimensions = 1004;
   // Set to true to read the filter attributes of the corpus.
   bool read_filter_attributes = 1005;
+
 }
 
 message ReadCorpusResponse {
@@ -249,6 +268,7 @@ message ReadCorpusResponse {
     repeated FilterAttribute filter_attribute = 6;
     // Status of the filter attributes.
     Status filter_attribute_status = 1006;
+
   }
 
   // Information about the requested corpora.

common.proto

@@ -96,3 +96,4 @@ message StorageQuota {
   // storage quota.
   int64 num_metadata_chars = 2;
 }
+

core_services.proto

@@ -22,6 +22,7 @@ message IndexCoreDocumentRequest {
   int64 corpus_id = 2;
   // The document being indexed.
   com.vectara.indexing.CoreDocument document = 3;
+
 }
 
 message IndexCoreDocumentResponse {

services.proto

@@ -31,6 +31,7 @@ message IndexDocumentRequest {
   int64 corpus_id = 2;
   // The Document to index.
   com.vectara.indexing.Document document = 3;
+
 }
 
 message IndexDocumentResponse {
@@ -104,6 +105,14 @@ service AdminService {
     };
   }
 
+  rpc UpdateCorpus(com.vectara.admin.UpdateCorpusRequest)
+    returns (com.vectara.admin.UpdateCorpusResponse) {
+    option (google.api.http) = {
+      post: "/v1/update-corpus"
+      body: "*"
+    };
+  }
+
   rpc DeleteCorpus(com.vectara.admin.DeleteCorpusRequest) returns (com.vectara.admin.DeleteCorpusResponse){
     option (google.api.http) = {
       post: "/v1/delete-corpus"

serving.proto

@@ -122,6 +122,10 @@ message SummarizationRequest {
     // Higher values penalize new tokens based on whether they appear in the text so far,
     // increasing the model's likelihood to talk about new topics.
     optional float presence_penalty = 20;
+    // This feature is in Beta. If specified, the LLM will make a best effort to sample deterministically,
+    // such that repeated requests with the same seed and parameters should return the same result.
+    // Note:- Seed value is only used when temperature is set to 0.
+    optional uint32 seed = 25;
   }
   ModelParams model_params = 215;
 
@@ -159,6 +163,8 @@ message QueryRequest {
     // chars_before is used for showing the end user the characters leading up
     // to the result snippet.  This can help the end-user understand the
     // context of that result. Ignored if sentences_before is set.
+    // Vectara will capture the full sentence that contains the captured characters,
+    // so as to not lose the meaning caused by a truncated word or sentence.
     int32 chars_before = 5 [
       (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_field) = {
         example: "30"
@@ -167,6 +173,8 @@ message QueryRequest {
     // chars_after is used for showing the end user the characters after the
     // result snippet.  This can help the end-user understand the context of
     // that result. Ignored if sentences_before is set.
+    // Vectara will capture the full sentence that contains the captured characters,
+    // so as to not lose the meaning caused by a truncated word or sentence.
     int32 chars_after = 10 [
       (grpc.gateway.protoc_gen_openapiv2.options.openapiv2_field) = {
         example: "30"
@@ -216,13 +224,17 @@ message QueryRequest {
   message RerankingConfig {
     // Which reranking model to use if reranking.  Currently, the only IDs
     // available are:
-    // - 272725717, HuggingFace Open Source x-attentional reranker
     // - 272725718, Maximum Marginal Relevance Reranker
+    // - 272725719, Vectara Multilingual Reranker v1 Reranker
     uint32 reranker_id = 5;
 
     // Reranker-specific parameters.  The numbering starts from 100, and moves
     // upwards in increments of 5.
     optional MMRConfig mmr_config = 100;
+
+    // User function that will be executed on each search result.
+    // May return a double or null. If this is null the item is skipped.
+    optional string user_function = 105;
   }
   RerankingConfig reranking_config = 30;
 
@@ -322,7 +334,6 @@ message Summary {
   repeated Status status = 1000;
   // Populated for streaming requests only.
   int32 future_id = 1010;
-
 }
 
 // A document part that matched a query.