Generate code based on 2024-10 open api spec and add rerank #155

rohanshah18 · 2024-10-12T01:45:07Z

Problem

Add support for reranking documents.

Solution

Rerank is a part of 2024-10 openAPI spec, so the underlying code was regenerated based on the latest spec. Following are the important changes made as a part of this PR.

Updated build-oas.sh to include all control plane, data plane, and inference modules.
Inference is now broken out of control plane so the underlying Inference code generated by openAPI spec is now updated. This also resulted in instantiating a separate ApiClient which is a part of Inference module now instead of the shared ApiClient class of control plane. As a part of this refactoring, the customOkHttpClient is now a part of the PineconeConfig class, so once the customOkHttpClient is set by the user at the time of instantiating Pinecone using its builder inner class, it'll be stored in the PineconeConfig instance which can further be leveraged for Inference api calls as well.
Control and data plane is now prepended with db_ so this resulted in updating import statements for a lot of existing classes.
Added the following two methods for rerank:
a. Method that accepts required parameters only and the optional parameters are set to default. More details on the api and its default parameters can be found here

      rerank(String model,
                       String query,
                       List<Map<String, String>> documents)

b. Method that accepts all parameters:

      rerank(String model,
                       String query,
                       List<Map<String, String>> documents,
                       List<String> rankFields,
                       int topN,
                       boolean returnDocuments,
                       Map<String, Object> parameters)

Added integration tests for both rerank methods.
Added an example in the README file.

Note: Everything under org/openapitools directory can be skipped as a part of the review process. All of the code under this directory was generated using openApi.

Type of Change

Bug fix (non-breaking change which fixes an issue)
New feature (non-breaking change which adds functionality)
Breaking change (fix or feature that would cause existing functionality to not work as expected)
This change requires a documentation update
Infrastructure change (CI configs, etc)
Non-code change (docs, etc)
None of the above: (explain here)

Test Plan

Added integration tests.

rohanshah18 · 2024-10-15T20:52:41Z

README.md

@@ -487,7 +487,7 @@ UpdateResponse updateResponse = index.update("v1", values, "example-namespace");

 # Collections

-Collections fall under data plane operations.
+Collections fall under control plane operations.


Typo from before

rohanshah18 · 2024-10-15T20:52:55Z

README.md

@@ -583,6 +583,66 @@ EmbeddingsList embeddings = inference.embed(embeddingModel, parameters, inputs);
 List<Embedding> embeddedData = embeddings.getData();
 ```

+## Rerank


Add example for rerank

rohanshah18 · 2024-10-15T20:55:46Z

codegen/build-oas.sh

@@ -3,7 +3,7 @@
 set -eux -o pipefail

 version=$1 # e.g. 2024-07
-modules=("control")
+modules=("db_control" "db_data" "inference")


replaced control with db_control

added db_data for bulk import (actual feature is going to be a part of the next PR).

added inference which is now broken out of control plane

rohanshah18 · 2024-10-15T20:56:10Z

src/integration/java/io/pinecone/helpers/TestUtilities.java

-import io.pinecone.proto.DescribeIndexStatsResponse;
-import io.pinecone.proto.NamespaceSummary;
-import org.openapitools.control.client.model.*;
+import org.openapitools.db_control.client.model.*;


example of updating import statement since db_ is now appended.

rohanshah18 · 2024-10-15T20:56:53Z

src/integration/java/io/pinecone/integration/inference/EmbedTest.java

@@ -4,8 +4,8 @@
 import io.pinecone.clients.Pinecone;
 import org.junit.jupiter.api.Assertions;
 import org.junit.jupiter.api.Test;
-import org.openapitools.control.client.ApiException;
-import org.openapitools.control.client.model.EmbeddingsList;
+import org.openapitools.inference.client.ApiException;


Example of inference being its own module instead of being a part of the control plane which resulted in updating the import statement

rohanshah18 · 2024-10-15T20:57:25Z

src/integration/java/io/pinecone/integration/inference/RerankTest.java

+    private static final Inference inference = pinecone.getInferenceClient();
+
+    @Test
+    public void testRerank() throws ApiException {


Test all parameters for rerank

rohanshah18 · 2024-10-15T20:57:37Z

src/integration/java/io/pinecone/integration/inference/RerankTest.java

+    }
+
+    @Test
+    public void testRerankWithRequiredParameters() throws ApiException {


Test required parameters for rerank

rohanshah18 · 2024-10-15T20:58:33Z

src/main/java/io/pinecone/clients/Inference.java

-
-import java.util.List;
-import java.util.Map;
+import io.pinecone.configs.PineconeConfig;


Since Inference is broken out of control plane, the ApiClient for control plane can't be used as done before

rohanshah18 · 2024-10-15T20:59:03Z

src/main/java/io/pinecone/clients/Inference.java

+     * @return RerankResult containing the ranked documents and their scores.
+     * @throws ApiException If the API call fails, an ApiException is thrown.
+     */
+    public RerankResult rerank(String model,


Rerank with required params only

rohanshah18 · 2024-10-15T20:59:12Z

src/main/java/io/pinecone/clients/Inference.java

+     * @return RerankResult containing the ranked documents and their scores.
+     * @throws ApiException If the API call fails, an ApiException is thrown.
+     */
+    public RerankResult rerank(String model,


Rerank with required + optional params

rohanshah18 · 2024-10-15T21:00:00Z

src/main/java/io/pinecone/clients/Pinecone.java

@@ -1033,7 +1033,7 @@ public Builder withProxy(String proxyHost, int proxyPort) {
         * @return A new {@link Pinecone} instance configured based on the builder parameters.
         */
        public Pinecone build() {
-            PineconeConfig config = new PineconeConfig(apiKey, sourceTag, proxyConfig);
+            PineconeConfig config = new PineconeConfig(apiKey, sourceTag, proxyConfig, customOkHttpClient);


customOkHttpClient being passed to the PineconeConfig constructor at the time of initializing PineconeConfig in Pinecone class

rohanshah18 · 2024-10-15T21:00:42Z

src/main/java/io/pinecone/configs/PineconeConfig.java

    }

    /**
-     * Constructs a {@link PineconeConfig} instance with the specified API key, source tag, control plane proxy
-     * configuration, and data plane proxy configuration.
+     * Constructs a {@link PineconeConfig} instance with the specified API key, source tag, a HTTP proxy configuration,


customOkHttpClient is now a member of PineconeConfig. It makes a lot of sense since gRPC custom channel is already a part of this config class.

src/main/java/org/openapitools/db_control/client/ApiClient.java

src/main/java/org/openapitools/inference/client/ApiClient.java

rohanshah18 added 5 commits October 11, 2024 14:05

generate code for 2024-10 db_data, db_control, & inference

1578580

refactor for adding db_ and breaking inference out

e0adfa8

refactor getInference() to prevent instantiating a new object

760c650

add support for reranking docs

61fe905

add example to the README

81ea401

rohanshah18 requested review from jhamon, aulorbe, austin-denoble and ssmith-pc October 15, 2024 20:52

rohanshah18 marked this pull request as ready for review October 15, 2024 20:52

rohanshah18 commented Oct 15, 2024

View reviewed changes

clean up

2d01d2a

github-advanced-security bot found potential problems Oct 17, 2024

View reviewed changes

src/main/java/org/openapitools/db_control/client/ApiClient.java Dismissed Show dismissed Hide dismissed

src/main/java/org/openapitools/inference/client/ApiClient.java Dismissed Show dismissed Hide dismissed

rohanshah18 changed the title ~~Rerank~~ Generate code based on 2024-10 open api spec and add rerank Oct 17, 2024

rohanshah18 closed this Oct 18, 2024

rohanshah18 deleted the rshah/rerank branch October 18, 2024 13:39

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Generate code based on 2024-10 open api spec and add rerank #155

Generate code based on 2024-10 open api spec and add rerank #155

rohanshah18 commented Oct 12, 2024 •

edited

Loading

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

rohanshah18 Oct 15, 2024

Generate code based on 2024-10 open api spec and add rerank #155

Generate code based on 2024-10 open api spec and add rerank #155

Conversation

rohanshah18 commented Oct 12, 2024 • edited Loading

Problem

Solution

Type of Change

Test Plan

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rohanshah18 commented Oct 12, 2024 •

edited

Loading