Remove format from double helper function

[rlazo]

We can omit the format for the double-precision floating point type

See firebase/firebase-ios-sdk#13990 for further context

[github-actions]

Release note changes

The following release notes were modified. Please ensure they look correct.

Release Notes

firebase-vertexai

### {{firebase_vertexai}} version 16.0.2 {: #vertex-ai_v16-0-2}

* {{fixed}} Improved error message when using an invalid location. GitHub [#6428](//github.com/firebase/firebase-android-sdk/issues/6428){: .external}

* {{fixed}} Fixed issue where Firebase App Check error tokens were unintentionally missing from the requests. GitHub [#6409](//github.com/firebase/firebase-android-sdk/issues/6409){: .external}

* {{fixed}} Clarified in the documentation that `Schema.integer` and `Schema.float` only provide hints to the model. GitHub [#6420](//github.com/firebase/firebase-android-sdk/issues/6420){: .external}

* {{fixed}} Fixed issue were `Schema.double` set the format parameter in `Schema`. GitHub [#6432](//github.com/firebase/firebase-android-sdk/issues/6432){: .external}

[github-actions]

Vertex AI Mock Responses Check ⚠️

A newer major version of the mock responses for Vertex AI unit tests is available. update_responses.sh should be updated to clone the latest version of the responses: v5.1

[github-actions]

Test Results

 20 files   -  34   20 suites   - 34   14s ⏱️ - 1m 48s
113 tests  - 402  113 ✅  - 401  0 💤  - 1  0 ❌ ±0 
226 runs   - 804  226 ✅  - 802  0 💤  - 2  0 ❌ ±0 

Results for commit a63d93d. ± Comparison against base commit 80019ca.

This pull request removes 515 and adds 113 tests. Note that renamed tests count towards both.

com.google.firebase.dataconnect.AnyValueSerializerUnitTest ‑ descriptor should have expected values
com.google.firebase.dataconnect.AnyValueSerializerUnitTest ‑ deserialize() should throw UnsupportedOperationException
com.google.firebase.dataconnect.AnyValueSerializerUnitTest ‑ serialize() should throw UnsupportedOperationException
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(Boolean) creates an object with the expected value
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(Double) creates an object with the expected value (edge cases)
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(Double) creates an object with the expected value (normal cases)
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(List) creates an object with the expected value (edge cases)
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(List) creates an object with the expected value (normal cases)
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(Map) creates an object with the expected value (edge cases)
com.google.firebase.dataconnect.AnyValueUnitTest ‑ constructor(Map) creates an object with the expected value (normal cases)
…

com.google.firebase.vertexai.GenerativeModelTesting ‑ exception thrown when using invalid location
com.google.firebase.vertexai.GenerativeModelTesting ‑ system calling in request
com.google.firebase.vertexai.SchemaTests ‑ basic schema declaration
com.google.firebase.vertexai.SchemaTests ‑ full schema declaration
com.google.firebase.vertexai.StreamingSnapshotTests ‑ citation parsed correctly
com.google.firebase.vertexai.StreamingSnapshotTests ‑ empty content
com.google.firebase.vertexai.StreamingSnapshotTests ‑ http errors
com.google.firebase.vertexai.StreamingSnapshotTests ‑ image rejected
com.google.firebase.vertexai.StreamingSnapshotTests ‑ invalid api key
com.google.firebase.vertexai.StreamingSnapshotTests ‑ invalid json
…

♻️ This comment has been updated with latest results.

[rlazo]

cc @andrewheard

[daymxn]

Is double the default? Is this behavior we should rely on?

e: I just saw the ios issue which helped provide context.

Although, isn't this an issue with the backend? Shouldn't it enforce it?

[rlazo]

Is double the default? Is this behavior we should rely on?

e: I just saw the ios issue which helped provide context.

Although, isn't this an issue with the backend? Shouldn't it enforce it?

double is the largest floating-point number type in Kotlin (not counting Big* numbers). As in the case of long, we assume the type with the largest capacity is generic enough to represent the range of values the model can return if no is format specified.

As documented in the kdoc, even if everything is working as expected, integer and float are hints in the Schema.

[daymxn]

SGTM. May be worth directly linking to andrew's PR in the PR desc for context/auditing