fix(ai): fix serialization for Content with no parts field
#6964
Conversation
📝 PRs merging into main branchOur main branch should always be in a releasable state. If you are working on a larger change, or if you don't want this change to see the light of the day just yet, consider using a feature branch first, and only merge into the main branch when the code complete and ready to be released. |
Test Results 12 files - 12 12 suites - 12 19s ⏱️ -15s For more details on these failures, see this check. Results for commit 75e5e14. ± Comparison against base commit af24598. ♻️ This comment has been updated with latest results. |
Coverage Report 1Affected ProductsNo changes between base commit (af24598) and merge commit (8c73fb4).Test Logs |
Size Report 1Affected Products
Test Logs |
| @@ -90,7 +90,7 @@ constructor(public val role: String? = "user", public val parts: List<Part>) { | |||
| @Serializable | |||
| internal data class Internal( | |||
| @EncodeDefault val role: String? = "user", | |||
| val parts: List<InternalPart> | |||
| @EncodeDefault val parts: List<InternalPart> = emptyList() | |||
There was a problem hiding this comment.
can you add a comment here noting why we need the encodeDefault in this case, as it's not immediately apparent.
There was a problem hiding this comment.
I actually can't remember why I added this here 😳 the tests still pass even if we don't use encodeDefault.
There was a problem hiding this comment.
Devs can send a request with an empty list, so makes sense that @EncodeDefault that is used for encoding, not decoding, isn't 100% necessary. Should be safe to remove
There was a problem hiding this comment.
SGTM. Don't forget to add an entry to the changelog
… into rpf/fix-content-part-serialization
In the Gemini Dev API,
Content#partsis optional, while in the Vertex AI API it is marked as required.This PR should settle the difference by exposing a list of parts with at least 1 element whenever
Content#partsis not present.Golden file PR: FirebaseExtended/vertexai-sdk-test-data#40