refactor(android): drop the single-prompt merge once ml kit sends older gemini nano whole conversations

Abierto
#364 1 comentario 0 reacciones 0 asignados Ver en GitHub

Los mantenedores suelen responder en 1 día

Nadie ha tomado este issue todavía.

Evaluación

Dificultad
3/5
Tiempo estimado
1-2 días
Aptitud para principiantes
30/100
Tipo de issue
Refactorización
Claridad
Bien especificado
Estado de actividad
Activo
Stack tecnológico
android, react-native, typescript
Área
ai, mobile-dev

Línea de trabajo

Start by reading GeminiNano.buildRequest in @react-native-ai/android and the change made in PR #367, then compare it with ADK's NanoSingleTextPrompt reference in #358. The work is blocked until ML Kit or AICore sends older Gemini Nano models the full multi-Content conversation or until those models age out. Done means the single TextPart merge is removed, system instructions and turns are sent separately, and behavior is verified on an affected nano-v2 device such as a Pixel 9 Pro.

Escrito por el modelo de indexación a partir del texto del issue.

Descripción

Blocked enhancement

Gemini Nano without system prompt support (isSystemPromptAvailable() false, such as nano-v2 on a Pixel 9 Pro with AICore aicore_20260219) reads only the first text part of an ML Kit genai-prompt request. A conversation sent as several Content turns, or with a separate system instruction, gets an answer to the first part only.

@react-native-ai/android (PR #367) therefore merges the system instruction and the text of every turn into one TextPart on those models, with images first, like ADK's NanoSingleTextPrompt (#358). Models with system prompt support get the instruction and turns separately.

genai-agent's AgentSession is no alternative: it rebuilds the same multi-Content request.

Remove the merge in GeminiNano.buildRequest once ML Kit or AICore sends older Gemini Nano the whole conversation (or those models age out).

Lenguaje dominante
TypeScript
Estrellas
1.4k
Forks
64
Merge medio
11 h 24 min
PR fusionados (30 d)
91

Preparar el entorno

Primeros pasos

  1. Lee el issue completo y luego la guía de contribución del proyecto.
  2. Comenta en el issue que vas a ocuparte — evita que dos personas hagan lo mismo.
  3. Haz un fork del repositorio y trabaja en una rama.
  4. Abre un pull request que haga referencia al número del issue.

Más de callstackincubator/ai

Todos los issues de callstackincubator/ai

Issues similares

Más issues de TypeScript

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.