fix: gemini provider not support display name for image input

Author: bbohlender

README.md

@@ -103,3 +103,8 @@ Explore the complete **prai** documentation to learn everything from basic conce
 - **[Models](https://pmndrs.github.io/prai/concepts/modeltypescript)** - AI provider abstraction supporting OpenAI, Groq, Gemini, and mock providers
 - **[History](https://pmndrs.github.io/prai/concepts/historytypescript)** - Conversation tracking, reference system, subtasks, and multimodal support
 
+
+TODO:
+
+[ ] write model to log
+[ ] allow overriding response and cost field through step options 

packages/prai/src/provider/gemini.ts

@@ -79,13 +79,12 @@ export function gemini(
 function messageContentToPartUnion(content: Message['content'][number]): Part {
   switch (content.type) {
     case 'image_url':
-      return { fileData: { fileUri: content.image_url.url, displayName: 'input.png', mimeType: 'image/png' } }
+      return { fileData: { fileUri: content.image_url.url, mimeType: 'image/png' } }
     case 'input_audio':
       return {
         inlineData: {
           data: content.input_audio.data,
           mimeType: `audio/${content.input_audio.format}`,
-          displayName: `input.${content.input_audio.format}`,
         },
       }
     case 'text':