gemini 1.5 + vision fix

README.md

@@ -4,7 +4,7 @@
 # Witsy
 
 Generative AI desktop application:
-- OpenAI, Ollama, Anthropic, MistralAI and Groq models supported
+- OpenAI, Ollama, Anthropic, MistralAI, Google and Groq models supported
 - Chat completion and image generation with Vision models support
 - Prompt anywhere allows to generate content directly in any application
 - AI commands runnable on highlighted text in almost any application
@@ -59,10 +59,11 @@ npm start
 
 ## Prerequisites
 
-To use OpenAI, Anthropic or Mistral AI models, you need to enter your API key:
+To use OpenAI, Anthropic, Google or Mistral AI models, you need to enter your API key:
 - [OpenAI](https://platform.openai.com/api-keys)
 - [Anthropic](https://console.anthropic.com/settings/keys)
 - [MistralAI](https://console.mistral.ai/api-keys/)
+- [Google](https://aistudio.google.com/app/apikey)
 
 To use Ollama models, you need to install [Ollama](https://ollama.com) and download some [models](https://ollama.com/library).
 
@@ -76,13 +77,13 @@ To use Internet search you need a [Tavily API key](https://app.tavily.com/home).
 
 ## TODO
 
-- [ ] Google Gemini API
 - [ ] i18n
 - [ ] File upload for retrieval (??)
 - [ ] Proper database (SQLite3) storage (??)
 
 ## DONE
 
+- [x] Google Gemini API
 - [x] Prompt anywhere
 - [x] Cancel commands
 - [x] GPT-4o support

package.json

@@ -1,7 +1,7 @@
 {
   "name": "witsy",
   "productName": "Witsy",
-  "version": "1.6.2",
+  "version": "1.7.0",
   "description": "Witsy: desktop AI assistant",
   "repository": {
     "type": "git",

src/services/engine.ts

@@ -35,11 +35,6 @@ export default class LlmEngine {
     throw new Error('Not implemented')
   }
 
-  // eslint-disable-next-line @typescript-eslint/no-unused-vars
-  isVisionModel(model: string): boolean {
-    return false
-  }
-
   async getModels(): Promise<any[]> {
     throw new Error('Not implemented')
   }
@@ -86,6 +81,15 @@ export default class LlmEngine {
     return this.config.engines[this.getName()].models.chat
   }
 
+  isVisionModel(model: string): boolean {
+    for (const filter of this.getVisionModels()) {
+      if (minimatch(model, filter)) {
+        return true
+      }
+    }
+    return false
+  }
+
   requiresVisionModelSwitch(thread: Message[], currentModel: string): boolean {
 
     // if we already have a vision or auto switch is disabled

src/services/google.ts

@@ -31,10 +31,6 @@ export default class extends LlmEngine {
     return ['gemini-1.5-flash-latest', 'models/gemini-1.5-pro-latest', 'models/gemini-pro-vision', '*vision*']
   }
 
-  isVisionModel(model: string): boolean {
-    return this.getVisionModels().includes(model)
-  }
-
   getRountingModel(): string | null {
     return null
   }

src/services/groq.ts

@@ -31,11 +31,6 @@ export default class extends LlmEngine {
     return []
   }
 
-  // eslint-disable-next-line @typescript-eslint/no-unused-vars
-  isVisionModel(model: string): boolean {
-    return false
-  }
-
   getRountingModel(): string|null {
     return null
   }

src/services/mistralai.ts

@@ -32,10 +32,6 @@ export default class extends LlmEngine {
     return []
   }
 
-  isVisionModel(model: string): boolean {
-    return this.getVisionModels().includes(model)
-  }
-
   getRountingModel(): string|null {
     return null
   }

src/services/ollama.ts

@@ -35,10 +35,6 @@ export default class extends LlmEngine {
     return ['llava-llama3:latest', 'llava:latest', '*llava*']
   }
 
-  isVisionModel(model: string): boolean {
-    return this.getVisionModels().includes(model) || model.includes('llava')
-  }
-
   getRountingModel(): string|null {
     return null
   }

src/services/openai.ts

@@ -36,10 +36,6 @@ export default class extends LlmEngine {
     return [/*'*4o*', */ 'gpt-4-turbo', 'gpt-4-vision', '*vision*']
   }
 
-  isVisionModel(model: string): boolean {
-    return this.getVisionModels().includes(model)
-  }
-
   getRountingModel(): string | null {
     return 'gpt-3.5-turbo'
   }