refactor(config.yaml): update client configuration for openai-compatible

- Updated `type` to `openai-compatible`
- Added `name` field for clarity
- Adjusted `api_base` URL to include `/v1`
- Simplified and updated model configurations for consistency

roles/home/aichat/config.yaml

@@ -1,23 +1,12 @@
 clients:
-  - type: ollama
+  - type: openai-compatible
-    api_base: https://ollama.giugl.io
+    name: ollama
+    api_base: https://ollama.giugl.io/v1
     models:
-    - name: mistral:7b-instruct-v0.3-fp16
+      - name: pino 
-      max_input_tokens: 32000
-      max_output_tokens: 8192
-
-    - name: llama3:8b-instruct-fp16
         max_input_tokens: 8192
-      max_output_tokens: 8192
+        max_output_tokens: 16000
-      
-    - name: phi3:14b-medium-4k-instruct-q8_0
-      max_input_tokens: 128000
-      max_output_tokens: 8192
 
       - name: pino-coder          
         max_input_tokens: 8192
-      max_output_tokens: 8192
+        max_output_tokens: 16000
-
-  - type: openai
-    api_key: null
-    api_base: https://api.openai.com/v1