feat: Add GigaChat v2 models and fix /models endpoint for Xcode 26

Currently, the /models endpoint returns only OpenAI models, which causes issues when iOS developers use gpt2giga with Xcode 26's AI assistant integration. When Xcode requests the model list and users select an OpenAI model (like gpt-4), subsequent /completions requests fail because GigaChat API doesn't recognize these model names.

Additionally, Xcode 26 sends requests to /models? (with a trailing question mark), which currently returns invalid_request_error due to strict path matching.

This PR adds real GigaChat v2 models to the models list and fixes the endpoint handling:

- Added GigaChat v2 models: GigaChat-2, GigaChat-2-Max, GigaChat-2-Pro
- Fixed /models endpoint: Now handles URLs with query parameters (e.g., /models?) for Xcode 26 compatibility

Author: artemkolyadin

gpt2giga/__init__.py

@@ -362,7 +362,7 @@ def send_to_gigachat_stream(self, data: dict) -> Iterator[dict]:
             yield process_gigachat_stream(chunk, gpt_model, is_tool_call)
 
     def do_GET(self):
-        if self.path in ("/models", "/v1/models"):
+        if self.path.startswith(("/models", "/v1/models")):
             self.handle_models_request()
         else:
             self.handle_proxy_chat()

gpt2giga/gpt2giga_models.json

@@ -1,6 +1,24 @@
 {
     "object": "list",
     "data": [
+      {
+        "id": "GigaChat-2",
+        "object": "model",
+        "created": 1700000001,
+        "owned_by": "salutedevices"
+      },
+      {
+        "id": "GigaChat-2-Max",
+        "object": "model",
+        "created": 1700000002,
+        "owned_by": "salutedevices"
+      },
+      {
+        "id": "GigaChat-2-Pro",
+        "object": "model",
+        "created": 1700000003,
+        "owned_by": "salutedevices"
+      },
       {
         "id": "gpt-4-turbo",
         "object": "model",