GithubBackup/gemini-balance
1
0
Fork 0
mirror of https://github.com/snailyp/gemini-balance.git synced 2026-06-05 23:59:45 +08:00
Code Issues Packages Projects Releases Wiki Activity

feat: add support for the n parameter in OpenAI-compatible requests. Now, when you make a request to the /v1/chat/completions endpoint with the n parameter, it will be correctly mapped to candidateCount in the Gemini API request, allowing you to receive multiple completions.

Browse Source
This commit is contained in:
zenyanbo
2025-08-11 17:06:38 +08:00
parent a6558b4668
commit f58ae2b340
3 changed files with 49 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

69
app/handler/response_handler.py
View File

@@ -42,21 +42,35 @@ class GeminiResponseHandler(ResponseHandler):
def _handle_openai_stream_response(
response: Dict[str, Any], model: str, finish_reason: str, usage_metadata: Optional[Dict[str, Any]]
) -> Dict[str, Any]:
text, reasoning_content, tool_calls, _ = _extract_result(
response, model, stream=True, gemini_format=False
)
if not text and not tool_calls and not reasoning_content:
delta = {}
else:
delta = {"content": text, "reasoning_content": reasoning_content, "role": "assistant"}
if tool_calls:
delta["tool_calls"] = tool_calls
choices = []
candidates = response.get("candidates", [])
for candidate in candidates:
index = candidate.get("index", 0)
text, reasoning_content, tool_calls, _ = _extract_result(
{"candidates": [candidate]}, model, stream=True, gemini_format=False
)
if not text and not tool_calls and not reasoning_content:
delta = {}
else:
delta = {"content": text, "reasoning_content": reasoning_content, "role": "assistant"}
if tool_calls:
delta["tool_calls"] = tool_calls
choice = {
"index": index,
"delta": delta,
"finish_reason": finish_reason
}
choices.append(choice)
template_chunk = {
"id": f"chatcmpl-{uuid.uuid4()}",
"object": "chat.completion.chunk",
"created": int(time.time()),
"model": model,
"choices": [{"index": 0, "delta": delta, "finish_reason": finish_reason}],
"choices": choices,
}
if usage_metadata:
template_chunk["usage"] = {"prompt_tokens": usage_metadata.get("promptTokenCount", 0), "completion_tokens": usage_metadata.get("candidatesTokenCount",0), "total_tokens": usage_metadata.get("totalTokenCount", 0)}
@@ -66,26 +80,31 @@ def _handle_openai_stream_response(
def _handle_openai_normal_response(
response: Dict[str, Any], model: str, finish_reason: str, usage_metadata: Optional[Dict[str, Any]]
) -> Dict[str, Any]:
text, reasoning_content, tool_calls, _ = _extract_result(
response, model, stream=False, gemini_format=False
)
choices = []
candidates = response.get("candidates", [])
for i, candidate in enumerate(candidates):
text, reasoning_content, tool_calls, _ = _extract_result(
{"candidates": [candidate]}, model, stream=False, gemini_format=False
)
choice = {
"index": i,
"message": {
"role": "assistant",
"content": text,
"reasoning_content": reasoning_content,
"tool_calls": tool_calls,
},
"finish_reason": finish_reason,
}
choices.append(choice)
return {
"id": f"chatcmpl-{uuid.uuid4()}",
"object": "chat.completion",
"created": int(time.time()),
"model": model,
"choices": [
{
"index": 0,
"message": {
"role": "assistant",
"content": text,
"reasoning_content": reasoning_content,
"tool_calls": tool_calls,
},
"finish_reason": finish_reason,
}
],
"choices": choices,
"usage": {"prompt_tokens": usage_metadata.get("promptTokenCount", 0), "completion_tokens": usage_metadata.get("candidatesTokenCount",0), "total_tokens": usage_metadata.get("totalTokenCount", 0)},
}