Public/koboldcpp
1
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2026-09-19 17:25:07 +02:00
Code Issues Packages Projects Releases Wiki Activity

improve responses api

Browse Source
This commit is contained in:
Concedo
2026-03-28 18:42:15 +08:00
parent f768b2a4bd
commit 1e787cd03a
1 changed files with 71 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files
koboldcpp.py
+71 -7
View File
@@ -3249,7 +3249,9 @@ def determine_tool_json_to_use(genparams, curr_ctx, assistant_message_start, is_
def compress_tools_array(tools_array):
tools_array_filtered = []
for tool_dict in tools_array:
tool_data = tool_dict['function']
tool_data = tool_dict
if 'function' in tool_dict:
tool_data = tool_dict['function']
tool_props = {}
params = tool_data.get("parameters", {})
props = params.get("properties", {})
@@ -3644,13 +3646,30 @@ ws ::= | " " | "\n" [ \t]{0,20}
raw_instructions = genparams.get('instructions', '')
if isinstance(raw_input, str):
genparams['messages'] = [{"role": "user", "content": raw_input}]
elif isinstance(raw_input, list):
genparams['messages'] = raw_input
elif isinstance(raw_input, list): # Convert Responses API input items to chat messages format
converted = []
for item in raw_input:
if isinstance(item, dict):
role = item.get("role", "user")
content = item.get("content", "")
# content can itself be a list of typed parts
if isinstance(content, list):
parts = []
for part in content:
if part.get("type") == "input_text":
parts.append({"type": "text", "text": part.get("text", "")})
elif part.get("type") == "input_image":
img = part.get("image_url", part.get("source", {}))
parts.append({"type": "image_url", "image_url": {"url": img}})
content = parts
converted.append({"role": role, "content": content})
elif isinstance(item, str):
converted.append({"role": "user", "content": item})
genparams['messages'] = converted
else:
genparams['messages'] = []
if raw_instructions and isinstance(raw_instructions, str) and raw_instructions!="":
if raw_instructions and isinstance(raw_instructions, str):
genparams['messages'].insert(0, {"role": "system", "content": raw_instructions})
genparams['stream'] = False
transform_genparams(genparams, 4, use_jinja) # Delegate to the chat-completions transform by re-running as format 4
return genparams
@@ -4069,7 +4088,7 @@ class KcppServerRequestHandler(http.server.SimpleHTTPRequestHandler):
#tool calls resolution
tool_calls = []
if api_format == 4 or api_format == 2:
if api_format == 4 or api_format == 2 or api_format == 8:
using_openai_tools = genparams.get('using_openai_tools', False)
if using_openai_tools:
# first, check and potentially segment multiple tags for multi-tool calls
@@ -4115,7 +4134,52 @@ class KcppServerRequestHandler(http.server.SimpleHTTPRequestHandler):
res = {"model": friendlymodelname,"created_at": str(datetime.now(timezone.utc).isoformat()),"message":{"role":"assistant","content":recvtxt},"done": True,"done_reason":currfinishreason,"total_duration": 1,"load_duration": 1,"prompt_eval_count": prompttokens,"prompt_eval_duration": 1,"eval_count": comptokens,"eval_duration": 1}
elif api_format == 8:
resp_id = f"resp-A{genparams.get('oai_uniqueid', 1)}"
res = {"id": resp_id, "object": "response", "created_at": int(time.time()), "model": friendlymodelname, "output": [ { "type": "message", "role": "assistant", "content": [ {"type": "output_text", "text": recvtxt} ] } ], "usage": { "input_tokens": prompttokens, "output_tokens": comptokens, "total_tokens": prompttokens + comptokens } }
output_item_id = f"msg_0{genparams.get('oai_uniqueid', 1)}"
output_items = []
if recvtxt is not None: # Add text message if there's content
output_items.append({
"type": "message",
"id": output_item_id,
"status": "completed",
"role": "assistant",
"content": [{"type": "output_text", "text": recvtxt, "annotations": [], "logprobs": []}]
})
if tool_calls and len(tool_calls) > 0: # Add function call items if tool calls exist
for tc in tool_calls:
output_items.append({"type": "function_call", "id": tc.get("id", ""), "call_id": tc.get("id", ""), "name": tc.get("function", {}).get("name", ""), "arguments": tc.get("function", {}).get("arguments", "{}"), "status": "completed"})
res = {
"id": resp_id,
"object": "response",
"created_at": int(time.time()),
"completed_at": int(time.time()),
"incomplete_details": None,
"previous_response_id": None,
"truncation": "disabled",
"parallel_tool_calls": False,
"text": {"format": {"type": "text"},"verbosity": "medium"},
"instructions": genparams.get('instructions', None),
"model": friendlymodelname,
"status": "completed" if currfinishreason != "error" else "failed",
"error": None,
"metadata": {},
"tools": genparams.get('tools', []),
"tool_choice": "auto",
"background": False,
"service_tier": "default",
"safety_identifier": None,
"prompt_cache_key": None,
"max_tool_calls": None,
"store": False,
"output": output_items,
"top_p": genparams.get("top_p", 0.92),
"max_output_tokens":genparams.get("max_length", None),
"presence_penalty": genparams.get("presence_penalty", 0),
"frequency_penalty": genparams.get("frequency_penalty", 0),
"top_logprobs": 0,
"temperature": genparams.get("temperature", 1),
"reasoning": {"effort": None, "summary": None},
"usage": {"input_tokens": prompttokens, "output_tokens": comptokens, "total_tokens": prompttokens + comptokens, "input_tokens_details": {"cached_tokens": 0}, "output_tokens_details": {"reasoning_tokens": 0}}
}
else: #kcpp format
res = {"results": [{"text": recvtxt, "tool_calls": tool_calls, "finish_reason": currfinishreason, "logprobs":logprobsdict, "prompt_tokens": prompttokens, "completion_tokens": comptokens}]}