happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

translating structured generation request parameters from responses api format to completions api format

This commit is contained in:
Batuhan Bozkurt 2026-03-30 19:19:26 +03:00
parent 64ac9ab66a
commit 6cbefe4a0e
1 changed files with 29 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

29
tools/server/server-common.cpp
View File

@ -1430,6 +1430,35 @@ json convert_responses_to_chatcmpl(const json & response_body) {
chatcmpl_body["max_tokens"] = response_body["max_output_tokens"];
}
// Convert Responses API text.format to Chat Completions response_format
if (response_body.contains("text") && response_body.at("text").contains("format")) {
const json & text_format = response_body.at("text").at("format");
std::string format_type = json_value(text_format, "type", std::string());
if (format_type == "json_schema") {
json json_schema_wrapper = {
{"name", json_value(text_format, "name", std::string("default"))},
{"schema", json_value(text_format, "schema", json::object())},
};
if (text_format.contains("strict")) {
json_schema_wrapper["strict"] = text_format.at("strict");
}
if (text_format.contains("description")) {
json_schema_wrapper["description"] = text_format.at("description");
}
chatcmpl_body["response_format"] = {
{"type", "json_schema"},
{"json_schema", json_schema_wrapper},
};
} else if (format_type == "json_object") {
chatcmpl_body["response_format"] = {
{"type", "json_object"},
};
}
chatcmpl_body.erase("text");
}
return chatcmpl_body;
}