Felipe97/llama-cpp-compiled

1#pragma once2 3#include <functional>4#include <string>5 6// openai-like client for CLI7struct cli_client {8    std::string server_base; // base url, for example "http://127.0.0.1:8080"9    std::string last_error;  // set when wait_health() fails10 11    std::string model; // optional, set when the server has multiple models (router mode)12 13    // simple GET request, returns the raw response body14    // throws std::runtime_error on transport error or non-2xx status15    std::string get(const std::string & path);16 17    // simple POST request, returns the raw response body18    // throws std::runtime_error on transport error or non-2xx status19    std::string post(const std::string & path, const std::string & body);20 21    // POST request with an SSE streaming response22    // on_data is invoked per "data:" event with the raw event payload23    // returns after the stream is finished (empty string on graceful exit)24    // otherwise, the raw error response body25    std::string post_sse(const std::string & path,26                          const std::string & body,27                          const std::function<bool()> & should_stop,28                          const std::function<void(const std::string &)> & on_data);29 30    // poll /health until the server is ready to accept requests31    // returns false if is_aborted returned true or the server is unreachable32    bool wait_health(const std::function<bool()> & is_aborted);33};34