Function create_lm_server

src/server/server.cpp:864–1090

@brief create lm server @param models the model list @param downloader the downloader @param default_tag the default tag @param port the port @return the server

///@param port the port
///@return the server
std::unique_ptr<WebServer> create_lm_server(model_list& models, ModelDownloader& downloader, program_args_t& args) {
    auto server = std::make_unique<WebServer>(args.host, args.port, args.cors);
    auto rest_handler = std::make_shared<RestHandler>(models, downloader, args);
    
    // Register Ollama-compatible routes
    server->register_handler("POST", "/api/show",
        [rest_handler](const http::request<http::string_body>& req,
            std::function<void(const json&)> send_response,
            std::function<void(const json&, bool)> send_streaming_response,
            std::shared_ptr<HttpSession> session,
            std::shared_ptr<CancellationToken> cancellation_token) {
                json request_json;
                if (!req.body().empty()) {
                    request_json = json::parse(req.body());
                }
                rest_handler->handle_show(request_json, send_response, send_streaming_response);
        });

    server->register_handler("POST", "/api/generate", 
        [rest_handler](const http::request<http::string_body>& req, 
                      std::function<void(const json&)> send_response,
                      std::function<void(const json&, bool)> send_streaming_response,
                      std::shared_ptr<HttpSession> session,
                      std::shared_ptr<CancellationToken> cancellation_token) {
            json request_json;
            if (!req.body().empty()) {
                request_json = json::parse(req.body());
            }
            rest_handler->handle_generate(request_json, send_response, send_streaming_response, cancellation_token);
        });
    
    server->register_handler("POST", "/api/chat",
        [rest_handler](const http::request<http::string_body>& req,
                      std::function<void(const json&)> send_response,
                      std::function<void(const json&, bool)> send_streaming_response,
                      std::shared_ptr<HttpSession> session,
                      std::shared_ptr<CancellationToken> cancellation_token) {
            json request_json;
            if (!req.body().empty()) {
                request_json = json::parse(req.body());
            }
            rest_handler->handle_chat(request_json, send_response, send_streaming_response, cancellation_token);
        });

    server->register_handler("GET", "/api/ps",
        [rest_handler](const http::request<http::string_body>& req,
                      std::function<void(const json&)> send_response,
                      std::function<void(const json&, bool)> send_streaming_response,
                      std::shared_ptr<HttpSession> session,
                      std::shared_ptr<CancellationToken> cancellation_token) {
            json request_json;
            rest_handler->handle_ps(request_json, send_response, send_streaming_response);
        });

    server->register_handler("POST", "/api/embeddings",
        [rest_handler](const http::request<http::string_body>& req,
                      std::function<void(const json&)> send_response,
                      std::function<void(const json&, bool)> send_streaming_response,