343 lines
14 KiB
C++
343 lines
14 KiB
C++
#ifndef HUMANUS_LLM_H
|
|
#define HUMANUS_LLM_H
|
|
|
|
#include "config.h"
|
|
#include "logger.h"
|
|
#include "schema.h"
|
|
#include "mcp/common/httplib.h"
|
|
#include <map>
|
|
#include <string>
|
|
#include <memory>
|
|
#include <vector>
|
|
#include <functional>
|
|
#include <stdexcept>
|
|
#include <future>
|
|
|
|
namespace humanus {
|
|
|
|
class LLM {
|
|
private:
|
|
static std::map<std::string, std::shared_ptr<LLM>> _instances;
|
|
|
|
std::unique_ptr<httplib::Client> client_;
|
|
|
|
std::shared_ptr<LLMSettings> llm_config_;
|
|
|
|
std::shared_ptr<ToolHelper> tool_helper_;
|
|
|
|
public:
|
|
// Constructor
|
|
LLM(const std::string& config_name, const std::shared_ptr<LLMSettings>& llm_config = nullptr, const std::shared_ptr<ToolHelper>& tool_helper = nullptr) : llm_config_(llm_config), tool_helper_(tool_helper) {
|
|
if (!llm_config_) {
|
|
if (Config::get_instance().llm().find(config_name) == Config::get_instance().llm().end()) {
|
|
throw std::invalid_argument("LLM config not found: " + config_name);
|
|
}
|
|
llm_config_ = std::make_shared<LLMSettings>(Config::get_instance().llm().at(config_name));
|
|
}
|
|
if (!llm_config_->oai_tool_support && !tool_helper_) {
|
|
if (Config::get_instance().tool_helper().find(config_name) == Config::get_instance().tool_helper().end()) {
|
|
throw std::invalid_argument("Tool helper config not found: " + config_name);
|
|
}
|
|
tool_helper_ = std::make_shared<ToolHelper>(Config::get_instance().tool_helper().at(config_name));
|
|
}
|
|
client_ = std::make_unique<httplib::Client>(llm_config_->base_url);
|
|
client_->set_default_headers({
|
|
{"Authorization", "Bearer " + llm_config_->api_key}
|
|
});
|
|
client_->set_read_timeout(llm_config_->timeout);
|
|
}
|
|
|
|
// Get the singleton instance
|
|
static std::shared_ptr<LLM> get_instance(const std::string& config_name = "default", const std::shared_ptr<LLMSettings>& llm_config = nullptr) {
|
|
if (_instances.find(config_name) == _instances.end()) {
|
|
_instances[config_name] = std::make_shared<LLM>(config_name, llm_config);
|
|
}
|
|
return _instances[config_name];
|
|
}
|
|
|
|
/**
|
|
* @brief Format the message list to the format that LLM can accept
|
|
* @param messages Message object message list
|
|
* @return The formatted message list
|
|
* @throws std::invalid_argument If the message format is invalid or missing necessary fields
|
|
* @throws std::runtime_error If the message type is not supported
|
|
*/
|
|
json format_messages(const std::vector<Message>& messages) {
|
|
json formatted_messages = json::array();
|
|
|
|
auto concat_content = [](const json& lhs, const json& rhs) -> json {
|
|
if (lhs.is_string() && rhs.is_string()) {
|
|
return lhs.get<std::string>() + "\n" + rhs.get<std::string>(); // Maybe other delimiter?
|
|
}
|
|
json res = json::array();
|
|
if (lhs.is_string()) {
|
|
res.push_back({
|
|
{"type", "text"},
|
|
{"text", lhs.get<std::string>()}
|
|
});
|
|
} else if (lhs.is_array()) {
|
|
res.insert(res.end(), lhs.begin(), lhs.end());
|
|
}
|
|
if (rhs.is_string()) {
|
|
res.push_back({
|
|
{"type", "text"},
|
|
{"text", rhs.get<std::string>()}
|
|
});
|
|
} else if (rhs.is_array()) {
|
|
res.insert(res.end(), rhs.begin(), rhs.end());
|
|
}
|
|
return res;
|
|
};
|
|
|
|
for (const auto& message : messages) {
|
|
if (message.content.empty() && message.tool_calls.empty()) {
|
|
continue;
|
|
}
|
|
formatted_messages.push_back(message.to_json());
|
|
if (!llm_config_->oai_tool_support) {
|
|
if (formatted_messages.back()["role"] == "tool") {
|
|
std::string tool_results_str = formatted_messages.back().dump(2);
|
|
formatted_messages.back() = {
|
|
{"role", "user"},
|
|
{"content", tool_results_str}
|
|
};
|
|
} else if (!formatted_messages.back()["tool_calls"].empty()) {
|
|
if (formatted_messages.back()["content"].is_null()) {
|
|
formatted_messages.back()["content"] = "";
|
|
}
|
|
std::string tool_calls_str = tool_helper_->dump(formatted_messages.back()["tool_calls"]);
|
|
formatted_messages.back().erase("tool_calls");
|
|
formatted_messages.back()["content"] = concat_content(formatted_messages.back()["content"], tool_calls_str);
|
|
}
|
|
}
|
|
}
|
|
|
|
for (const auto& message : formatted_messages) {
|
|
if (message["role"] != "user" && message["role"] != "assistant" && message["role"] != "system" && message["role"] != "tool") {
|
|
throw std::invalid_argument("Invalid role: " + message["role"].get<std::string>());
|
|
}
|
|
}
|
|
|
|
size_t i = 0, j = -1;
|
|
for (; i < formatted_messages.size(); i++) {
|
|
if (i == 0 || formatted_messages[i]["role"] != formatted_messages[j]["role"]) {
|
|
formatted_messages[++j] = formatted_messages[i];
|
|
} else {
|
|
formatted_messages[j]["content"] = concat_content(formatted_messages[j]["content"], formatted_messages[i]["content"]);
|
|
}
|
|
}
|
|
|
|
formatted_messages.erase(formatted_messages.begin() + j + 1, formatted_messages.end());
|
|
|
|
return formatted_messages;
|
|
}
|
|
|
|
/**
|
|
* @brief Send a request to the LLM and get the reply
|
|
* @param messages The conversation message list
|
|
* @param system_prompt Optional system message
|
|
* @param max_retries The maximum number of retries
|
|
* @return The generated assistant content
|
|
* @throws std::invalid_argument If the message is invalid or the reply is empty
|
|
* @throws std::runtime_error If the API call fails
|
|
*/
|
|
std::string ask(
|
|
const std::vector<Message>& messages,
|
|
const std::string& system_prompt = "",
|
|
int max_retries = 3
|
|
) {
|
|
json formatted_messages = json::array();
|
|
|
|
if (!system_prompt.empty()) {
|
|
formatted_messages.push_back({
|
|
{"role", "system"},
|
|
{"content", system_prompt}
|
|
});
|
|
}
|
|
|
|
json _formatted_messages = format_messages(messages);
|
|
formatted_messages.insert(formatted_messages.end(), _formatted_messages.begin(), _formatted_messages.end());
|
|
|
|
json body = {
|
|
{"model", llm_config_->model},
|
|
{"messages", formatted_messages},
|
|
{"temperature", llm_config_->temperature},
|
|
{"max_tokens", llm_config_->max_tokens}
|
|
};
|
|
|
|
std::string body_str = body.dump();
|
|
|
|
int retry = 0;
|
|
|
|
while (retry <= max_retries) {
|
|
// send request
|
|
auto res = client_->Post(llm_config_->end_point, body_str, "application/json");
|
|
|
|
if (!res) {
|
|
logger->error("Failed to send request: " + httplib::to_string(res.error()));
|
|
} else if (res->status == 200) {
|
|
try {
|
|
json json_data = json::parse(res->body);
|
|
return json_data["choices"][0]["message"]["content"].get<std::string>();
|
|
} catch (const std::exception& e) {
|
|
logger->error("Failed to parse response: " + std::string(e.what()));
|
|
}
|
|
} else {
|
|
logger->error("Failed to send request: status=" + std::to_string(res->status) + ", body=" + res->body);
|
|
}
|
|
|
|
retry++;
|
|
|
|
if (retry > max_retries) {
|
|
break;
|
|
}
|
|
|
|
// wait for a while before retrying
|
|
std::this_thread::sleep_for(std::chrono::milliseconds(500));
|
|
|
|
logger->info("Retrying " + std::to_string(retry) + "/" + std::to_string(max_retries));
|
|
}
|
|
|
|
throw std::runtime_error("Failed to get response from LLM");
|
|
}
|
|
|
|
/**
|
|
* @brief Send a request to the LLM with tool functions
|
|
* @param messages The conversation message list
|
|
* @param system_prompt Optional system message
|
|
* @param next_step_prompt Optinonal prompt message for the next step
|
|
* @param timeout The request timeout (seconds)
|
|
* @param tools The tool list
|
|
* @param tool_choice The tool choice strategy
|
|
* @param max_retries The maximum number of retries
|
|
* @return The generated assistant message (content, tool_calls)
|
|
* @throws std::invalid_argument If the tool, tool choice or message is invalid
|
|
* @throws std::runtime_error If the API call fails
|
|
*/
|
|
json ask_tool(
|
|
const std::vector<Message>& messages,
|
|
const std::string& system_prompt = "",
|
|
const std::string& next_step_prompt = "",
|
|
const json& tools = {},
|
|
const std::string& tool_choice = "auto",
|
|
int max_retries = 3
|
|
) {
|
|
if (tool_choice != "none" && tool_choice != "auto" && tool_choice != "required") {
|
|
throw std::invalid_argument("Invalid tool_choice: " + tool_choice);
|
|
}
|
|
|
|
json formatted_messages = json::array();
|
|
|
|
if (!system_prompt.empty()) {
|
|
formatted_messages.push_back({
|
|
{"role", "system"},
|
|
{"content", system_prompt}
|
|
});
|
|
}
|
|
|
|
json _formatted_messages = format_messages(messages);
|
|
formatted_messages.insert(formatted_messages.end(), _formatted_messages.begin(), _formatted_messages.end());
|
|
|
|
if (!next_step_prompt.empty()) {
|
|
if (formatted_messages.empty() || formatted_messages.back()["role"] != "user") {
|
|
formatted_messages.push_back({
|
|
{"role", "user"},
|
|
{"content", next_step_prompt}
|
|
});
|
|
} else {
|
|
if (formatted_messages.back()["content"].is_string()) {
|
|
formatted_messages.back()["content"] = formatted_messages.back()["content"].get<std::string>() + "\n\n" + next_step_prompt;
|
|
} else if (formatted_messages.back()["content"].is_array()) {
|
|
formatted_messages.back()["content"].push_back({
|
|
{"type", "text"},
|
|
{"text", next_step_prompt}
|
|
});
|
|
}
|
|
}
|
|
}
|
|
|
|
if (!tools.empty()) {
|
|
for (const json& tool : tools) {
|
|
if (!tool.contains("type")) {
|
|
throw std::invalid_argument("Tool must contain 'type' field but got: " + tool.dump(2));
|
|
}
|
|
}
|
|
if (tool_choice == "required" && tools.empty()) {
|
|
throw std::invalid_argument("No tool available for required tool choice");
|
|
}
|
|
if (!tools.is_array()) {
|
|
throw std::invalid_argument("Tools must be an array");
|
|
}
|
|
}
|
|
|
|
json body = {
|
|
{"model", llm_config_->model},
|
|
{"messages", formatted_messages},
|
|
{"temperature", llm_config_->temperature},
|
|
{"max_tokens", llm_config_->max_tokens},
|
|
{"tool_choice", tool_choice}
|
|
};
|
|
|
|
if (llm_config_->oai_tool_support) {
|
|
body["tools"] = tools;
|
|
} else {
|
|
if (body["messages"].empty() || body["messages"].back()["role"] != "user") {
|
|
body["messages"].push_back({
|
|
{"role", "user"},
|
|
{"content", tool_helper_->hint(tools.dump(2))}
|
|
});
|
|
} else if (body["messages"].back()["content"].is_string()) {
|
|
body["messages"].back()["content"] = body["messages"].back()["content"].get<std::string>() + "\n\n" + tool_helper_->hint(tools.dump(2));
|
|
} else if (body["messages"].back()["content"].is_array()) {
|
|
body["messages"].back()["content"].push_back({
|
|
{"type", "text"},
|
|
{"text", tool_helper_->hint(tools.dump(2))}
|
|
});
|
|
}
|
|
}
|
|
|
|
std::string body_str = body.dump();
|
|
|
|
int retry = 0;
|
|
|
|
while (retry <= max_retries) {
|
|
// send request
|
|
auto res = client_->Post(llm_config_->end_point, body_str, "application/json");
|
|
|
|
if (!res) {
|
|
logger->error("Failed to send request: " + httplib::to_string(res.error()));
|
|
} else if (res->status == 200) {
|
|
try {
|
|
json json_data = json::parse(res->body);
|
|
json message = json_data["choices"][0]["message"];
|
|
if (!llm_config_->oai_tool_support && message["content"].is_string()) {
|
|
message = tool_helper_->parse(message["content"].get<std::string>());
|
|
}
|
|
return message;
|
|
} catch (const std::exception& e) {
|
|
logger->error("Failed to parse response: " + std::string(e.what()));
|
|
}
|
|
} else {
|
|
logger->error("Failed to send request: status=" + std::to_string(res->status) + ", body=" + res->body);
|
|
}
|
|
|
|
retry++;
|
|
|
|
if (retry > max_retries) {
|
|
break;
|
|
}
|
|
|
|
// wait for a while before retrying
|
|
std::this_thread::sleep_for(std::chrono::milliseconds(500));
|
|
|
|
logger->info("Retrying " + std::to_string(retry) + "/" + std::to_string(max_retries));
|
|
}
|
|
|
|
throw std::runtime_error("Failed to get response from LLM");
|
|
}
|
|
};
|
|
|
|
} // namespace humanus
|
|
|
|
#endif // HUMANUS_LLM_H
|