llm_chat/llm.h at main · hzy0117/llm_chat · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74

/*
 * Copyright (c) 2006-2025, RT-Thread Development Team
 *
 * SPDX-License-Identifier: Apache-2.0
 *
 * Change Logs:
 * Date           Author       Notes
 * 2025/02/01     Rbb666       Add license info
 * 2025/02/01     Rbb666       Add history support
 * 2025/02/10     CXSforHPU    Add llm history support
 */
#ifndef __LLM_H_
#define __LLM_H_

#include <rtthread.h>
#include <stdio.h>
#include <string.h>
#include <cJSON.h>

#ifndef PKG_LLMCHAT_DBG
#define LLM_DBG(fmt, ...)
#else
#define LLM_DBG(fmt, ...)                    \
    do                                       \
    {                                        \
        rt_kprintf("[\033[32mLLM\033[0m] "); \
        rt_kprintf(fmt, ##__VA_ARGS__);      \
    } while (0)
#endif

#ifndef LLM_HISTORY_LINES
#define LLM_HISTORY_LINES 5
#endif

enum llm_input_stat
{
    LLM_WAIT_NORMAL,
    LLM_WAIT_SPEC_KEY,
    LLM_WAIT_FUNC_KEY,
};

struct llm_obj
{
    struct rt_thread thread;
    rt_uint8_t thread_stack[PKG_LLM_THREAD_STACK_SIZE];

    struct rt_semaphore rx_sem;

    enum llm_input_stat stat;

    int argc;

    char llm_history[LLM_HISTORY_LINES][PKG_LLM_CMD_BUFFER_SIZE];
    rt_uint16_t history_count;
    rt_uint16_t history_current;

    char line[PKG_LLM_CMD_BUFFER_SIZE];
    rt_uint16_t line_position;
    rt_uint8_t line_curpos;

    rt_device_t device;
    rt_err_t (*rx_indicate)(rt_device_t dev, rt_size_t size);

    cJSON *messages;
    char *(*get_answer)(cJSON *messages);
};
typedef struct llm_obj *llm_t;

char *get_llm_answer(cJSON *messages);
void add_message2messages(const char *input_buffer, char *role, struct llm_obj *handle);
cJSON *create_message(const char *input_buffer, char *role);
void clear_messages(struct llm_obj *handle);
#endif