File size: 3,438 Bytes
af72a8e
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
import os
import requests
import gradio as gr
from datetime import datetime
from llama_cpp import Llama  # GPT4ALL용

# ───────────────────────── λͺ¨λΈ λ‘œλ”© ─────────────────────────
llm = Llama(model_path="./ggjt-model.bin")

# ───────────────────────── NEIS API 호좜 ν•¨μˆ˜ ─────────────────────────
def get_school_info(region_code, school_name, api_key):
    url = f"https://open.neis.go.kr/hub/schoolInfo?KEY={api_key}&Type=json&pIndex=1&pSize=1&SCHUL_NM={school_name}&ATPT_OFCDC_SC_CODE={region_code}"
    res = requests.get(url)
    data = res.json()
    school = data.get("schoolInfo", [{}])[1].get("row", [{}])[0]
    return school.get("SD_SCHUL_CODE"), school.get("ATPT_OFCDC_SC_CODE")

def get_schedule(region_code, school_code, year, month, api_key):
    from_ymd = f"{year}{month:02}01"
    to_ymd = f"{year}{month:02}31"
    url = f"https://open.neis.go.kr/hub/SchoolSchedule?KEY={api_key}&Type=json&pIndex=1&pSize=100&ATPT_OFCDC_SC_CODE={region_code}&SD_SCHUL_CODE={school_code}&AA_FROM_YMD={from_ymd}&AA_TO_YMD={to_ymd}"
    res = requests.get(url)
    data = res.json()
    return data.get("SchoolSchedule", [{}])[1].get("row", [])

# ───────────────────────── GPT4ALL 응닡 생성 ─────────────────────────
def generate_gpt_answer(user_question, events):
    context = "\n".join([f"{e['EVENT_NM']} : {e['AA_YMD']}" for e in events])
    prompt = f"""
### Instruction:
λ‹€μŒμ€ μƒλ¦¬μ΄ˆλ“±ν•™κ΅μ˜ 2024λ…„ 7μ›” ν•™μ‚¬μΌμ •μž…λ‹ˆλ‹€:

{context}

질문: "{user_question}"

μœ„ 학사일정 μ •λ³΄λ§Œ μ°Έκ³ ν•΄μ„œ μ§ˆλ¬Έμ— ν•΄λ‹Ήν•˜λŠ” λ‚ μ§œλ₯Ό μ°Ύμ•„, μžμ—°μŠ€λŸ½κ³  κ°„κ²°ν•œ ν•œκ΅­μ–΄ λ¬Έμž₯으둜 λ‹΅ν•΄μ£Όμ„Έμš”. 예: "2024λ…„ 7μ›” 24μΌμž…λ‹ˆλ‹€."
### Response:
"""

    result = ""
    for output in llm(prompt, stop=["###"], stream=True):
        result += output["choices"][0]["text"]
    return result.strip()

# ───────────────────────── 전체 처리 ν•¨μˆ˜ ─────────────────────────
def answer_question(user_question):
    api_key = os.environ.get("NEIS_API_KEY", "a69e08342c8947b4a52cd72789a5ecaf")
    school_name = "μƒλ¦¬μ΄ˆλ“±ν•™κ΅"
    region_code = "R10"
    year = 2024
    month = 7

    school_code, region_code = get_school_info(region_code, school_name, api_key)
    events = get_schedule(region_code, school_code, year, month, api_key)

    if not events:
        return "학사일정 정보λ₯Ό λΆˆλŸ¬μ˜€μ§€ λͺ»ν–ˆμŠ΅λ‹ˆλ‹€."

    return generate_gpt_answer(user_question, events)

# ───────────────────────── Gradio μΈν„°νŽ˜μ΄μŠ€ ─────────────────────────
description = "πŸ“… μƒλ¦¬μ΄ˆλ“±ν•™κ΅ 2024λ…„ 7μ›” 학사일정 기반 GPT μ±—λ΄‡μž…λ‹ˆλ‹€. 예: '여름방학은 μ–Έμ œμ•Ό?'"

gr.Interface(fn=answer_question,
             inputs=gr.Textbox(lines=2, placeholder="예: 여름방학식은 μ–Έμ œμ•Ό?", label="질문 μž…λ ₯"),
             outputs=gr.Textbox(label="λ‹΅λ³€"),
             title="πŸŽ“ GPT4ALL 학사일정 Q&A",
             description=description).launch()