pluem9 commited on
Commit ·
f87ae4b
1
Parent(s): 31e3978
demo1
Browse files- src/streamlit_app.py +97 -84
- src/utils/__pycache__/loader.cpython-313.pyc +0 -0
- src/utils/__pycache__/qg_baseline.cpython-313.pyc +0 -0
- src/utils/extractors.py +0 -12
- src/utils/loader.py +137 -0
- src/utils/qg_baseline.py +307 -0
src/streamlit_app.py
CHANGED
|
@@ -1,111 +1,124 @@
|
|
| 1 |
import streamlit as st
|
| 2 |
import pandas as pd
|
| 3 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 4 |
|
| 5 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 6 |
|
| 7 |
# ---------- CSS ----------
|
| 8 |
st.markdown("""
|
| 9 |
<style>
|
| 10 |
-
[data-testid="stTextInput"],
|
| 11 |
-
[data-testid="stDecoratedTextInput"] { display: none !important; }
|
| 12 |
-
|
| 13 |
.block-container { max-width: 980px; margin: 2rem auto; }
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
|
| 21 |
-
}
|
| 22 |
-
|
| 23 |
-
.
|
| 24 |
-
|
| 25 |
-
font-weight: 700; font-size: 22px;
|
| 26 |
-
}
|
| 27 |
-
|
| 28 |
-
div[data-testid="stFileUploaderDropzone"]{
|
| 29 |
-
border: 2px dashed rgba(0,0,0,.25);
|
| 30 |
-
border-radius: 14px;
|
| 31 |
-
padding: 26px;
|
| 32 |
-
transition: border-color .2s, background .2s;
|
| 33 |
-
}
|
| 34 |
-
div[data-testid="stFileUploaderDropzone"]:hover{
|
| 35 |
-
border-color: #4f8bf9;
|
| 36 |
-
background: rgba(79,139,249,0.06);
|
| 37 |
-
}
|
| 38 |
-
|
| 39 |
-
.hint{ color: #8a8a8a; font-size: 13px; margin-top: 4px; }
|
| 40 |
-
.file-chip{
|
| 41 |
-
display:inline-flex; align-items:center; gap:8px;
|
| 42 |
-
background: rgba(79,139,249,0.12);
|
| 43 |
-
color:#1f4fbf; border-radius: 999px; padding:6px 12px; font-size:13px;
|
| 44 |
-
}
|
| 45 |
</style>
|
| 46 |
""", unsafe_allow_html=True)
|
| 47 |
|
| 48 |
# ---------- Title ----------
|
| 49 |
st.markdown("<h1 style='margin:0 0 10px 0;'>Quiz Generation</h1>", unsafe_allow_html=True)
|
| 50 |
-
st.caption("อัปโหลดสไลด์/ไฟล์ข้อความ เพื่อสร้างแบบทดสอบอัตโนมัติ")
|
| 51 |
|
| 52 |
-
# ---------- Upload
|
| 53 |
st.markdown('<div class="upload-title">📘 Upload File</div>', unsafe_allow_html=True)
|
| 54 |
-
st.write("")
|
| 55 |
-
|
| 56 |
uploaded = st.file_uploader(
|
| 57 |
"ลากไฟล์มาวาง หรือกดเพื่อเลือกไฟล์",
|
| 58 |
type=["txt", "pdf", "docx", "csv"],
|
| 59 |
label_visibility="collapsed",
|
| 60 |
accept_multiple_files=False
|
| 61 |
)
|
| 62 |
-
|
| 63 |
st.markdown('<div class="hint">รองรับ .txt .pdf .docx .csv • ขนาดไม่ควรเกิน ~50MB</div>', unsafe_allow_html=True)
|
| 64 |
-
st.markdown("</div>", unsafe_allow_html=True)
|
| 65 |
-
|
| 66 |
-
# ---------- Session State ----------
|
| 67 |
-
if "uploaded_file" not in st.session_state:
|
| 68 |
-
st.session_state.uploaded_file = None
|
| 69 |
-
if "quiz_generated" not in st.session_state:
|
| 70 |
-
st.session_state.quiz_generated = False
|
| 71 |
|
| 72 |
-
|
| 73 |
-
if uploaded
|
| 74 |
-
st.session_state.uploaded_file = uploaded
|
| 75 |
-
|
| 76 |
-
# ---------- Preview & Actions ----------
|
| 77 |
-
if st.session_state.uploaded_file:
|
| 78 |
-
uploaded = st.session_state.uploaded_file
|
| 79 |
st.markdown(f'<div class="file-chip">📄 {uploaded.name}</div>', unsafe_allow_html=True)
|
| 80 |
st.divider()
|
| 81 |
-
|
| 82 |
-
|
| 83 |
-
|
| 84 |
-
|
| 85 |
-
|
| 86 |
-
|
| 87 |
-
|
| 88 |
-
|
| 89 |
-
|
| 90 |
-
|
| 91 |
-
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
|
| 95 |
-
|
| 96 |
-
df = pd.read_csv(uploaded)
|
| 97 |
-
st.subheader("Preview (CSV)")
|
| 98 |
-
st.dataframe(df, use_container_width=True)
|
| 99 |
-
|
| 100 |
-
else:
|
| 101 |
-
st.info("ไฟล์ PDF/DOCX จะอ่านได้หลังจากเพิ่มไลบรารีอ่านไฟล์ (เช่น pdfplumber / python-docx)")
|
| 102 |
|
| 103 |
# ---------- Action ----------
|
| 104 |
-
|
| 105 |
-
|
| 106 |
-
|
| 107 |
-
|
| 108 |
-
|
| 109 |
-
|
| 110 |
-
|
| 111 |
-
st.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import streamlit as st
|
| 2 |
import pandas as pd
|
| 3 |
+
import io
|
| 4 |
+
import os
|
| 5 |
+
import sys
|
| 6 |
+
import csv
|
| 7 |
+
import json
|
| 8 |
|
| 9 |
+
# ---------- Setup Import Path ----------
|
| 10 |
+
BASE_DIR = os.path.dirname(__file__)
|
| 11 |
+
UTILS_DIR = os.path.join(BASE_DIR, "src", "utils")
|
| 12 |
+
if UTILS_DIR not in sys.path:
|
| 13 |
+
sys.path.append(UTILS_DIR)
|
| 14 |
+
|
| 15 |
+
from utils.loader import load_to_text
|
| 16 |
+
from utils.qg_baseline import generate_quiz, to_csv_rows
|
| 17 |
+
|
| 18 |
+
# ---------- Page Config ----------
|
| 19 |
+
st.set_page_config(page_title="Quiz Generation", page_icon="🧠", layout="wide")
|
| 20 |
|
| 21 |
# ---------- CSS ----------
|
| 22 |
st.markdown("""
|
| 23 |
<style>
|
| 24 |
+
[data-testid="stTextInput"], [data-testid="stDecoratedTextInput"] { display:none!important; }
|
|
|
|
|
|
|
| 25 |
.block-container { max-width: 980px; margin: 2rem auto; }
|
| 26 |
+
.upload-title { display:flex; gap:12px; align-items:center; margin-bottom:8px; font-weight:700; font-size:22px; }
|
| 27 |
+
div[data-testid="stFileUploaderDropzone"]{ border:2px dashed rgba(0,0,0,.25); border-radius:14px; padding:26px;
|
| 28 |
+
transition:border-color .2s, background .2s; }
|
| 29 |
+
div[data-testid="stFileUploaderDropzone"]:hover{ border-color:#4f8bf9; background:rgba(79,139,249,0.06); }
|
| 30 |
+
.hint{ color:#8a8a8a; font-size:13px; margin-top:4px; }
|
| 31 |
+
.file-chip{ display:inline-flex; align-items:center; gap:8px; background:rgba(79,139,249,0.12); color:#1f4fbf;
|
| 32 |
+
border-radius:999px; padding:6px 12px; font-size:13px; }
|
| 33 |
+
.kpi { display:grid; grid-template-columns: repeat(3, minmax(0,1fr)); gap:12px; margin:8px 0 4px 0; }
|
| 34 |
+
.kpi .card{ border:1px solid rgba(0,0,0,.08); border-radius:12px; padding:14px 16px; background:var(--secondary-background-color); }
|
| 35 |
+
.kpi .title{ font-size:12px; color:#666; }
|
| 36 |
+
.kpi .value{ font-size:20px; font-weight:700; margin-top:2px; }
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 37 |
</style>
|
| 38 |
""", unsafe_allow_html=True)
|
| 39 |
|
| 40 |
# ---------- Title ----------
|
| 41 |
st.markdown("<h1 style='margin:0 0 10px 0;'>Quiz Generation</h1>", unsafe_allow_html=True)
|
| 42 |
+
st.caption("อัปโหลดสไลด์/ไฟล์ข้อความ → ตั้งค่า → กดปุ่มเพื่อสร้างแบบทดสอบอัตโนมัติ")
|
| 43 |
|
| 44 |
+
# ---------- Upload ----------
|
| 45 |
st.markdown('<div class="upload-title">📘 Upload File</div>', unsafe_allow_html=True)
|
|
|
|
|
|
|
| 46 |
uploaded = st.file_uploader(
|
| 47 |
"ลากไฟล์มาวาง หรือกดเพื่อเลือกไฟล์",
|
| 48 |
type=["txt", "pdf", "docx", "csv"],
|
| 49 |
label_visibility="collapsed",
|
| 50 |
accept_multiple_files=False
|
| 51 |
)
|
|
|
|
| 52 |
st.markdown('<div class="hint">รองรับ .txt .pdf .docx .csv • ขนาดไม่ควรเกิน ~50MB</div>', unsafe_allow_html=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
+
text = ""
|
| 55 |
+
if uploaded:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 56 |
st.markdown(f'<div class="file-chip">📄 {uploaded.name}</div>', unsafe_allow_html=True)
|
| 57 |
st.divider()
|
| 58 |
+
uploaded.seek(0)
|
| 59 |
+
text = load_to_text(uploaded)
|
| 60 |
+
|
| 61 |
+
st.subheader("📄 Preview")
|
| 62 |
+
st.text_area("เนื้อหาไฟล์ (ตัวอย่าง)", text[:5000] if text else "(ไฟล์ไม่มีข้อความอ่านได้)", height=260, label_visibility="collapsed")
|
| 63 |
+
if not text:
|
| 64 |
+
st.info("ไม่พบข้อความที่อ่านได้จากไฟล์ที่อัปโหลด (อาจเป็นสแกนภาพหรือมีการป้องกัน).")
|
| 65 |
+
|
| 66 |
+
# ---------- Options ----------
|
| 67 |
+
st.divider()
|
| 68 |
+
st.subheader("⚙️ Quiz Options")
|
| 69 |
+
col1, col2, col3 = st.columns([1, 1, 2])
|
| 70 |
+
n_questions = col1.slider("จำนวนข้อ", 5, 50, 10)
|
| 71 |
+
q_types = col2.multiselect("ประเภทคำถาม", ["CLOZE", "T/F", "MCQ"], default=["CLOZE", "T/F", "MCQ"])
|
| 72 |
+
seed_value = col2.number_input("Seed (กำหนดให้ผลเหมือนเดิม)", min_value=0, value=42, step=1)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
|
| 74 |
# ---------- Action ----------
|
| 75 |
+
generate_btn = st.button("🚀 สร้าง Quiz", use_container_width=True, disabled=not uploaded)
|
| 76 |
+
|
| 77 |
+
# ---------- Results Area (ใต้ปุ่ม) ----------
|
| 78 |
+
if generate_btn:
|
| 79 |
+
if not uploaded:
|
| 80 |
+
st.warning("กรุณาอัปโหลดไฟล์ก่อน")
|
| 81 |
+
elif not text:
|
| 82 |
+
st.error("ไม่พบข้อความสำหรับสร้างแบบทดสอบ")
|
| 83 |
+
else:
|
| 84 |
+
with st.spinner("กำลังสร้างคำถาม..."):
|
| 85 |
+
quiz = generate_quiz(text, n_questions, mix=tuple(q_types), seed=seed_value)
|
| 86 |
+
rows = to_csv_rows(quiz)
|
| 87 |
+
df = pd.DataFrame(rows)
|
| 88 |
+
|
| 89 |
+
# KPIs
|
| 90 |
+
counts = df["Type"].value_counts().to_dict() if not df.empty else {}
|
| 91 |
+
k_cloze = counts.get("CLOZE", 0)
|
| 92 |
+
k_tf = counts.get("T/F", 0)
|
| 93 |
+
k_mcq = counts.get("MCQ", 0)
|
| 94 |
+
|
| 95 |
+
st.markdown('<div class="kpi">', unsafe_allow_html=True)
|
| 96 |
+
st.markdown(f'<div class="card"><div class="title">รวมข้อ</div><div class="value">{len(df)}</div></div>', unsafe_allow_html=True)
|
| 97 |
+
st.markdown(f'<div class="card"><div class="title">CLOZE</div><div class="value">{k_cloze}</div></div>', unsafe_allow_html=True)
|
| 98 |
+
st.markdown(f'<div class="card"><div class="title">T/F • MCQ</div><div class="value">{k_tf} • {k_mcq}</div></div>', unsafe_allow_html=True)
|
| 99 |
+
st.markdown('</div>', unsafe_allow_html=True)
|
| 100 |
+
|
| 101 |
+
# Tabs: ตาราง / CSV / JSON
|
| 102 |
+
tab_table, tab_csv, tab_json = st.tabs(["📋 ตารางคำถาม", "🧾 พรีวิว CSV", "🧩 JSON"])
|
| 103 |
+
|
| 104 |
+
with tab_table:
|
| 105 |
+
st.dataframe(df, use_container_width=True)
|
| 106 |
+
|
| 107 |
+
with tab_csv:
|
| 108 |
+
csv_buf = io.StringIO()
|
| 109 |
+
# ใช้ ; เพื่อให้ Excel (locale ไทย) แยกคอลัมน์อัตโนมัติ + UTF-8 BOM
|
| 110 |
+
df.to_csv(csv_buf, index=False, sep=';', quoting=csv.QUOTE_ALL)
|
| 111 |
+
csv_bytes = csv_buf.getvalue().encode("utf-8-sig")
|
| 112 |
+
st.code(csv_buf.getvalue()[:1000] + ("\n... (ตัดท่อนี้ในการแสดงตัวอย่าง)" if len(csv_buf.getvalue()) > 1000 else ""), language="csv")
|
| 113 |
+
st.download_button("⬇️ ดาวน์โหลด CSV", csv_bytes, file_name="quiz_generated.csv", mime="text/csv", use_container_width=True)
|
| 114 |
+
|
| 115 |
+
with tab_json:
|
| 116 |
+
json_str = json.dumps(rows, ensure_ascii=False, indent=2)
|
| 117 |
+
st.code(json_str[:2000] + ("\n... (ตัดท่อนี้ในการแสดงตัวอย่าง)" if len(json_str) > 2000 else ""), language="json")
|
| 118 |
+
st.download_button(
|
| 119 |
+
"⬇️ ดาวน์โหลด JSON",
|
| 120 |
+
json_str.encode("utf-8"),
|
| 121 |
+
file_name="quiz_generated.json",
|
| 122 |
+
mime="application/json",
|
| 123 |
+
use_container_width=True
|
| 124 |
+
)
|
src/utils/__pycache__/loader.cpython-313.pyc
ADDED
|
Binary file (7.43 kB). View file
|
|
|
src/utils/__pycache__/qg_baseline.cpython-313.pyc
ADDED
|
Binary file (16.6 kB). View file
|
|
|
src/utils/extractors.py
DELETED
|
@@ -1,12 +0,0 @@
|
|
| 1 |
-
from pypdf import PdfReader
|
| 2 |
-
|
| 3 |
-
def extract_text_from_pdf(file_path):
|
| 4 |
-
reader = PdfReader(file_path)
|
| 5 |
-
all_text = ""
|
| 6 |
-
|
| 7 |
-
for i, page in enumerate(reader.pages):
|
| 8 |
-
text = page.extract_text()
|
| 9 |
-
if text: # ถ้ามีข้อความในหน้านั้น
|
| 10 |
-
all_text += f"\n\n--- Page {i+1} ---\n{text}"
|
| 11 |
-
|
| 12 |
-
return all_text.strip()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
src/utils/loader.py
ADDED
|
@@ -0,0 +1,137 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# src/utils/loader.py
|
| 2 |
+
from pypdf import PdfReader
|
| 3 |
+
from docx import Document
|
| 4 |
+
import pandas as pd
|
| 5 |
+
import io
|
| 6 |
+
import re
|
| 7 |
+
import os
|
| 8 |
+
|
| 9 |
+
# ---------------------------------------------
|
| 10 |
+
# 🧩 1. อ่านไฟล์ PDF
|
| 11 |
+
# ---------------------------------------------
|
| 12 |
+
def extract_text_from_pdf(file_path: str) -> str:
|
| 13 |
+
"""
|
| 14 |
+
ใช้เมื่อมี path ของไฟล์ PDF จริง (เช่นตอนรันบนเครื่อง)
|
| 15 |
+
"""
|
| 16 |
+
reader = PdfReader(file_path)
|
| 17 |
+
all_text = ""
|
| 18 |
+
for i, page in enumerate(reader.pages):
|
| 19 |
+
text = page.extract_text()
|
| 20 |
+
if text:
|
| 21 |
+
all_text += f"\n\n--- Page {i+1} ---\n{text}"
|
| 22 |
+
return all_text.strip()
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
def extract_text_from_pdf_bytes(file_bytes: bytes) -> str:
|
| 26 |
+
"""
|
| 27 |
+
ใช้เมื่อไฟล์มาจาก Streamlit uploader (เป็น bytes)
|
| 28 |
+
"""
|
| 29 |
+
reader = PdfReader(io.BytesIO(file_bytes))
|
| 30 |
+
texts = []
|
| 31 |
+
for i, page in enumerate(reader.pages):
|
| 32 |
+
text = page.extract_text()
|
| 33 |
+
if text:
|
| 34 |
+
texts.append(f"--- Page {i+1} ---\n{text}")
|
| 35 |
+
return "\n\n".join(texts).strip()
|
| 36 |
+
|
| 37 |
+
# ---------------------------------------------
|
| 38 |
+
# 🧩 2. อ่านไฟล์ DOCX
|
| 39 |
+
# ---------------------------------------------
|
| 40 |
+
def extract_text_from_docx_bytes(file_bytes: bytes) -> str:
|
| 41 |
+
doc = Document(io.BytesIO(file_bytes))
|
| 42 |
+
paras = [p.text for p in doc.paragraphs if p.text.strip()]
|
| 43 |
+
return "\n".join(paras).strip()
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
def extract_text_from_docx_path(file_path: str) -> str:
|
| 47 |
+
doc = Document(file_path)
|
| 48 |
+
paras = [p.text for p in doc.paragraphs if p.text.strip()]
|
| 49 |
+
return "\n".join(paras).strip()
|
| 50 |
+
|
| 51 |
+
# ---------------------------------------------
|
| 52 |
+
# 🧩 3. อ่านไฟล์ TXT / CSV
|
| 53 |
+
# ---------------------------------------------
|
| 54 |
+
def extract_text_from_txt_bytes(file_bytes: bytes, encoding="utf-8") -> str:
|
| 55 |
+
try:
|
| 56 |
+
return file_bytes.decode(encoding, errors="ignore")
|
| 57 |
+
except Exception:
|
| 58 |
+
return file_bytes.decode("utf-8", errors="ignore")
|
| 59 |
+
|
| 60 |
+
|
| 61 |
+
def extract_text_from_txt_path(file_path: str, encoding="utf-8") -> str:
|
| 62 |
+
with open(file_path, "r", encoding=encoding, errors="ignore") as f:
|
| 63 |
+
return f.read()
|
| 64 |
+
|
| 65 |
+
|
| 66 |
+
def extract_text_from_csv_bytes(file_bytes: bytes) -> str:
|
| 67 |
+
df = pd.read_csv(io.BytesIO(file_bytes))
|
| 68 |
+
# รวมทุกคอลัมน์ทุกแถวให้เป็นข้อความเดียว (เพื่อสร้าง quiz ภายหลัง)
|
| 69 |
+
lines = []
|
| 70 |
+
for _, row in df.astype(str).iterrows():
|
| 71 |
+
lines.append(" | ".join(row.values))
|
| 72 |
+
return "\n".join(lines)
|
| 73 |
+
|
| 74 |
+
|
| 75 |
+
def extract_text_from_csv_path(file_path: str) -> str:
|
| 76 |
+
df = pd.read_csv(file_path)
|
| 77 |
+
lines = []
|
| 78 |
+
for _, row in df.astype(str).iterrows():
|
| 79 |
+
lines.append(" | ".join(row.values))
|
| 80 |
+
return "\n".join(lines)
|
| 81 |
+
|
| 82 |
+
# ---------------------------------------------
|
| 83 |
+
# 🧩 4. Clean ข้อความ (ตัดช่องว่าง / จัดรูปแบบ)
|
| 84 |
+
# ---------------------------------------------
|
| 85 |
+
def clean_text(text: str) -> str:
|
| 86 |
+
text = re.sub(r'\s+\n', '\n', text) # เว้นวรรคก่อนขึ้นบรรทัด
|
| 87 |
+
text = re.sub(r'\n{3,}', '\n\n', text) # ตัดบรรทัดว่างเกิน 2 บรรทัด
|
| 88 |
+
text = re.sub(r'[ \t]{2,}', ' ', text) # ตัดช่องว่างยาว
|
| 89 |
+
return text.strip()
|
| 90 |
+
|
| 91 |
+
# ---------------------------------------------
|
| 92 |
+
# 🧩 5. ฟังก์ชันรวมหลัก (Auto-detect)
|
| 93 |
+
# ---------------------------------------------
|
| 94 |
+
def load_to_text(uploaded_file) -> str:
|
| 95 |
+
"""
|
| 96 |
+
ใช้ใน Streamlit: รับ uploaded_file แล้วคืนข้อความทั้งหมดเป็น string เดียว
|
| 97 |
+
"""
|
| 98 |
+
file_bytes = uploaded_file.read()
|
| 99 |
+
name = uploaded_file.name.lower()
|
| 100 |
+
mime = uploaded_file.type or ""
|
| 101 |
+
|
| 102 |
+
if name.endswith(".pdf"):
|
| 103 |
+
text = extract_text_from_pdf_bytes(file_bytes)
|
| 104 |
+
elif name.endswith(".docx"):
|
| 105 |
+
text = extract_text_from_docx_bytes(file_bytes)
|
| 106 |
+
elif name.endswith(".csv"):
|
| 107 |
+
text = extract_text_from_csv_bytes(file_bytes)
|
| 108 |
+
elif name.endswith(".txt") or mime.startswith("text/"):
|
| 109 |
+
text = extract_text_from_txt_bytes(file_bytes)
|
| 110 |
+
else:
|
| 111 |
+
# fallback พยายามอ่านเป็นข้อความ
|
| 112 |
+
try:
|
| 113 |
+
text = extract_text_from_txt_bytes(file_bytes)
|
| 114 |
+
except Exception:
|
| 115 |
+
text = ""
|
| 116 |
+
return clean_text(text)
|
| 117 |
+
|
| 118 |
+
|
| 119 |
+
def load_from_path(file_path: str) -> str:
|
| 120 |
+
"""
|
| 121 |
+
ใช้นอก Streamlit (อ่านจาก path ปกติ)
|
| 122 |
+
"""
|
| 123 |
+
file_path = str(file_path)
|
| 124 |
+
if not os.path.exists(file_path):
|
| 125 |
+
raise FileNotFoundError(file_path)
|
| 126 |
+
|
| 127 |
+
if file_path.lower().endswith(".pdf"):
|
| 128 |
+
text = extract_text_from_pdf(file_path)
|
| 129 |
+
elif file_path.lower().endswith(".docx"):
|
| 130 |
+
text = extract_text_from_docx_path(file_path)
|
| 131 |
+
elif file_path.lower().endswith(".csv"):
|
| 132 |
+
text = extract_text_from_csv_path(file_path)
|
| 133 |
+
elif file_path.lower().endswith(".txt"):
|
| 134 |
+
text = extract_text_from_txt_path(file_path)
|
| 135 |
+
else:
|
| 136 |
+
text = ""
|
| 137 |
+
return clean_text(text)
|
src/utils/qg_baseline.py
ADDED
|
@@ -0,0 +1,307 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# src/utils/qg_baseline.py
|
| 2 |
+
# -*- coding: utf-8 -*-
|
| 3 |
+
"""
|
| 4 |
+
Question Generation (Baseline) — Thai/EN friendly, seed-able, and with simple QC.
|
| 5 |
+
|
| 6 |
+
Features:
|
| 7 |
+
- split_blocks() : แยกเนื้อหาเป็นบล็อกคล้ายสไลด์ (รองรับ --- Page, หัวข้อ, บรรทัดว่าง)
|
| 8 |
+
- to_sentences() : แยกประโยคแบบ heuristic (ไทย/อังกฤษ), รวม bullet
|
| 9 |
+
- pick_keywords(): เลือกคีย์เวิร์ดสำคัญ (กรอง stopwords ไทย/อังกฤษแบบหยาบ)
|
| 10 |
+
- make_cloze(): เติมคำ (CLOZE) พร้อมตัวลวง 3 ตัว (รวมเป็น 4 ตัวเลือกเสมอ)
|
| 11 |
+
- make_true_false(): กลับค่าตัวเลข/ปี (ถ้ามี) ไม่งั้นใช้ True
|
| 12 |
+
- make_mcq_from_bullets(): “ข้อใดไม่ใช่หัวข้อย่อยในสไลด์นี้?”
|
| 13 |
+
- generate_quiz(): ประกอบทั้งหมด + QC (กรองซ้ำ, จำกัดความยาว, ครอบจำนวนข้อ)
|
| 14 |
+
- to_csv_rows(): ส่งออกในรูปแถวสำหรับ CSV
|
| 15 |
+
|
| 16 |
+
ใช้ random.Random(seed) ภายใน (ไม่ยุ่งกับ global random)
|
| 17 |
+
"""
|
| 18 |
+
|
| 19 |
+
from __future__ import annotations
|
| 20 |
+
import re
|
| 21 |
+
import random
|
| 22 |
+
from typing import List, Dict, Optional
|
| 23 |
+
from collections import Counter
|
| 24 |
+
|
| 25 |
+
|
| 26 |
+
# -----------------------------
|
| 27 |
+
# Utilities
|
| 28 |
+
# -----------------------------
|
| 29 |
+
def _rng(seed: Optional[int] = None) -> random.Random:
|
| 30 |
+
"""คืนอ็อบเจ็กต์สุ่มเฉพาะตัว (isolated RNG)"""
|
| 31 |
+
return random.Random(seed) if seed is not None else random
|
| 32 |
+
|
| 33 |
+
_TH_STOP = set("""
|
| 34 |
+
การ ของ และ ที่ ใน เป็น ได้ ให้ หรือ จาก ด้วย มี ไป มา ซึ่ง เช่น ค่า ตัว ข้อ ความ จะ ถูก คือ
|
| 35 |
+
ว่า ให้ โดย เมื่อ หาก เพราะ แต่ ยัง จึง กับ อย่าง เช่นนี้ เช่นนั้น อีก ทั้ง นี้ นั้น นี่ เลย ก็ ก็ยัง ยังไง
|
| 36 |
+
ถึง เพราะว่า ดังนั้น แล้ว โดยที่ โดยเฉพาะ โดยรวม ประมาณ อาจ จะ ได้ ไม่ ใช่ ก็ได้ กัน บ้าง จน กระทั่ง ต่อ
|
| 37 |
+
ระหว่าง สำหรับ ตาม ด้วยกัน เท่านั้น เท่านี้ นอกจากนี้ ทั้งหมด บน ใต้ หน้า หลัง ก่อน หลังจาก ต่อไป
|
| 38 |
+
""".split())
|
| 39 |
+
|
| 40 |
+
_EN_STOP = set("""
|
| 41 |
+
the of and to in is are was were be been being a an for with on at by from this that these those as it its
|
| 42 |
+
or if then than not no yes do does did have has had can could should would may might you your we our they them
|
| 43 |
+
itself themselves ourselves themselves he she his her theirs their i me my mine
|
| 44 |
+
""".split())
|
| 45 |
+
|
| 46 |
+
def _normalize_whitespace(s: str) -> str:
|
| 47 |
+
s = re.sub(r'[ \t]+', ' ', s)
|
| 48 |
+
s = re.sub(r'\s*\n\s*', '\n', s)
|
| 49 |
+
return s.strip()
|
| 50 |
+
|
| 51 |
+
def _len_ok(s: str, min_len=12, max_len=280) -> bool:
|
| 52 |
+
L = len(s or "")
|
| 53 |
+
return min_len <= L <= max_len
|
| 54 |
+
|
| 55 |
+
def _uniq_keep_order(items: List[str]) -> List[str]:
|
| 56 |
+
seen = set()
|
| 57 |
+
out = []
|
| 58 |
+
for x in items:
|
| 59 |
+
if x not in seen:
|
| 60 |
+
out.append(x)
|
| 61 |
+
seen.add(x)
|
| 62 |
+
return out
|
| 63 |
+
|
| 64 |
+
|
| 65 |
+
# -----------------------------
|
| 66 |
+
# Block & sentence splitting
|
| 67 |
+
# -----------------------------
|
| 68 |
+
def split_blocks(text: str) -> List[str]:
|
| 69 |
+
"""
|
| 70 |
+
แบ่งเป็นบล็อกโดยดูจาก:
|
| 71 |
+
- ตัวแบ่งหน้าแบบ --- Page X ---
|
| 72 |
+
- หัวข้อ/บรรทัดว่างติดกัน
|
| 73 |
+
"""
|
| 74 |
+
if not text:
|
| 75 |
+
return []
|
| 76 |
+
# ตัดด้วย "--- Page n ---" ก่อน
|
| 77 |
+
parts = re.split(r'(?im)^---\s*Page\s*\d+\s*---\s*$', text)
|
| 78 |
+
# ถ้าไม่มี page marker ใช้บรรทัดว่างคู่/สาม
|
| 79 |
+
if len(parts) == 1:
|
| 80 |
+
parts = re.split(r'\n{2,}', text)
|
| 81 |
+
|
| 82 |
+
blocks = []
|
| 83 |
+
for p in parts:
|
| 84 |
+
p = _normalize_whitespace(p)
|
| 85 |
+
if p and len(p) > 40:
|
| 86 |
+
blocks.append(p)
|
| 87 |
+
return blocks or [text.strip()]
|
| 88 |
+
|
| 89 |
+
|
| 90 |
+
def to_sentences(block: str) -> List[str]:
|
| 91 |
+
"""
|
| 92 |
+
แยกประโยคไทย/อังกฤษแบบง่าย:
|
| 93 |
+
- ตัดตาม . ! ? หรือขึ้นบรรทัดใหม่
|
| 94 |
+
- รวม bullet • -
|
| 95 |
+
- กรองสั้น/ยาวเกินไป
|
| 96 |
+
"""
|
| 97 |
+
lines = [ln.strip() for ln in block.splitlines() if ln.strip()]
|
| 98 |
+
# รวม bullet ให้เป็นประโยคเดียว (ช่วย MCQ ภายหลัง)
|
| 99 |
+
bullets = [re.sub(r'^[•\-\*\u2022]\s*', '', ln).strip() for ln in lines if re.match(r'^[•\-\*\u2022]\s*', ln)]
|
| 100 |
+
# แยกประโยคทั่วไป
|
| 101 |
+
raw = re.split(r'(?<=[.!?])\s+|\n+', block)
|
| 102 |
+
sents = [s.strip('•-* \t') for s in raw if _len_ok(s.strip())]
|
| 103 |
+
|
| 104 |
+
# ถ้าไม่มีประโยคเลย เอา bullet มาช่วย
|
| 105 |
+
if not sents and bullets:
|
| 106 |
+
sents = [b for b in bullets if _len_ok(b)]
|
| 107 |
+
|
| 108 |
+
return _uniq_keep_order(sents)
|
| 109 |
+
|
| 110 |
+
|
| 111 |
+
# -----------------------------
|
| 112 |
+
# Keyword extraction (simple)
|
| 113 |
+
# -----------------------------
|
| 114 |
+
def pick_keywords(text: str, topk=30) -> List[str]:
|
| 115 |
+
tokens = re.findall(r'[A-Za-zก-๙0-9\-_/]+', text or "")
|
| 116 |
+
cnt = Counter(w.lower() for w in tokens if len(w) > 2)
|
| 117 |
+
cand = []
|
| 118 |
+
for w, _ in cnt.most_common():
|
| 119 |
+
if w in _TH_STOP or w in _EN_STOP:
|
| 120 |
+
continue
|
| 121 |
+
# กรองที่เป็นตัวเลขยาวเกิน (โทรศัพท์/โค้ดยืด)
|
| 122 |
+
if re.fullmatch(r'\d{1,6}', w) or re.search(r'[A-Za-zก-๙]', w):
|
| 123 |
+
cand.append(w)
|
| 124 |
+
return cand[:topk]
|
| 125 |
+
|
| 126 |
+
|
| 127 |
+
# -----------------------------
|
| 128 |
+
# Question builders
|
| 129 |
+
# -----------------------------
|
| 130 |
+
def _pad_options(opts: List[str], n=4, pad='—') -> List[str]:
|
| 131 |
+
opts = [o if o is not None else '' for o in opts]
|
| 132 |
+
while len(opts) < n:
|
| 133 |
+
opts.append(pad)
|
| 134 |
+
return opts[:n]
|
| 135 |
+
|
| 136 |
+
def _dedup_options(opts: List[str]) -> List[str]:
|
| 137 |
+
# ลบซ้ำแบบไม่แคร์เคส
|
| 138 |
+
seen = set()
|
| 139 |
+
out = []
|
| 140 |
+
for o in opts:
|
| 141 |
+
key = o.strip().lower()
|
| 142 |
+
if key not in seen and key != '':
|
| 143 |
+
out.append(o)
|
| 144 |
+
seen.add(key)
|
| 145 |
+
return out
|
| 146 |
+
|
| 147 |
+
def make_cloze(sentence: str, keywords: List[str], rng: random.Random) -> Dict:
|
| 148 |
+
# เลือกคำที่อยู่ในประโยคจริง ๆ
|
| 149 |
+
present = [k for k in keywords if re.search(rf'(?i)\b{re.escape(k)}\b', sentence)]
|
| 150 |
+
if not present:
|
| 151 |
+
return {}
|
| 152 |
+
answer = rng.choice(present)
|
| 153 |
+
# ปิดคำแบบ case-insensitive
|
| 154 |
+
q = re.sub(rf'(?i)\b{re.escape(answer)}\b', '____', sentence)
|
| 155 |
+
# ตัวลวง: คำอื่นที่มี type ใกล้เคียง (เลขกับเลข / คำกับคำ)
|
| 156 |
+
is_num = bool(re.fullmatch(r'\d+', answer))
|
| 157 |
+
pool = [k for k in keywords if k.lower() != answer.lower() and (bool(re.fullmatch(r'\d+', k)) == is_num)]
|
| 158 |
+
rng.shuffle(pool)
|
| 159 |
+
distractors = _dedup_options(pool)[:3]
|
| 160 |
+
options = [answer] + distractors
|
| 161 |
+
rng.shuffle(options)
|
| 162 |
+
options = _pad_options(options)
|
| 163 |
+
return {
|
| 164 |
+
"type": "CLOZE",
|
| 165 |
+
"question": q,
|
| 166 |
+
"options": options,
|
| 167 |
+
"answer": answer,
|
| 168 |
+
"explanation": ""
|
| 169 |
+
}
|
| 170 |
+
|
| 171 |
+
def make_true_false(sentence: str, rng: random.Random) -> Dict:
|
| 172 |
+
# ถ้ามีตัวเลข ให้กลับค่า (เช่น ปี/จำนวน)
|
| 173 |
+
nums = re.findall(r'\b\d+\b', sentence)
|
| 174 |
+
if nums:
|
| 175 |
+
n = rng.choice(nums)
|
| 176 |
+
# เลือก offset เล็กน้อยเพื่อไม่ให้หลุดกรอบ
|
| 177 |
+
offset = rng.choice([1, 2, 3, 5, 10])
|
| 178 |
+
fake = str(int(n) + offset)
|
| 179 |
+
q = re.sub(rf'\b{re.escape(n)}\b', fake, sentence, count=1)
|
| 180 |
+
return {"type": "T/F", "question": q, "options": ["True", "False"], "answer": "False",
|
| 181 |
+
"explanation": f"ค่าจริงควรเป็น {n}"}
|
| 182 |
+
# ไม่มีตัวเลข → ใช้เป็น True
|
| 183 |
+
return {"type": "T/F", "question": sentence, "options": ["True", "False"], "answer": "True", "explanation": ""}
|
| 184 |
+
|
| 185 |
+
def make_mcq_from_bullets(block: str, keywords: List[str], rng: random.Random) -> List[Dict]:
|
| 186 |
+
# รวบรวม bullet ในบล็อกนี้
|
| 187 |
+
bullets = [re.sub(r'^[•\-\*\u2022]\s*', '', ln).strip()
|
| 188 |
+
for ln in block.splitlines()
|
| 189 |
+
if re.match(r'^[•\-\*\u2022]\s*', ln)]
|
| 190 |
+
out: List[Dict] = []
|
| 191 |
+
if len(bullets) >= 4:
|
| 192 |
+
# “ข้อใดไม่ใช่หัวข้อย่อยในสไลด์นี้?”
|
| 193 |
+
# ตัวลวงมาจาก keywords ที่ไม่ใช่ bullet
|
| 194 |
+
pool = [w for w in pick_keywords(" ".join(keywords), topk=80) if w not in bullets]
|
| 195 |
+
rng.shuffle(pool)
|
| 196 |
+
wrong = pool[:1] or ["(ไม่มี)"]
|
| 197 |
+
# เลือกถูก 3 ข้อจาก bullets
|
| 198 |
+
right = bullets[:3]
|
| 199 |
+
options = right + wrong
|
| 200 |
+
rng.shuffle(options)
|
| 201 |
+
ans = next((o for o in options if o not in bullets), options[0])
|
| 202 |
+
options = _pad_options(options)
|
| 203 |
+
out.append({
|
| 204 |
+
"type": "MCQ",
|
| 205 |
+
"question": "ต่อไปนี้ข้อใด 'ไม่ใช่' หัวข้อย่อยในสไลด์นี้?",
|
| 206 |
+
"options": options,
|
| 207 |
+
"answer": ans,
|
| 208 |
+
"explanation": ""
|
| 209 |
+
})
|
| 210 |
+
return out
|
| 211 |
+
|
| 212 |
+
|
| 213 |
+
# -----------------------------
|
| 214 |
+
# Main generator + QC
|
| 215 |
+
# -----------------------------
|
| 216 |
+
def _question_fingerprint(q: Dict) -> str:
|
| 217 |
+
base = f"{q.get('type','')}|{q.get('question','')}".strip().lower()
|
| 218 |
+
base = re.sub(r'\s+', ' ', base)
|
| 219 |
+
return base
|
| 220 |
+
|
| 221 |
+
def _valid_question(q: Dict) -> bool:
|
| 222 |
+
if not q: return False
|
| 223 |
+
if not _len_ok(q.get("question","")): return False
|
| 224 |
+
opts = q.get("options", [])
|
| 225 |
+
if q["type"] == "T/F":
|
| 226 |
+
return True
|
| 227 |
+
# อย่างน้อย 2 ตัวเลือกไม่ซ้ำ/ไม่ว่าง
|
| 228 |
+
clean = _dedup_options(opts)
|
| 229 |
+
return len(clean) >= 2 and q.get("answer","") != ""
|
| 230 |
+
|
| 231 |
+
def generate_quiz(
|
| 232 |
+
text: str,
|
| 233 |
+
n_questions: int = 10,
|
| 234 |
+
mix: tuple[str, ...] = ("CLOZE", "T/F", "MCQ"),
|
| 235 |
+
seed: Optional[int] = None
|
| 236 |
+
) -> List[Dict]:
|
| 237 |
+
"""
|
| 238 |
+
สร้างชุดคำถามจากข้อความ
|
| 239 |
+
- seed: ตั้งค่าเพื่อให้ผลเดิมซ้ำได้
|
| 240 |
+
"""
|
| 241 |
+
rng = _rng(seed)
|
| 242 |
+
blocks = split_blocks(text) or [text]
|
| 243 |
+
global_keywords = pick_keywords(text, topk=80)
|
| 244 |
+
|
| 245 |
+
questions: List[Dict] = []
|
| 246 |
+
seen_fp = set()
|
| 247 |
+
|
| 248 |
+
for b in blocks:
|
| 249 |
+
if len(questions) >= n_questions:
|
| 250 |
+
break
|
| 251 |
+
|
| 252 |
+
sents = to_sentences(b)
|
| 253 |
+
local_kw = _uniq_keep_order(pick_keywords(b, topk=30) + global_keywords)[:80]
|
| 254 |
+
|
| 255 |
+
# MCQ จาก bullets
|
| 256 |
+
if "MCQ" in mix:
|
| 257 |
+
for q in make_mcq_from_bullets(b, local_kw, rng):
|
| 258 |
+
fp = _question_fingerprint(q)
|
| 259 |
+
if fp not in seen_fp and _valid_question(q):
|
| 260 |
+
questions.append(q)
|
| 261 |
+
seen_fp.add(fp)
|
| 262 |
+
if len(questions) >= n_questions: break
|
| 263 |
+
|
| 264 |
+
# สร้าง Cloze + T/F จากประโยค (สุ่มเลือกบางประโยค)
|
| 265 |
+
rng.shuffle(sents)
|
| 266 |
+
for s in sents:
|
| 267 |
+
if len(questions) >= n_questions: break
|
| 268 |
+
|
| 269 |
+
if "CLOZE" in mix and rng.random() < 0.7:
|
| 270 |
+
q = make_cloze(s, local_kw, rng)
|
| 271 |
+
if q and _valid_question(q):
|
| 272 |
+
fp = _question_fingerprint(q)
|
| 273 |
+
if fp not in seen_fp:
|
| 274 |
+
questions.append(q); seen_fp.add(fp)
|
| 275 |
+
if len(questions) >= n_questions: break
|
| 276 |
+
|
| 277 |
+
if "T/F" in mix and len(questions) < n_questions:
|
| 278 |
+
q = make_true_false(s, rng)
|
| 279 |
+
if _valid_question(q):
|
| 280 |
+
fp = _question_fingerprint(q)
|
| 281 |
+
if fp not in seen_fp:
|
| 282 |
+
questions.append(q); seen_fp.add(fp)
|
| 283 |
+
|
| 284 |
+
# ตัดให้พอดีจำนวน
|
| 285 |
+
return questions[:n_questions]
|
| 286 |
+
|
| 287 |
+
|
| 288 |
+
# -----------------------------
|
| 289 |
+
# Export helpers
|
| 290 |
+
# -----------------------------
|
| 291 |
+
def to_csv_rows(quiz: List[Dict]) -> List[Dict]:
|
| 292 |
+
rows = []
|
| 293 |
+
for i, q in enumerate(quiz, 1):
|
| 294 |
+
opts = q.get("options", [])
|
| 295 |
+
opts = _pad_options(opts)
|
| 296 |
+
rows.append({
|
| 297 |
+
"No": i,
|
| 298 |
+
"Type": q.get("type",""),
|
| 299 |
+
"Question": q.get("question",""),
|
| 300 |
+
"OptionA": opts[0],
|
| 301 |
+
"OptionB": opts[1],
|
| 302 |
+
"OptionC": opts[2],
|
| 303 |
+
"OptionD": opts[3],
|
| 304 |
+
"Answer": q.get("answer",""),
|
| 305 |
+
"Explanation": q.get("explanation",""),
|
| 306 |
+
})
|
| 307 |
+
return rows
|