Spaces:
Sleeping
Sleeping
Update parser/assembler.py
Browse files- parser/assembler.py +71 -89
parser/assembler.py
CHANGED
|
@@ -1,116 +1,98 @@
|
|
| 1 |
from typing import List, Dict
|
| 2 |
from .section_extractor import extract_sections
|
| 3 |
from helpers.cleaner import clean_text
|
| 4 |
-
from helpers.utils import normalize_digits, is_article, extract_article_number
|
| 5 |
|
| 6 |
|
| 7 |
-
def
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
"""
|
| 9 |
-
ت
|
| 10 |
-
واستخراج المواد بشكل صحيح باستخدام الأنماط الحديثة من utils.
|
| 11 |
"""
|
| 12 |
-
|
| 13 |
-
# -----------------------------------
|
| 14 |
-
# 1) تنظيف النصوص الخام + تحويل الأرقام
|
| 15 |
-
# -----------------------------------
|
| 16 |
-
pure_texts = [
|
| 17 |
-
clean_text(normalize_digits(block.get("text", "").strip()))
|
| 18 |
-
for block in text_blocks
|
| 19 |
-
if block.get("text")
|
| 20 |
-
]
|
| 21 |
-
|
| 22 |
-
# -----------------------------------
|
| 23 |
-
# 2) استخراج العنوان
|
| 24 |
-
# -----------------------------------
|
| 25 |
title = ""
|
| 26 |
-
|
| 27 |
-
|
| 28 |
-
|
| 29 |
-
|
| 30 |
-
if candidate.lower() != "html":
|
| 31 |
-
title = candidate
|
| 32 |
break
|
| 33 |
|
| 34 |
-
|
| 35 |
-
|
| 36 |
-
|
| 37 |
-
while pure_texts and not any(
|
| 38 |
-
keyword in pure_texts[0]
|
| 39 |
-
for keyword in ["الباب", "الفصل", "القسم"]
|
| 40 |
-
):
|
| 41 |
-
preamble_lines.append(pure_texts.pop(0))
|
| 42 |
|
| 43 |
preamble = "\n".join(preamble_lines)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 44 |
|
| 45 |
-
# -----------------------------------
|
| 46 |
-
# 4) استخراج الأقسام الخام
|
| 47 |
-
# -----------------------------------
|
| 48 |
-
sections_raw = extract_sections(pure_texts)
|
| 49 |
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
sections = []
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 54 |
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
# -------- المحتوى غير المادة --------
|
| 59 |
-
content_lines = [t for t in raw_texts if not is_article(t)]
|
| 60 |
-
content = "\n".join(content_lines).strip()
|
| 61 |
-
|
| 62 |
-
# -------- استخراج المواد --------
|
| 63 |
-
articles = []
|
| 64 |
-
current_article = None
|
| 65 |
-
|
| 66 |
-
for t in raw_texts:
|
| 67 |
-
t = t.strip()
|
| 68 |
-
|
| 69 |
-
if is_article(t):
|
| 70 |
-
# إضافة المادة السابقة إن وجدت
|
| 71 |
-
if current_article:
|
| 72 |
-
current_article["text"] = current_article["text"].strip()
|
| 73 |
-
articles.append(current_article)
|
| 74 |
-
|
| 75 |
-
# بدء مادة جديدة
|
| 76 |
-
current_article = {
|
| 77 |
-
"number": extract_article_number(t),
|
| 78 |
-
"text": t,
|
| 79 |
-
}
|
| 80 |
-
else:
|
| 81 |
-
# إلحاق النص بالمادة الحالية
|
| 82 |
-
if current_article:
|
| 83 |
-
current_article["text"] += "\n" + t
|
| 84 |
-
|
| 85 |
-
# إضافة آخر مادة
|
| 86 |
-
if current_article:
|
| 87 |
-
current_article["text"] = current_article["text"].strip()
|
| 88 |
-
articles.append(current_article)
|
| 89 |
-
|
| 90 |
-
# -------- تنظيف بيانات القسم --------
|
| 91 |
-
clean_section = {
|
| 92 |
-
"title": clean_text(normalize_digits(s["name"])),
|
| 93 |
-
"content": clean_text(normalize_digits(content)),
|
| 94 |
"articles": [
|
| 95 |
-
{
|
| 96 |
-
"number": a["number"],
|
| 97 |
-
"text": clean_text(normalize_digits(a["text"]))
|
| 98 |
-
}
|
| 99 |
for a in articles
|
| 100 |
]
|
| 101 |
-
}
|
| 102 |
-
|
| 103 |
-
sections.append(clean_section)
|
| 104 |
|
| 105 |
-
#
|
| 106 |
-
# 6) إرجاع المستند القانوني الكامل
|
| 107 |
-
# -----------------------------------
|
| 108 |
return {
|
| 109 |
"message": "تم التحليل بنجاح",
|
| 110 |
"saved_to_db": False,
|
| 111 |
"law": {
|
| 112 |
-
"title":
|
| 113 |
-
"preamble":
|
| 114 |
"sections": sections
|
| 115 |
}
|
| 116 |
}
|
|
|
|
| 1 |
from typing import List, Dict
|
| 2 |
from .section_extractor import extract_sections
|
| 3 |
from helpers.cleaner import clean_text
|
| 4 |
+
from helpers.utils import normalize_digits, is_article, extract_article_number, is_section
|
| 5 |
|
| 6 |
|
| 7 |
+
def clean_text_block(text: str) -> str:
|
| 8 |
+
"""تنظيف النص وتحويل الأرقام الهندية إلى عربية."""
|
| 9 |
+
return clean_text(normalize_digits(text.strip()))
|
| 10 |
+
|
| 11 |
+
|
| 12 |
+
def extract_title_and_preamble(texts: List[str]) -> (str, str, List[str]):
|
| 13 |
"""
|
| 14 |
+
استخراج عنوان القانون والمقدمة، وإرجاع بقية النصوص بعد المقدمة.
|
|
|
|
| 15 |
"""
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
title = ""
|
| 17 |
+
while texts:
|
| 18 |
+
t = texts.pop(0)
|
| 19 |
+
if t.lower() != "html":
|
| 20 |
+
title = t
|
|
|
|
|
|
|
| 21 |
break
|
| 22 |
|
| 23 |
+
preamble_lines = []
|
| 24 |
+
while texts and not any(is_section(k) for k in [texts[0]]):
|
| 25 |
+
preamble_lines.append(texts.pop(0))
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
preamble = "\n".join(preamble_lines)
|
| 28 |
+
return title, preamble, texts
|
| 29 |
+
|
| 30 |
+
|
| 31 |
+
def extract_articles_from_texts(texts: List[str]) -> List[Dict]:
|
| 32 |
+
"""استخراج المواد من قائمة نصوص القسم باستخدام is_article و extract_article_number."""
|
| 33 |
+
articles = []
|
| 34 |
+
current = None
|
| 35 |
+
|
| 36 |
+
for t in texts:
|
| 37 |
+
t = t.strip()
|
| 38 |
+
if is_article(t):
|
| 39 |
+
if current:
|
| 40 |
+
current["text"] = current["text"].strip()
|
| 41 |
+
articles.append(current)
|
| 42 |
+
current = {"number": extract_article_number(t), "text": t}
|
| 43 |
+
else:
|
| 44 |
+
if current:
|
| 45 |
+
current["text"] += "\n" + t
|
| 46 |
+
|
| 47 |
+
if current:
|
| 48 |
+
current["text"] = current["text"].strip()
|
| 49 |
+
articles.append(current)
|
| 50 |
+
|
| 51 |
+
return articles
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
|
| 54 |
+
def parse_law_from_texts(text_blocks: List[Dict[str, str]]) -> Dict:
|
| 55 |
+
"""
|
| 56 |
+
تحويل نصوص القانون إلى هيكل JSON مع تنظيف واستخراج مواد وأقسام
|
| 57 |
+
بالاعتماد على دوال utils.py.
|
| 58 |
+
"""
|
| 59 |
+
|
| 60 |
+
# 1) تنظيف النصوص الخام
|
| 61 |
+
pure_texts = [clean_text_block(b.get("text", "")) for b in text_blocks if b.get("text")]
|
| 62 |
+
|
| 63 |
+
# 2) استخراج العنوان والمقدمة
|
| 64 |
+
title, preamble, remaining_texts = extract_title_and_preamble(pure_texts)
|
| 65 |
+
|
| 66 |
+
# 3) استخراج الأقسام الخام
|
| 67 |
+
sections_raw = extract_sections(remaining_texts)
|
| 68 |
+
|
| 69 |
+
# 4) بناء الأقسام مع المواد
|
| 70 |
sections = []
|
| 71 |
+
for sec in sections_raw:
|
| 72 |
+
raw_texts = sec["texts"]
|
| 73 |
+
|
| 74 |
+
# المحتوى غير المواد
|
| 75 |
+
content = "\n".join(t for t in raw_texts if not is_article(t)).strip()
|
| 76 |
+
|
| 77 |
+
# استخراج المواد
|
| 78 |
+
articles = extract_articles_from_texts(raw_texts)
|
| 79 |
|
| 80 |
+
sections.append({
|
| 81 |
+
"title": clean_text_block(sec["name"]),
|
| 82 |
+
"content": clean_text_block(content),
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 83 |
"articles": [
|
| 84 |
+
{"number": a["number"], "text": clean_text_block(a["text"])}
|
|
|
|
|
|
|
|
|
|
| 85 |
for a in articles
|
| 86 |
]
|
| 87 |
+
})
|
|
|
|
|
|
|
| 88 |
|
| 89 |
+
# 5) إرجاع المستند القانوني الكامل
|
|
|
|
|
|
|
| 90 |
return {
|
| 91 |
"message": "تم التحليل بنجاح",
|
| 92 |
"saved_to_db": False,
|
| 93 |
"law": {
|
| 94 |
+
"title": clean_text_block(title),
|
| 95 |
+
"preamble": clean_text_block(preamble),
|
| 96 |
"sections": sections
|
| 97 |
}
|
| 98 |
}
|