Views
No views yet
| Fitur | Deskripsi |
|---|---|
similarity_score() | Skor kemiripan semantik antar dua pasal (cross-lingual EN↔ID) |
compare() | Gap analysis lengkap antara dua dokumen hukum |
search() | Cari pasal berdasarkan topik/query bebas |
summarize_gap() | Ringkasan perbedaan dalam format naratif/poin/tabel |
pip install sentence-transformers scikit-learn keybert numpy1from src.legal_comparator import LegalComparator
2
3# 1. Inisialisasi model
4model = LegalComparator()
5
6# 2. Load corpus
7model.load(
8 corpus_path = "corpus_structured.json",
9 embedding_path = "embeddings_labse.npy"
10)
11
12# 3. Similarity score dua pasal spesifik
13hasil = model.similarity_score(
14 pasal_a_id = "UNCAC_article_15",
15 pasal_b_id = "UU_31_1999_pasal_5"
16)
17print(hasil["score"], hasil["interpretasi"])
18
19# 4. Gap analysis penuh
20compare = model.compare(doc_a="UNCAC", doc_b="UU_31_1999")
21
22# 5. Ringkasan perbedaan
23print(model.summarize_gap(compare, gaya="naratif"))
24
25# 6. Cari pasal berdasarkan topik
26search = model.search("suap kepada pejabat negara", top_n=5)| Label | Dokumen | Tahun | Bahasa |
|---|---|---|---|
UNCAC | UN Convention Against Corruption | 2003 | EN |
UU_7_2006 | UU Ratifikasi UNCAC | 2006 | ID |
UU_31_1999 | UU Tipikor | 1999 | ID |
UU_20_2001 | Amandemen UU Tipikor | 2001 | ID |
UU_28_1999 | UU Penyelenggaraan Negara Bersih KKN | 1999 | ID |
UU_30_2002 | UU KPK | 2002 | ID |
UU_19_2019 | UU Revisi KPK | 2019 | ID |
sentence-transformers/LaBSE (109 bahasa, cross-lingual)KeyBERT + paraphrase-multilingual-MiniLM-L12-v2