fetch_documents.py

#!/usr/bin/env python3
"""
Fetch Ancient Legal Document Images and Texts

Sources:
- Wikimedia Commons (public domain images)
- Yale Avalon Project (translations)
- Internet Archive (scans)
- Museum APIs (Louvre, British Museum, etc.)
"""

import os
import json
import requests
import time
from pathlib import Path
from typing import Dict, List, Optional
from dataclasses import dataclass, asdict
from urllib.parse import urljoin
import logging

logging.basicConfig(level=logging.INFO, format='%(asctime)s | %(message)s')
logger = logging.getLogger(__name__)

# Base paths
BASE_DIR = Path(__file__).parent
IMAGES_DIR = BASE_DIR / "images"
TRANSCRIPTIONS_DIR = BASE_DIR / "transcriptions"
TRANSLATIONS_DIR = BASE_DIR / "translations"

# Ensure directories exist
IMAGES_DIR.mkdir(exist_ok=True)
TRANSCRIPTIONS_DIR.mkdir(exist_ok=True)
TRANSLATIONS_DIR.mkdir(exist_ok=True)


@dataclass
class AncientDocument:
    """Metadata for an ancient legal document"""
    id: str
    title: str
    alternate_titles: List[str]
    date_composed: str
    language_original: str
    script: str
    physical_location: str
    image_sources: List[Dict[str, str]]  # [{"url": ..., "description": ...}]
    translation_sources: List[Dict[str, str]]  # [{"url": ..., "language": ...}]
    provisions_count: Optional[int] = None
    topics: List[str] = None
    significance: str = ""


# ============================================================================
# DOCUMENT DEFINITIONS - Known sources for ancient legal documents
# ============================================================================

ANCIENT_DOCUMENTS = [
    AncientDocument(
        id="hammurabi",
        title="Code of Hammurabi",
        alternate_titles=["Codex Hammurabi", "Laws of Hammurabi"],
        date_composed="-1754",
        language_original="akk",
        script="cuneiform",
        physical_location="Louvre Museum, Paris",
        provisions_count=282,
        topics=["property", "family", "commerce", "crime", "slavery", "wages"],
        significance="One of the oldest deciphered writings of significant length in the world",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/6/64/Milkau_Oberer_Teil_der_Stele_mit_dem_Text_von_Hammurapis_Gesetzescode_369-2.jpg",
                "description": "Full stele showing Hammurabi receiving laws from Shamash"
            },
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/a/a7/Code_de_Hammurabi_1.jpg",
                "description": "Upper portion detail"
            },
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/e/e3/P1050771_Louvre_code_Hammurabi_face_rwk.JPG",
                "description": "Front face with cuneiform text"
            },
        ],
        translation_sources=[
            {"url": "https://avalon.law.yale.edu/ancient/hamframe.asp", "language": "en"},
        ]
    ),
    AncientDocument(
        id="ur_nammu",
        title="Code of Ur-Nammu",
        alternate_titles=["Laws of Ur-Nammu"],
        date_composed="-2100",
        language_original="sux",
        script="cuneiform",
        physical_location="Istanbul Archaeological Museums / University of Pennsylvania",
        provisions_count=57,
        topics=["murder", "robbery", "adultery", "injury", "slavery"],
        significance="Oldest known law code, predates Hammurabi by 300 years",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/5/5e/Code_of_Ur-Nammu_%28Istanbul%29.jpg",
                "description": "Ur-Nammu tablet"
            },
        ],
        translation_sources=[
            {"url": "https://www.ancient.eu/article/1000/the-ur-nammu-law-code/", "language": "en"},
        ]
    ),
    AncientDocument(
        id="eshnunna",
        title="Laws of Eshnunna",
        alternate_titles=["Codex Eshnunna"],
        date_composed="-1930",
        language_original="akk",
        script="cuneiform",
        physical_location="Iraq Museum, Baghdad",
        provisions_count=60,
        topics=["prices", "wages", "family", "property", "slavery"],
        significance="Predates Hammurabi, shows Akkadian legal tradition",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/8/8a/Law_tablet_Eshnunna_Louvre_AO6676.jpg",
                "description": "Eshnunna law tablet"
            },
        ],
        translation_sources=[]
    ),
    AncientDocument(
        id="twelve_tables",
        title="Twelve Tables",
        alternate_titles=["Law of the Twelve Tables", "Lex Duodecim Tabularum"],
        date_composed="-450",
        language_original="la",
        script="latin",
        physical_location="Original destroyed; reconstructed from citations",
        provisions_count=100,  # approximate
        topics=["procedure", "family", "property", "inheritance", "crime"],
        significance="Foundation of Roman law, basis of Western legal tradition",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/0/0b/Twelve_Tables.jpg",
                "description": "Artistic reconstruction"
            },
        ],
        translation_sources=[
            {"url": "https://avalon.law.yale.edu/ancient/twelve_tables.asp", "language": "en"},
        ]
    ),
    AncientDocument(
        id="magna_carta",
        title="Magna Carta",
        alternate_titles=["Great Charter", "Magna Carta Libertatum"],
        date_composed="1215",
        language_original="la",
        script="latin",
        physical_location="British Library / Salisbury Cathedral / Lincoln Castle",
        provisions_count=63,
        topics=["rights", "justice", "taxation", "church", "feudal"],
        significance="Foundation of constitutional law and individual rights",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/e/ee/Magna_Carta_%28British_Library_Cotton_MS_Augustus_II.106%29.jpg",
                "description": "British Library copy (Cotton MS Augustus II.106)"
            },
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/0/06/Magna_Carta_%28An_Emblematic_Picture%29.jpg",
                "description": "Emblematic picture"
            },
        ],
        translation_sources=[
            {"url": "https://avalon.law.yale.edu/medieval/magframe.asp", "language": "en"},
        ]
    ),
    AncientDocument(
        id="draco",
        title="Draconian Constitution",
        alternate_titles=["Laws of Draco", "Draco's Code"],
        date_composed="-621",
        language_original="grc",
        script="greek",
        physical_location="Lost; known only from later sources",
        topics=["homicide", "punishment", "debt"],
        significance="First written Athenian laws; notoriously harsh",
        image_sources=[],  # No original survives
        translation_sources=[]
    ),
    AncientDocument(
        id="solon",
        title="Laws of Solon",
        alternate_titles=["Solonian Constitution"],
        date_composed="-594",
        language_original="grc",
        script="greek",
        physical_location="Lost; known from citations",
        topics=["debt", "citizenship", "courts", "assembly"],
        significance="Reformed Draconian laws; foundation of Athenian democracy",
        image_sources=[],
        translation_sources=[]
    ),
    AncientDocument(
        id="hittite_laws",
        title="Hittite Laws",
        alternate_titles=["Code of the Nesilim"],
        date_composed="-1650",
        language_original="hit",
        script="cuneiform",
        physical_location="Various museums",
        provisions_count=200,
        topics=["homicide", "theft", "marriage", "land", "animals"],
        significance="Shows legal diversity in ancient Near East",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/8/8a/Hittite_law_code.jpg",
                "description": "Hittite law tablet"
            },
        ],
        translation_sources=[]
    ),
    AncientDocument(
        id="manu",
        title="Laws of Manu",
        alternate_titles=["Manusmriti", "Manava-Dharmasastra"],
        date_composed="-200",
        language_original="sa",
        script="devanagari",
        physical_location="Various manuscripts",
        provisions_count=2685,  # verses
        topics=["dharma", "caste", "ritual", "kingship", "punishment"],
        significance="Foundational Hindu legal text",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/d/d6/Manusmriti.jpg",
                "description": "Manusmriti manuscript page"
            },
        ],
        translation_sources=[
            {"url": "https://www.sacred-texts.com/hin/manu.htm", "language": "en"},
        ]
    ),
    AncientDocument(
        id="justinian_digest",
        title="Digest of Justinian",
        alternate_titles=["Pandects", "Digesta"],
        date_composed="533",
        language_original="la",
        script="latin",
        physical_location="Various manuscripts",
        provisions_count=9000,  # excerpts
        topics=["persons", "property", "obligations", "succession", "procedure"],
        significance="Compilation of Roman legal writings; basis of civil law",
        image_sources=[
            {
                "url": "https://upload.wikimedia.org/wikipedia/commons/8/8f/Corpus_Iuris_Civilis_02.jpg",
                "description": "Medieval manuscript"
            },
        ],
        translation_sources=[]
    ),
]


def download_image(url: str, save_path: Path, description: str = "") -> bool:
    """Download an image from URL"""
    try:
        headers = {"User-Agent": "Mozilla/5.0 (research bot - ancient document collection)"}
        response = requests.get(url, headers=headers, timeout=60, stream=True)
        response.raise_for_status()

        with open(save_path, 'wb') as f:
            for chunk in response.iter_content(chunk_size=8192):
                f.write(chunk)

        logger.info(f"  Downloaded: {save_path.name} ({description})")
        return True

    except Exception as e:
        logger.error(f"  Failed to download {url}: {e}")
        return False


def fetch_avalon_translation(url: str, save_path: Path) -> bool:
    """Fetch translation from Yale Avalon Project"""
    try:
        from bs4 import BeautifulSoup

        headers = {"User-Agent": "Mozilla/5.0 (research bot)"}
        response = requests.get(url, headers=headers, timeout=30)
        response.raise_for_status()

        soup = BeautifulSoup(response.content, 'html.parser')

        # Extract main content (varies by page structure)
        content = soup.find('div', class_='document-content')
        if not content:
            content = soup.find('body')

        if content:
            text = content.get_text(separator='\n', strip=True)
            with open(save_path, 'w', encoding='utf-8') as f:
                f.write(text)
            logger.info(f"  Saved translation: {save_path.name}")
            return True

    except Exception as e:
        logger.error(f"  Failed to fetch translation from {url}: {e}")

    return False


def fetch_document(doc: AncientDocument) -> Dict:
    """Fetch all available resources for a document"""
    logger.info(f"\n{'='*60}")
    logger.info(f"Fetching: {doc.title} ({doc.date_composed})")
    logger.info(f"{'='*60}")

    result = {
        "id": doc.id,
        "title": doc.title,
        "images_downloaded": [],
        "translations_downloaded": [],
        "errors": []
    }

    # Create document directory for images
    doc_images_dir = IMAGES_DIR / doc.id
    doc_images_dir.mkdir(exist_ok=True)

    # Download images
    for i, img_source in enumerate(doc.image_sources):
        url = img_source["url"]
        ext = url.split('.')[-1].split('?')[0][:4]  # Get extension
        if ext not in ['jpg', 'jpeg', 'png', 'gif', 'tif', 'tiff']:
            ext = 'jpg'

        filename = f"{doc.id}_{i+1:02d}.{ext}"
        save_path = doc_images_dir / filename

        if save_path.exists():
            logger.info(f"  Already exists: {filename}")
            result["images_downloaded"].append(filename)
        elif download_image(url, save_path, img_source.get("description", "")):
            result["images_downloaded"].append(filename)
        else:
            result["errors"].append(f"Failed: {url}")

        time.sleep(1)  # Be polite

    # Fetch translations
    for trans_source in doc.translation_sources:
        url = trans_source["url"]
        lang = trans_source["language"]

        filename = f"{doc.id}_{lang}.txt"
        save_path = TRANSLATIONS_DIR / filename

        if save_path.exists():
            logger.info(f"  Already exists: {filename}")
            result["translations_downloaded"].append(filename)
        elif "avalon" in url:
            if fetch_avalon_translation(url, save_path):
                result["translations_downloaded"].append(filename)

        time.sleep(1)

    return result


def save_metadata():
    """Save document metadata to JSON"""
    metadata = {
        "dataset": "ancient-law",
        "version": "1.0.0",
        "documents": [asdict(doc) for doc in ANCIENT_DOCUMENTS]
    }

    metadata_path = BASE_DIR / "metadata.json"
    with open(metadata_path, 'w', encoding='utf-8') as f:
        json.dump(metadata, f, indent=2, ensure_ascii=False)

    logger.info(f"\nSaved metadata to {metadata_path}")


def main():
    """Run the fetch process"""
    print("=" * 70)
    print("ANCIENT LAW DOCUMENT FETCHER")
    print("Collecting document images and translations for HuggingFace")
    print("=" * 70)

    results = []

    for doc in ANCIENT_DOCUMENTS:
        result = fetch_document(doc)
        results.append(result)

    # Save metadata
    save_metadata()

    # Summary
    print("\n" + "=" * 70)
    print("FETCH COMPLETE")
    print("=" * 70)

    total_images = sum(len(r["images_downloaded"]) for r in results)
    total_translations = sum(len(r["translations_downloaded"]) for r in results)
    total_errors = sum(len(r["errors"]) for r in results)

    print(f"Documents processed: {len(results)}")
    print(f"Images downloaded: {total_images}")
    print(f"Translations saved: {total_translations}")
    print(f"Errors: {total_errors}")

    return results


if __name__ == "__main__":
    main()