273 lines
10 KiB
Python
273 lines
10 KiB
Python
# -*- coding: utf-8 -*-
|
||
"""
|
||
Importiert 10 freie Handwerk-/Technik-Bücher (Project Gutenberg) in die Bollwerk-Ressourcen.
|
||
Fokus: Nützliches Wissen für die Postapokalypse.
|
||
|
||
Verwendung:
|
||
python import-books-handwerk.py
|
||
python import-books-handwerk.py --dry-run
|
||
"""
|
||
|
||
import json
|
||
import os
|
||
import sys
|
||
import time
|
||
import uuid
|
||
import urllib.request
|
||
import urllib.error
|
||
|
||
SERVER_URL = "https://bollwerk.online"
|
||
|
||
BOOKS = [
|
||
{
|
||
"gutenberg_id": 10136,
|
||
"title": "The Book of Household Management",
|
||
"author": "Mrs. Beeton",
|
||
"description": "Das viktorianische Standardwerk für Haushaltsführung: Kochen, Konservieren, Reinigung, Krankenpflege, Tierhaltung, Budgetplanung. Ein Lexikon der praktischen Lebenskunst in 2000+ Seiten.",
|
||
"tags": ["nachschlagewerk", "handbuch"],
|
||
"language": "en",
|
||
"release_date": "1861-10-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 13510,
|
||
"title": "Knots, Splices and Rope Work",
|
||
"author": "A. Hyatt Verrill",
|
||
"description": "Vollständige Anleitung zu Knoten, Spleißen und Seilwerk: über 100 Knoten mit Abbildungen. Unverzichtbar für Bauen, Sichern, Transportieren und Überleben ohne moderne Hilfsmittel.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1917-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 47760,
|
||
"title": "Three Hundred Things a Bright Boy Can Do",
|
||
"author": "Various",
|
||
"description": "Praktisches DIY-Kompendium: Elektrik, Mechanik, Holzarbeiten, Metallbearbeitung, Fotografie, Modellbau. 300 Projekte mit detaillierten Bauanleitungen – ein Maker-Handbuch aus dem frühen 20. Jahrhundert.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1911-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 56601,
|
||
"title": "A Text-book of Tanning",
|
||
"author": "H. R. Procter",
|
||
"description": "Wissenschaftliches Lehrbuch der Lederherstellung: Gerbstoffe, Hautbehandlung, pflanzliches und chemisches Gerben. Essentielles Handwerk für Kleidung, Schuhe und Ausrüstung ohne Industrie.",
|
||
"tags": ["handbuch", "nachschlagewerk"],
|
||
"language": "en",
|
||
"release_date": "1885-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 67636,
|
||
"title": "The Useful Arts Employed in the Construction of Dwelling Houses",
|
||
"author": "Anonymous",
|
||
"description": "Detailliertes Handbuch über alle Gewerke beim Hausbau: Mauerwerk, Zimmerei, Dachdecken, Klempnerei, Verglasung, Putz und Anstrich. Praxis-Referenz für den Wiederaufbau.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1860-01-01",
|
||
"edition": "Second Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 56776,
|
||
"title": "Practical Hand Book of Gas, Oil and Steam Engines",
|
||
"author": "John B. Rathbun",
|
||
"description": "Praxishandbuch für Verbrennungs- und Dampfmotoren: Aufbau, Wartung, Reparatur, Fehlersuche. Wissen, um Energieerzeugung und Antriebstechnik ohne Stromnetz zu beherrschen.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1916-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 21682,
|
||
"title": "The Field and Garden Vegetables of America",
|
||
"author": "Fearing Burr",
|
||
"description": "Enzyklopädie mit fast 1100 Gemüsesorten: Beschreibung, Anbau, Ernte, Lagerung. Das umfassendste Nachschlagewerk für Selbstversorger-Gartenbau in der Public Domain.",
|
||
"tags": ["nachschlagewerk"],
|
||
"language": "en",
|
||
"release_date": "1863-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 50451,
|
||
"title": "A Boy's Workshop",
|
||
"author": "Harry Craigin",
|
||
"description": "Einführung in grundlegende Werkstattarbeiten: Holzverbindungen, Metallarbeiten, Werkzeugherstellung, einfache Maschinenbau-Projekte. Praktisches Handwerk von Grund auf erklärt.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1894-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 53139,
|
||
"title": "Maxims and Instructions for the Boiler Room",
|
||
"author": "N. Hawkins",
|
||
"description": "Praxiswissen für Dampfkesselbetrieb: Druckregelung, Brennstoffeffizienz, Sicherheit, Wartung. Schlüsselkompetenz für jede postapokalyptische Dampf-Infrastruktur.",
|
||
"tags": ["handbuch"],
|
||
"language": "en",
|
||
"release_date": "1903-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
{
|
||
"gutenberg_id": 65967,
|
||
"title": "The History of Silk, Cotton, Linen, Wool, and Other Fibrous Substances",
|
||
"author": "Clinton G. Gilroy",
|
||
"description": "Geschichte und Technik der Textilherstellung: Spinnen, Weben, Färben von der Rohfaser zum fertigen Stoff. Grundlagenwissen, um ohne Textilindustrie Kleidung zu produzieren.",
|
||
"tags": ["nachschlagewerk", "handbuch"],
|
||
"language": "en",
|
||
"release_date": "1845-01-01",
|
||
"edition": "First Edition",
|
||
},
|
||
]
|
||
|
||
# ---------------------------------------------------------------------------
|
||
RESET = "\033[0m"
|
||
GREEN = "\033[92m"
|
||
YELLOW = "\033[93m"
|
||
RED = "\033[91m"
|
||
|
||
def ok(msg): print(f"{GREEN}[OK] {msg}{RESET}", flush=True)
|
||
def fail(msg): print(f"{RED}[!!] {msg}{RESET}", flush=True)
|
||
def step(msg): print(f"\n{YELLOW}{msg}{RESET}", flush=True)
|
||
def info(msg): print(f" {msg}", flush=True)
|
||
|
||
|
||
def login(username: str, password: str) -> str:
|
||
payload = json.dumps({"username": username, "password": password}).encode()
|
||
req = urllib.request.Request(
|
||
f"{SERVER_URL}/api/auth/login",
|
||
data=payload,
|
||
headers={"Content-Type": "application/json"},
|
||
method="POST",
|
||
)
|
||
try:
|
||
with urllib.request.urlopen(req) as resp:
|
||
data = json.loads(resp.read())
|
||
return data["accessToken"]
|
||
except urllib.error.HTTPError as e:
|
||
body = e.read().decode() if e.fp else ""
|
||
fail(f"Login fehlgeschlagen: {e.code} {body}")
|
||
sys.exit(1)
|
||
|
||
|
||
def download_epub(gutenberg_id: int) -> bytes:
|
||
urls = [
|
||
f"https://www.gutenberg.org/ebooks/{gutenberg_id}.epub.noimages",
|
||
f"https://www.gutenberg.org/ebooks/{gutenberg_id}.epub3.images",
|
||
f"https://www.gutenberg.org/cache/epub/{gutenberg_id}/pg{gutenberg_id}.epub",
|
||
]
|
||
for url in urls:
|
||
try:
|
||
req = urllib.request.Request(url, headers={"User-Agent": "BollwerkImporter/1.0"})
|
||
with urllib.request.urlopen(req, timeout=60) as resp:
|
||
data = resp.read()
|
||
if len(data) > 1000:
|
||
return data
|
||
except (urllib.error.HTTPError, urllib.error.URLError):
|
||
continue
|
||
raise RuntimeError(f"Konnte Gutenberg #{gutenberg_id} nicht herunterladen")
|
||
|
||
|
||
def upload_resource(token: str, book: dict, file_bytes: bytes) -> dict:
|
||
guid = str(uuid.uuid4())
|
||
now = int(time.time() * 1000)
|
||
|
||
metadata = {
|
||
"guid": guid,
|
||
"title": book["title"],
|
||
"description": book["description"],
|
||
"tags": book["tags"],
|
||
"fileFormat": "epub",
|
||
"mimeType": "application/epub+zip",
|
||
"fileSize": len(file_bytes),
|
||
"releaseDate": book.get("release_date"),
|
||
"createdAt": now,
|
||
"updatedAt": now,
|
||
"author": book.get("author"),
|
||
"language": book.get("language", "en"),
|
||
"edition": book.get("edition"),
|
||
"downloadUrl": "",
|
||
}
|
||
|
||
boundary = f"----BollwerkBoundary{uuid.uuid4().hex[:16]}"
|
||
body = bytearray()
|
||
body += f"--{boundary}\r\n".encode()
|
||
body += b'Content-Disposition: form-data; name="metadata"\r\n'
|
||
body += b"Content-Type: application/json\r\n\r\n"
|
||
body += json.dumps(metadata).encode()
|
||
body += b"\r\n"
|
||
body += f"--{boundary}\r\n".encode()
|
||
body += f'Content-Disposition: form-data; name="file"; filename="{guid}.epub"\r\n'.encode()
|
||
body += b"Content-Type: application/epub+zip\r\n\r\n"
|
||
body += file_bytes
|
||
body += b"\r\n"
|
||
body += f"--{boundary}--\r\n".encode()
|
||
|
||
req = urllib.request.Request(
|
||
f"{SERVER_URL}/api/admin/resources",
|
||
data=bytes(body),
|
||
headers={
|
||
"Authorization": f"Bearer {token}",
|
||
"Content-Type": f"multipart/form-data; boundary={boundary}",
|
||
},
|
||
method="POST",
|
||
)
|
||
|
||
try:
|
||
with urllib.request.urlopen(req, timeout=120) as resp:
|
||
return json.loads(resp.read())
|
||
except urllib.error.HTTPError as e:
|
||
error_body = e.read().decode() if e.fp else ""
|
||
raise RuntimeError(f"Upload fehlgeschlagen: {e.code} {error_body}")
|
||
|
||
|
||
def main():
|
||
dry_run = "--dry-run" in sys.argv
|
||
count = len(BOOKS)
|
||
|
||
step(f"=== Bollwerk Bücher-Import ({count} Handwerk & Technik) ===")
|
||
|
||
if not dry_run:
|
||
username = os.environ.get("BOLLWERK_ADMIN_USER", "")
|
||
password = os.environ.get("BOLLWERK_ADMIN_PASS", "")
|
||
if not username or not password:
|
||
fail("Setze BOLLWERK_ADMIN_USER und BOLLWERK_ADMIN_PASS als Env-Vars")
|
||
sys.exit(1)
|
||
step("1/3 Login als Admin...")
|
||
token = login(username, password)
|
||
ok(f"Eingeloggt als '{username}'")
|
||
else:
|
||
token = ""
|
||
info("DRY-RUN: Kein Login, kein Upload")
|
||
|
||
step("2/3 Bücher herunterladen...")
|
||
downloads = []
|
||
for i, book in enumerate(BOOKS, 1):
|
||
info(f" [{i:2d}/{count}] {book['title']} (Gutenberg #{book['gutenberg_id']})...")
|
||
try:
|
||
data = download_epub(book["gutenberg_id"])
|
||
downloads.append((book, data))
|
||
ok(f" {book['title']} – {len(data) / 1024:.0f} KB")
|
||
except RuntimeError as e:
|
||
fail(f" {e}")
|
||
|
||
if not dry_run:
|
||
step("3/3 Upload auf Bollwerk-Server...")
|
||
success = 0
|
||
for i, (book, data) in enumerate(downloads, 1):
|
||
info(f" [{i:2d}/{len(downloads)}] {book['title']}...")
|
||
try:
|
||
result = upload_resource(token, book, data)
|
||
ok(f" {book['title']} → guid={result['guid']}")
|
||
success += 1
|
||
except RuntimeError as e:
|
||
fail(f" {e}")
|
||
time.sleep(0.5)
|
||
step(f"Fertig: {success}/{len(downloads)} Bücher erfolgreich importiert.")
|
||
else:
|
||
step(f"DRY-RUN abgeschlossen: {len(downloads)} Bücher heruntergeladen, 0 hochgeladen.")
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main()
|