support DLSite's multipart RAR archives for larger works

dlibrary/dlibrary.py

@@ -21,6 +21,7 @@ import fitz
 from libsixel import *
 from PIL import Image
 from jinja2 import Environment, PackageLoader, select_autoescape
+import rarfile
 import requests
 
 NUMBER_REGEX = re.compile('[0-9０-９]+')
@@ -84,32 +85,67 @@ PDF_PREVIEW_DPI = 72
 
 IRRELEVANT_PDF_BLOCK_REGEX = re.compile(r'\bTCPDF\b', re.I)
 
+MULTIPART_RAR_HEAD_REGEX = re.compile(r'^(.+)\.part0*1\.exe$', re.I)
+MULTIPART_RAR_TAIL_REGEX = re.compile(r'^(.+)\.part0*([^1]|[^0].+)\.rar$', re.I)
+
 def open_zipfile_with_encoding(path):
-    try:
-        return zipfile.ZipFile(path, metadata_encoding="utf-8")
-    except UnicodeDecodeError:
-        pass
+    for enc in ["utf-8", "shift-jis", "shift-jisx0213"]:
+        try:
+            return zipfile.ZipFile(path, metadata_encoding=enc)
+        except UnicodeDecodeError:
+            pass
 
-    try:
-        return zipfile.ZipFile(path, metadata_encoding="shift-jis")
-    except UnicodeDecodeError:
-        pass
+    print(f'{path} contains filenames with unknown character encoding!')
+    exit(1)
 
-    return zipfile.ZipFile(path, metadata_encoding="shift-jisx0213")
+def open_rarfile_with_encoding(path):
+    for enc in ["utf-8", "shift-jis", "shift-jisx0213"]:
+        rf = rarfile.RarFile(path, charset=enc)
+        if all('<EFBFBD>' not in info.filename for info in rf.infolist()):
+            return rf
+
+    print(f'{path} contains filenames with unknown character encoding!')
+    exit(1)
 
 def extract(args):
-    for zip_path in args.zipfiles:
-        work_id = zip_path.stem
-        work_extract_path = args.destdir / 'extract' / work_id
-        work_extract_path.mkdir(parents=True)
+    absolute_archive_paths = set(path.resolve(strict=True) for path in args.archives)
 
-        print(f'Extracting {zip_path} to {work_extract_path}')
+    for archive_path in args.archives:
+        if archive_path.suffix.lower() == '.zip':
+            work_id = archive_path.stem
+            work_extract_path = args.destdir / 'extract' / work_id
 
-        with open_zipfile_with_encoding(zip_path) as z:
-            z.extractall(path=work_extract_path)
+            print(f'Extracting {archive_path} to {work_extract_path}')
 
-        if args.remove:
-            zip_path.unlink()
+            with open_zipfile_with_encoding(archive_path) as z:
+                work_extract_path.mkdir(parents=True)
+                z.extractall(path=work_extract_path)
+
+            if args.remove:
+                archive_path.unlink()
+
+        elif rar_match := MULTIPART_RAR_HEAD_REGEX.fullmatch(archive_path.name):
+            work_id = rar_match.group(1)
+            work_extract_path = args.destdir / 'extract' / work_id
+
+            print(f'Extracting multipart RAR archive beginning with {archive_path} to {work_extract_path}')
+
+            with open_rarfile_with_encoding(archive_path) as r:
+                volumes = [Path(vol).resolve(strict=True) for vol in r.volumelist()]
+                if any(vol not in absolute_archive_paths for vol in volumes):
+                    print(f'Multipart RAR archive starting with {archive_path} contains volume files not listed on command-line, skipping')
+                    continue
+                work_extract_path.mkdir(parents=True)
+                r.extractall(path=work_extract_path)
+
+            if args.remove:
+                for vol in volumes:
+                    vol.unlink()
+
+        elif MULTIPART_RAR_TAIL_REGEX.fullmatch(archive_path.name):
+            pass
+        else:
+            print(f'Unknown archive file type {archive_path}, skipping')
 
 
 def manual_input_metadata(work_id):
@@ -1073,9 +1109,8 @@ argparser = argparse.ArgumentParser(
     that can be viewed in a web browser.
 
     Intended workflow:
-    - `extract` a collection of zipfiles downloaded from DLSite
-      into DLibrary's data directory, giving each work its own
-      subfolder.
+    - `extract` a collection of archive files into DLibrary's data
+      directory, automatically giving each work its own subfolder.
     - `fetch` metadata and thumbnail images for extracted works
       from DLSite.
     - `collate` extracted works, producing a single sequence of
@@ -1103,18 +1138,18 @@ argparser.add_argument(
 )
 subparsers = argparser.add_subparsers(title="subcommands", required=True)
 
-parser_extract = subparsers.add_parser('extract', aliases=['x'], help='extract zipfiles')
+parser_extract = subparsers.add_parser('extract', aliases=['x'], help='extract archive files')
 parser_extract.add_argument(
     '-r', '--remove',
     action='store_true',
-    help='remove original zipfiles after extraction',
+    help='remove original archive files after extraction',
 )
 parser_extract.add_argument(
-    'zipfiles',
+    'archives',
     metavar='FILE',
     type=Path,
     nargs='+',
-    help='zipfiles to extract',
+    help='archive files to extract',
 )
 parser_extract.set_defaults(func=extract)
 

flake.nix

@@ -6,6 +6,7 @@
   outputs = { self, nixpkgs }: let
     pkgs = import nixpkgs {
       system = "x86_64-linux";
+      config.allowUnfree = true;
     };
   in {
     packages.x86_64-linux = with pkgs.python3Packages; rec {
@@ -58,6 +59,7 @@
           importlib-resources
           setuptools
           libsixel
+          (rarfile.override { useUnrar = true; })
         ];
         src = ./.;
       };

pyproject.toml

@@ -12,6 +12,7 @@ dependencies = [
     "jinja2",
     "importlib_resources",
     "libsixel-python",
+    "rarfile",
 ]
 
 [project.scripts]