allow forcing init in dirty directory

archivebox/cli/archivebox_init.py

@@ -21,10 +21,18 @@ def main(args: Optional[List[str]]=None, stdin: Optional[IO]=None, pwd: Optional
         add_help=True,
         formatter_class=SmartFormatter,
     )
-    parser.parse_args(args or ())
+    parser.add_argument(
+        '--force', # '-f',
+        action='store_true',
+        help='Ignore unrecognized files in current directory and initialize anyway',
+    )
+    command = parser.parse_args(args or ())
     reject_stdin(__command__, stdin)
 
-    init(out_dir=pwd or OUTPUT_DIR)
+    init(
+        force=command.force,
+        out_dir=pwd or OUTPUT_DIR,
+    )
     
 
 if __name__ == '__main__':

archivebox/main.py

@@ -238,7 +238,7 @@ def run(subcommand: str,
 
 
 @enforce_types
-def init(out_dir: str=OUTPUT_DIR) -> None:
+def init(force: bool=False, out_dir: str=OUTPUT_DIR) -> None:
     """Initialize a new ArchiveBox collection in the current directory"""
     os.makedirs(out_dir, exist_ok=True)
 
@@ -253,6 +253,10 @@ def init(out_dir: str=OUTPUT_DIR) -> None:
         print('{green}[*] Updating existing ArchiveBox collection in this folder...{reset}'.format(**ANSI))
         print(f'    {out_dir}')
         print('{green}------------------------------------------------------------------{reset}'.format(**ANSI))
+    else:
+        if force:
+            stderr('[!] This folder appears to already have files in it, but no index.json is present.', color='lightyellow')
+            stderr('    Because --force was passed, ArchiveBox will initialize anyway (which may overwrite existing files).')
         else:
             stderr(
                 ("{red}[X] This folder appears to already have files in it, but no index.json is present.{reset}\n\n"
@@ -262,7 +266,7 @@ def init(out_dir: str=OUTPUT_DIR) -> None:
                 "    (Always make sure your data folder is backed up first before updating ArchiveBox)"
                 ).format(out_dir, **ANSI)
             )
-        raise SystemExit(1)
+            raise SystemExit(2)
 
     if existing_index:
         print('\n{green}[*] Verifying archive folder structure...{reset}'.format(**ANSI))