Processor resource discovery

ocrd/ocrd/__init__.py

@@ -20,3 +20,4 @@
 from ocrd_validators import *
 from ocrd.workspace import Workspace
 from ocrd.workspace_backup import WorkspaceBackupManager
+from ocrd.resource_manager import OcrdResourceManager

ocrd/ocrd/cli/resmgr.py

@@ -2,8 +2,9 @@
 from os import getcwd
 from os.path import join
 from pathlib import Path
-import requests
+from distutils.spawn import find_executable as which
 
+import requests
 import click
 
 from ocrd_utils import (
@@ -56,11 +57,12 @@ def list_installed(executable=None):
 
 @resmgr_cli.command('download')
 @click.option('-n', '--any-url', help='Allow downloading/copying unregistered resources', is_flag=True)
+@click.option('-a', '--allow-uninstalled', help="Allow installing resources for uninstalled processors", is_flag=True)
 @click.option('-o', '--overwrite', help='Overwrite existing resources', is_flag=True)
 @click.option('-l', '--location', help='Where to store resources', type=click.Choice(RESOURCE_LOCATIONS), default='data', show_default=True)
 @click.argument('executable', required=True)
-@click.argument('url_or_name', required=True)
-def download(any_url, overwrite, location, executable, url_or_name):
+@click.argument('url_or_name', required=False)
+def download(any_url, allow_uninstalled, overwrite, location, executable, url_or_name):
     """
     Download resource URL_OR_NAME for processor EXECUTABLE.
 
@@ -73,10 +75,22 @@ def download(any_url, overwrite, location, executable, url_or_name):
     log = getLogger('ocrd.cli.resmgr')
     resmgr = OcrdResourceManager()
     basedir = resmgr.location_to_resource_dir(location)
-    is_url = url_or_name.startswith('https://') or url_or_name.startswith('http://')
-    is_filename = Path(url_or_name).exists()
+    if executable != '*' and not url_or_name:
+        log.error("Unless EXECUTABLE ('%s') is the '*' wildcard, URL_OR_NAME is required" % executable)
+        sys.exit(1)
+    elif executable == '*':
+        executable = None
+    is_url = (url_or_name.startswith('https://') or url_or_name.startswith('http://')) if url_or_name else False
+    is_filename = Path(url_or_name).exists() if url_or_name else False
+    if executable and not which(executable):
+        if not allow_uninstalled:
+            log.error("Executable %s is not installed. Is there a typo in the executable? " \
+                "To install resources for uninstalled processor, use the -a/--allow-uninstalled flag" % executable)
+            sys.exit(1)
+        else:
+            log.warning("Executable %s is not installed but -a/--allow-uninstalled was given, so proceeding" % executable)
     find_kwargs = {'executable': executable}
-    if url_or_name != '*':
+    if url_or_name and url_or_name != '*':
         find_kwargs['url' if is_url else 'name'] = url_or_name
     reslist = resmgr.find_resources(**find_kwargs)
     if not reslist:
@@ -103,7 +117,10 @@ def download(any_url, overwrite, location, executable, url_or_name):
         else:
             sys.exit(1)
     else:
-        for _, resdict in reslist:
+        for executable, resdict in reslist:
+            if not allow_uninstalled and not which(executable):
+                log.info("Skipping installing resources for %s as it is not installed. (Use -a/--allow-uninstalled to force)")
+                continue
             if resdict['url'] == '???':
                 log.info("Cannot download user resource %s" % (resdict['name'])),
                 continue

ocrd/ocrd/resource_manager.py

@@ -70,7 +70,7 @@ def list_installed(self, executable=None):
             # resources we know about
             all_executables = list(self.database.keys())
             # resources in the file system
-            parent_dirs = [join(x, 'ocrd-resources') for x in [XDG_DATA_HOME, '/usr/local/share']]
+            parent_dirs = [join(x, 'ocrd-resources') for x in [XDG_DATA_HOME, '/usr/local/share', getcwd()]]
             for parent_dir in parent_dirs:
                 if Path(parent_dir).exists():
                     all_executables += [x for x in listdir(parent_dir) if x.startswith('ocrd-')]
@@ -131,6 +131,9 @@ def find_resources(self, executable=None, name=None, url=None, database=None):
                     ret.append((executable, resdict))
         return ret
 
+    @property
+    def default_resource_dir(self):
+        return self.location_to_resource_dir('data')
     def location_to_resource_dir(self, location):
         return '/usr/local/share/ocrd-resources' if location == 'system' else \
                 join(XDG_DATA_HOME, 'ocrd-resources') if location == 'data' else \