Ensure license index is buildable with extras

Somehow building the index for all languages was failing.
This is fixing this issue.

Signed-off-by: Philippe Ombredanne <[email protected]>

Author: pombredanne

src/licensedcode/cache.py

@@ -240,9 +240,12 @@ def build_index(
     if not licenses_db:
         # combine the licenses in these additional directories with the licenses in the original DB
         additional_license_dirs = get_license_dirs(additional_dirs=additional_directories)
-        combined_license_directories = [licenses_data_dir] + additional_license_dirs
         # generate a single combined license db with all licenses
-        licenses_db = load_licenses_from_multiple_dirs(license_dirs=combined_license_directories)
+        licenses_db = load_licenses_from_multiple_dirs(
+            builtin_license_data_dir=licenses_data_dir,
+            additional_license_data_dirs=additional_license_dirs,
+            with_deprecated=False,
+        )
 
     # if we have additional directories, extract the rules from them
     additional_rule_dirs = get_rule_dirs(additional_dirs=additional_directories)

src/licensedcode/models.py

@@ -38,6 +38,7 @@
 from licensedcode.frontmatter import dumps_frontmatter
 from licensedcode.frontmatter import load_frontmatter
 from licensedcode.languages import LANG_INFO as known_languages
+from licensedcode.tokenize import get_existing_required_phrase_spans
 from licensedcode.tokenize import index_tokenizer
 from licensedcode.tokenize import index_tokenizer_with_stopwords
 from licensedcode.tokenize import query_lines
@@ -890,6 +891,7 @@ def get_license_dirs(additional_dirs):
     Return a list of all subdirectories containing license files within the
     input list of additional directories. These directories do not have to be absolute paths.
     """
+    additional_dirs = additional_dirs or []
     return [f"{str(Path(path).absolute())}/licenses" for path in additional_dirs]
 
 
@@ -898,6 +900,7 @@ def get_rule_dirs(additional_dirs):
     Return a list of all subdirectories containing rule files within the
     input list of additional directories. These directories do not have to be absolute paths.
     """
+    additional_dirs = additional_dirs or []
     return [f"{str(Path(path).absolute())}/rules" for path in additional_dirs]
 
 
@@ -2295,7 +2298,6 @@ def build_required_phrase_spans(self):
         Return a list of Spans marking required phrases token positions of that must
         be present for this rule to be matched.
         """
-        from licensedcode.required_phrases import get_existing_required_phrase_spans
         if self.is_from_license:
             return []
         try: