Store filenames with extensions in name field #467

    * Update scancode-toolkit to 31.0.0rc3 #447
    * Update expected test results

Signed-off-by: Jono Yang <[email protected]>

Author: JonoYang

CHANGELOG.rst

@@ -56,6 +56,13 @@ v31.0.0 (next)
 
   https://github.com/nexB/scancode.io/issues/413
 
+- CodebaseResource.name now contains both the bare file name with extension, as
+  opposed to just the bare file name without extension.
+  https://github.com/nexB/scancode.io/issues/467
+
+  - Using a name stripped from its extension was something that was not used in
+    other AboutCode project or tools.
+
 v30.2.0 (2021-12-17)
 --------------------
 

scanpipe/migrations/0020_concat_codebaseresource_name_and_extension.py

@@ -0,0 +1,27 @@
+# Generated by Django 4.0.6 on 2022-07-27 16:23
+
+from django.db import migrations
+
+
+def concat_codebaseresource_name_and_extension(apps, schema_editor):
+    CodebaseResource = apps.get_model('scanpipe', 'CodebaseResource')
+    for cb in CodebaseResource.objects.filter(extension__isnull=False).iterator():
+        name = cb.name
+        extension = cb.extension
+        # The scan_package pipeline was putting the extension in the name field
+        # whereas the scan_codebase pipeline did not put the extension in the
+        # name field
+        if not name.endswith(extension):
+            cb.name = f"{name}{extension}"
+            cb.save()
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('scanpipe', '0019_codebaseresource_package_data_discovereddependency'),
+    ]
+
+    operations = [
+        migrations.RunPython(concat_codebaseresource_name_and_extension),
+    ]

scanpipe/models.py

@@ -1428,7 +1428,7 @@ class Type(models.TextChoices):
     name = models.CharField(
         max_length=255,
         blank=True,
-        help_text=_("File or directory name of this resource."),
+        help_text=_("File or directory name of this resource with its extension."),
     )
     extension = models.CharField(
         max_length=100,
@@ -1775,6 +1775,14 @@ def for_packages(self):
         """
         return [str(package) for package in self.discovered_packages.all()]
 
+    def for_packages_append(self, package_uid):
+        if not package_uid:
+            return
+        package = DiscoveredPackage.objects.get(package_uid=package_uid)
+        if package not in self.discovered_packages.all():
+            self.discovered_packages.add(package)
+            self.save()
+
 
 class DiscoveredPackageQuerySet(PackageURLQuerySetMixin, ProjectRelatedQuerySet):
     pass

scanpipe/pipes/scancode.py

@@ -139,7 +139,7 @@ def get_resource_info(location):
     file_info.update(
         {
             "type": resource_type,
-            "name": fileutils.file_base_name(location),
+            "name": fileutils.file_name(location),
             "extension": fileutils.file_extension(location),
         }
     )