loader.pypi: Separate computations and storage instructions

swh/loader/pypi/loader.py

@@ -98,26 +98,18 @@ class PyPiLoader(SWHLoader):
                                    self.origin_metadata_url)
 
     def fetch_data(self):
-        """(override) This will fetch and prepare the needed releases.
+        """(override) Fetch and collect swh objects.
 
         """
-        self.pypi_releases = self.project.releases()
-
-    def store_data(self):
-        """(override) This collects the necessary objects information and send
-           them to storage.
-
-        """
-        _snapshot = {
+        self._snapshot = {
             'branches': {}
         }
+        self._contents = []
+        self._directories = []
+        self._revisions = []
+        self._releases = []
 
-        _contents = []
-        _directories = []
-        _revisions = []
-        _releases = []
-
-        for version, _release in self.pypi_releases:
+        for version, _release in self.project.releases():
             info = self.project.info(version)
             author = self.project.author(version)
             logging.debug('author: %s' % author)
@@ -127,8 +119,8 @@ class PyPiLoader(SWHLoader):
             directory = Directory.from_disk(path=_dir_path, data=True)
             _objects = directory.collect()
 
-            _contents.extend(_objects['content'].values())
-            _directories.extend(_objects['directory'].values())
+            self._contents.extend(_objects['content'].values())
+            self._directories.extend(_objects['directory'].values())
 
             date = normalize_timestamp(
                 int(arrow.get(release['date']).timestamp))
@@ -153,7 +145,7 @@ class PyPiLoader(SWHLoader):
             }
             _revision['id'] = identifier_to_bytes(
                 revision_identifier(_revision))
-            _revisions.append(_revision)
+            self._revisions.append(_revision)
 
             _release = {
                 'name': name,
@@ -166,18 +158,22 @@ class PyPiLoader(SWHLoader):
             }
             _release['id'] = identifier_to_bytes(
                 release_identifier(_release))
-            _releases.append(_release)
+            self._releases.append(_release)
 
-            _snapshot['branches'][name] = {
+            self._snapshot['branches'][name] = {
                 'target': _release['id'],
                 'target_type': 'release',
             }
 
-        _snapshot['id'] = identifier_to_bytes(
-            snapshot_identifier(_snapshot))
+        self._snapshot['id'] = identifier_to_bytes(
+            snapshot_identifier(self._snapshot))
+
+    def store_data(self):
+        """(override) This sends collected objects to storage.
 
-        self.maybe_load_contents(_contents)
-        self.maybe_load_directories(_directories)
-        self.maybe_load_revisions(_revisions)
-        self.maybe_load_releases(_releases)
-        self.maybe_load_snapshot(_snapshot)
+        """
+        self.maybe_load_contents(self._contents)
+        self.maybe_load_directories(self._directories)
+        self.maybe_load_revisions(self._revisions)
+        self.maybe_load_releases(self._releases)
+        self.maybe_load_snapshot(self._snapshot)