[Libosinfo] [PATCH osinfo-db 06/10] tests: util: Cache xml file list and parsed OS objects

[Cole Robinson]

These should be reusable across test cases. Cache the lookup results on
the first access

Signed-off-by: Cole Robinson <crobinso at redhat.com>
---
 tests/util.py | 42 ++++++++++++++++++++++++++----------------
 1 file changed, 26 insertions(+), 16 deletions(-)

diff --git a/tests/util.py b/tests/util.py
index 15ec3cc..418b9d2 100644
--- a/tests/util.py
+++ b/tests/util.py
@@ -15,31 +15,41 @@ class _DataFiles():
     def __init__(self):
         self.datadir = os.environ['INTERNAL_OSINFO_DB_DATA_DIR']
         self.schema = os.path.join(self.datadir, 'schema', 'osinfo.rng')
+        self._all_xml_cache = []
+        self._oses_cache = []
 
         if not os.path.exists(self.datadir):
             raise RuntimeError("INTERNAL_OSINFO_DB_DATA_DIR=%s "
                 "doesn't exist" % self.datadir)
 
-    def _get_files(self, directory):
-        files = []
-        root = os.path.join(self.datadir, directory)
-        for (dirpath, _, filenames) in os.walk(root):
-            for filename in filenames:
-                if not filename.endswith('.xml'):
-                    continue
-                files.append(os.path.join(dirpath, filename))
-        return files
+    def _get_all_xml(self):
+        """
+        Get and cache the full list of all DATA_DIR .xml paths
+        """
+        if not self._all_xml_cache:
+            for (dirpath, _, filenames) in os.walk(self.datadir):
+                for filename in filenames:
+                    if not filename.endswith('.xml'):
+                        continue
+                    self._all_xml_cache.append(os.path.join(dirpath, filename))
+        return self._all_xml_cache
+
+    def _filter_xml(self, dirname):
+        """
+        Filter XML paths by those in $DATA_DIR/$dirname
+        """
+        return [p for p in self._get_all_xml() if
+                p.startswith(os.path.join(self.datadir, dirname))]
 
     def oses(self):
-        ret = []
-        files = self._get_files('os')
-        for path in files:
-            osroot = ET.parse(path).getroot().find('os')
-            ret.append(osinfo.Os(osroot))
-        return ret
+        if not self._oses_cache:
+            for path in self._filter_xml('os'):
+                osroot = ET.parse(path).getroot().find('os')
+                self._oses_cache.append(osinfo.Os(osroot))
+        return self._oses_cache
 
     def xmls(self):
-        return self._get_files('')
+        return self._get_all_xml()
 
 
 DataFiles = _DataFiles()
-- 
2.21.0