Implement a simpler format than JSON for the metadata. Should be much faster.

1 files changed, 31 insertions, 23 deletions

diff --git a/src/olpc/datastore/metadatastore.py b/src/olpc/datastore/metadatastore.py
index f15f1af..112b693 100644
--- a/src/olpc/datastore/metadatastore.py
+++ b/src/olpc/datastore/metadatastore.py
@@ -1,12 +1,6 @@
 import os
 import logging
-
-try:
-    import cjson
-    has_cjson = True
-except ImportError:
-    import simplejson
-    has_cjson = False
+import urllib
 
 from olpc.datastore import layoutmanager
 
@@ -16,15 +10,18 @@ class MetadataStore(object):
     def store(self, uid, metadata):
         metadata = metadata.copy()
 
+        for key in metadata.keys():
+            if ' ' in key:
+                raise ValueError('Property names cannot include spaces. '
+                                 'Wrong name: %r' % key)
+
         dir_path = layoutmanager.get_instance().get_entry_path(uid)
         if not os.path.exists(dir_path):
             os.makedirs(dir_path)
 
         for key, value in metadata.items():
-            if isinstance(value, str) and len(value) > MAX_SIZE:
-                self._write_external(key, value, dir_path)
-                del metadata[key]
-            elif isinstance(value, str) and not self._is_unicode(value):
+            if isinstance(value, str) and \
+                    (len(value) > MAX_SIZE or not self._is_unicode(value)):
                 self._write_external(key, value, dir_path)
                 del metadata[key]
 
@@ -96,19 +93,30 @@ class MetadataStore(object):
             return False
 
     def _encode(self, metadata, file_path):
-        if has_cjson:
-            f = open(file_path, 'w')
-            f.write(cjson.encode(metadata))
-            f.close()
-        else:
-            simplejson.dump(metadata, open(file_path, 'w'))
+        f = open(file_path, 'w')
+        for key, value in metadata.items():
+            if value is None:
+                value = ''
+            f.write('%s %s\n' % (key, urllib.quote(str(value))))
+        f.close()
 
     def _decode(self, file_path):
-        if has_cjson:
-            f = open(file_path, 'r')
-            metadata = cjson.decode(f.read())
-            f.close()
-            return metadata
+        f = open(file_path, 'r')
+        metadata = {}
+        for line in f.readlines():
+            logging.debug(line)
+            key, value = line.split(' ', 1)
+            value = value[:-1] # Take out the trailing '\n'
+            value = self._cast_for_journal(key, urllib.unquote(value))
+            metadata[key] = value
+        f.close()
+        return metadata
+
+    def _cast_for_journal(self, key, value):
+        # Hack because the current journal expects these properties to have some
+        # predefined types
+        if key in ['timestamp', 'keep']:
+            return int(value)
         else:
-            return simplejson.load(open(file_path))
+            return value