Store each metadata property in its own file

1 files changed, 31 insertions, 104 deletions

diff --git a/src/olpc/datastore/metadatastore.py b/src/olpc/datastore/metadatastore.py
index 5b607be..810ccb8 100644
--- a/src/olpc/datastore/metadatastore.py
+++ b/src/olpc/datastore/metadatastore.py
@@ -1,5 +1,6 @@
 import os
 import logging
+import errno
 
 from olpc.datastore import layoutmanager
 
@@ -7,25 +8,21 @@ MAX_SIZE = 256
 
 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 or not self._is_unicode(value)):
-                self._write_external(key, value, dir_path)
-                del metadata[key]
+        metadata_path = os.path.join(dir_path, 'metadata')
+        if not os.path.exists(metadata_path):
+            os.makedirs(metadata_path)
 
         metadata['uid'] = uid
-        self._encode(metadata, os.path.join(dir_path, 'metadata'))
+        for key, value in metadata.items():
+            f = open(os.path.join(metadata_path, key), 'w+')
+            try:
+                f.write(str(value))
+            finally:
+                f.close()
 
     def retrieve(self, uid, properties=None):
         dir_path = layoutmanager.get_instance().get_entry_path(uid)
@@ -33,103 +30,33 @@ class MetadataStore(object):
             raise ValueError('Unknown object: %r' % uid)
 
         metadata_path = os.path.join(dir_path, 'metadata')
-        metadata = self._decode(metadata_path, properties)
+        metadata = {}
+        if properties is None or not properties:
+            properties = os.listdir(metadata_path)
 
-        if properties is None or len(properties) != len(metadata):
-            extra_metadata_dir = os.path.join(dir_path, 'extra_metadata')
-            if os.path.exists(extra_metadata_dir):
-                for key in os.listdir(extra_metadata_dir):
-                    if properties is not None and key not in properties:
-                        continue
-                    file_path = os.path.join(extra_metadata_dir, key)
-                    if os.path.exists(file_path):
-                        # TODO: This class shouldn't know anything about dbus.
-                        import dbus
-                        metadata[key] = dbus.ByteArray(open(file_path).read())
+        for key in properties:
+            property_path = metadata_path + '/' + key
+            try:
+                value = open(property_path, 'r').read()
+            except IOError, e:
+                if e.errno != errno.ENOENT:
+                    raise
+            else:
+                if not value:
+                    metadata[key] = ''
+                else:
+                    # TODO: This class shouldn't know anything about dbus.
+                    import dbus
+                    metadata[key] = dbus.ByteArray(value)
+            
         return metadata
 
     def delete(self, uid):
         dir_path = layoutmanager.get_instance().get_entry_path(uid)
         metadata_path = os.path.join(dir_path, 'metadata')
-        if os.path.isfile(metadata_path):
-            os.remove(os.path.join(dir_path, 'metadata'))
-        else:
-            logging.warning('%s is not a valid path' % metadata_path)
-
-        extra_metadata_path = os.path.join(dir_path, 'extra_metadata')
-        if os.path.isdir(extra_metadata_path):
-            for key in os.listdir(extra_metadata_path):
-                os.remove(os.path.join(extra_metadata_path, key))
-            os.rmdir(os.path.join(dir_path, 'extra_metadata'))
-        else:
-            logging.warning('%s is not a valid path' % extra_metadata_path)
-
-    def _write_external(self, key, value, dir_path):
-        extra_metadata_dir = os.path.join(dir_path, 'extra_metadata')
-        if not os.path.exists(extra_metadata_dir):
-            os.makedirs(extra_metadata_dir)
-        f = open(os.path.join(extra_metadata_dir, key), 'w')
-        f.write(value)
-        f.close()
-
-    def _is_unicode(self, string):
-        try:
-            string.decode('utf-8')
-            return True
-        except UnicodeDecodeError:
-            return False
-
-    def _encode(self, metadata, file_path):
-        f = open(file_path, 'w')
-        f.write('v1\n')
-        for key, value in metadata.items():
-            if not key:
-                raise ValueError('Property keys cannot be empty')
-            if (' ' in key) or ('\t' in key):
-                raise ValueError('Property keys cannot contain tabulators: %r' \
-                                 % key)
-            if value is None:
-                value = ''
-            else:
-                value = str(value)
-            f.write('%s\t%d\t%s\n' % (key, len(value), value))
-        f.close()
-
-    def _decode(self, file_path, properties):
-        f = open(file_path, 'r')
-        version_line = f.readline()
-        try:
-            version = int(version_line[1:-1])
-            if version != 1:
-                raise ValueError('Incompatible version %r' % version)
-        except:
-            logging.error('Invalid version line: %s' % version_line)
-            raise
-
-        metadata = {}
-        while True:
-            line = f.readline()
-            if not line:
-                break
-
-            key, value_len, value = line.split('\t', 2)
-            value_len = int(value_len)
-
-            if len(value) == value_len + 1:
-                value = value[:-1] # skip the newline
-            else:
-                value += f.read(value_len - len(value))
-                f.seek(1, 1) # skip the newline
-
-            if properties is None:
-                metadata[key] = value
-            elif key in properties:
-                metadata[key] = value
-                if len(properties) == len(metadata):
-                    break
-
-        f.close()
-        return metadata
+        for key in os.listdir(metadata_path):
+            os.remove(os.path.join(metadata_path, key))
+        os.rmdir(metadata_path)
 
     def _cast_for_journal(self, key, value):
         # Hack because the current journal expects these properties to have some