Merge branch 'develop' of github.com:matrix-org/synapse into erikj/admin_api_cmd

1 files changed, 23 insertions, 87 deletions

diff --git a/synapse/handlers/admin.py b/synapse/handlers/admin.py
index e424fc46bd..e8a651e231 100644
--- a/synapse/handlers/admin.py
+++ b/synapse/handlers/admin.py
@@ -14,10 +14,6 @@
 # limitations under the License.
 
 import logging
-import os
-import tempfile
-
-from canonicaljson import json
 
 from twisted.internet import defer
 
@@ -99,15 +95,16 @@ class AdminHandler(BaseHandler):
         defer.returnValue(ret)
 
     @defer.inlineCallbacks
-    def exfiltrate_user_data(self, user_id, writer):
-        """Write all data we have of the user to the specified directory.
+    def export_user_data(self, user_id, writer):
+        """Write all data we have on the user to the given writer.
 
         Args:
             user_id (str)
             writer (ExfiltrationWriter)
 
         Returns:
-            defer.Deferred
+            defer.Deferred: Resolves when all data for a user has been written.
+            The returned value is that returned by `writer.finished()`.
         """
         # Get all rooms the user is in or has been in
         rooms = yield self.store.get_rooms_for_user_where_membership_is(
@@ -134,7 +131,7 @@ class AdminHandler(BaseHandler):
 
             forgotten = yield self.store.did_forget(user_id, room_id)
             if forgotten:
-                logger.info("[%s] User forgot room %d, ignoring", room_id)
+                logger.info("[%s] User forgot room %d, ignoring", user_id, room_id)
                 continue
 
             if room_id not in rooms_user_has_been_in:
@@ -172,13 +169,14 @@ class AdminHandler(BaseHandler):
             # dict[str, set[str]].
             event_to_unseen_prevs = {}
 
-            # The reverse mapping to above, i.e. map from unseen event to parent
-            # events. dict[str, set[str]]
-            unseen_event_to_parents = {}
+            # The reverse mapping to above, i.e. map from unseen event to events
+            # that have the unseen event in their prev_events, i.e. the unseen
+            # events "children". dict[str, set[str]]
+            unseen_to_child_events = {}
 
             # We fetch events in the room the user could see by fetching *all*
             # events that we have and then filtering, this isn't the most
-            # efficient method perhaps but it does guarentee we get everything.
+            # efficient method perhaps but it does guarantee we get everything.
             while True:
                 events, _ = yield self.store.paginate_room_events(
                     room_id, from_key, to_key, limit=100, direction="f"
@@ -200,16 +198,14 @@ class AdminHandler(BaseHandler):
                     if unseen_events:
                         event_to_unseen_prevs[event.event_id] = unseen_events
                         for unseen in unseen_events:
-                            unseen_event_to_parents.setdefault(unseen, set()).add(
+                            unseen_to_child_events.setdefault(unseen, set()).add(
                                 event.event_id
                             )
 
                     # Now check if this event is an unseen prev event, if so
                     # then we remove this event from the appropriate dicts.
-                    for event_id in unseen_event_to_parents.pop(event.event_id, []):
-                        event_to_unseen_prevs.get(event_id, set()).discard(
-                            event.event_id
-                        )
+                    for child_id in unseen_to_child_events.pop(event.event_id, []):
+                        event_to_unseen_prevs[child_id].discard(event.event_id)
 
                     written_events.add(event.event_id)
 
@@ -233,7 +229,7 @@ class AdminHandler(BaseHandler):
 
 
 class ExfiltrationWriter(object):
-    """Interfaced used to specify how to write exfilrated data.
+    """Interface used to specify how to write exported data.
     """
 
     def write_events(self, room_id, events):
@@ -254,7 +250,7 @@ class ExfiltrationWriter(object):
         Args:
             room_id (str)
             event_id (str)
-            state (list[FrozenEvent])
+            state (dict[tuple[str, str], FrozenEvent])
         """
         pass
 
@@ -263,76 +259,16 @@ class ExfiltrationWriter(object):
 
         Args:
             room_id (str)
-            invite (FrozenEvent)
-            state (list[dict]): A subset of the state at the invite, with a
-                subset of the event keys (type, state_key, content and sender)
+            event (FrozenEvent)
+            state (dict[tuple[str, str], dict]): A subset of the state at the
+                invite, with a subset of the event keys (type, state_key
+                content and sender)
         """
 
     def finished(self):
-        """Called when exfiltration is complete, and the return valus is passed
-        to the requester.
+        """Called when all data has succesfully been exported and written.
+
+        This functions return value is passed to the caller of
+        `export_user_data`.
         """
         pass
-
-
-class FileExfiltrationWriter(ExfiltrationWriter):
-    """An ExfiltrationWriter that writes the users data to a directory.
-
-    Returns the directory location on completion.
-
-    Args:
-        user_id (str): The user whose data is being exfiltrated.
-        directory (str|None): The directory to write the data to, if None then
-            will write to a temporary directory.
-    """
-
-    def __init__(self, user_id, directory=None):
-        self.user_id = user_id
-
-        if directory:
-            self.base_directory = directory
-        else:
-            self.base_directory = tempfile.mkdtemp(
-                prefix="synapse-exfiltrate__%s__" % (user_id,)
-            )
-
-        os.makedirs(self.base_directory, exist_ok=True)
-        if list(os.listdir(self.base_directory)):
-            raise Exception("Directory must be empty")
-
-    def write_events(self, room_id, events):
-        room_directory = os.path.join(self.base_directory, "rooms", room_id)
-        os.makedirs(room_directory, exist_ok=True)
-        events_file = os.path.join(room_directory, "events")
-
-        with open(events_file, "a") as f:
-            for event in events:
-                print(json.dumps(event.get_pdu_json()), file=f)
-
-    def write_state(self, room_id, event_id, state):
-        room_directory = os.path.join(self.base_directory, "rooms", room_id)
-        state_directory = os.path.join(room_directory, "state")
-        os.makedirs(state_directory, exist_ok=True)
-
-        event_file = os.path.join(state_directory, event_id)
-
-        with open(event_file, "a") as f:
-            for event in state.values():
-                print(json.dumps(event.get_pdu_json()), file=f)
-
-    def write_invite(self, room_id, event, state):
-        self.write_events(room_id, [event])
-
-        # We write the invite state somewhere else as they aren't full events
-        # and are only a subset of the state at the event.
-        room_directory = os.path.join(self.base_directory, "rooms", room_id)
-        os.makedirs(room_directory, exist_ok=True)
-
-        invite_state = os.path.join(room_directory, "invite_state")
-
-        with open(invite_state, "a") as f:
-            for event in state.values():
-                print(json.dumps(event), file=f)
-
-    def finished(self):
-        return self.base_directory