Project

General

Profile

Actions

Bug #1956

open

one-to-one relationship for ApSelectionRun DEF_id to its rundir

Added by Anchi Cheng about 12 years ago. Updated about 12 years ago.

Status:
New
Priority:
Normal
Assignee:
-
Category:
-
Target version:
-
Start date:
07/27/2012
Due date:
% Done:

0%

Estimated time:
Affected Version:
Appion/Leginon 2.2.0 (trunk)
Show in known bugs:
No
Workaround:

Description

Liz got a sinedon error that ends with

File "/usr/lib/python2.4/site-packages/MySQLdb/connections.py", line 35, in defaulterrorhandler
raise errorclass, errorvalue
_mysql_exceptions.IntegrityError: (1062, "Duplicate entry '6813' for key 2")

It turns out that earlier appion database imposes a unique flag on apSelectionData.`REF|ApPath|path`. Gabe's helical stuff wants to add a new selection run in the same path even as kind of subrun. Shall we remove uniqueness constraint? Does anyone use that property in their code , i.e., assumes one-to-one relationship of selectionrun and the rundir? I think the most likely place that this is required is in the workflow where a run needs to be repeated but the subrun in this case can not berepeated in one step, anyway.

I will remove the constraint in the appion database Liz works on (ap90) so that she can move on but can change that back later.

Actions #1

Updated by Gabriel Lander about 12 years ago

Alternatively we could set the new SelectionRun path to None, since no particle files are written or stored anywhere. Only new particle picks for the filled in helical picks are inserted into the database, there is no new rundir is created. Would a "None" for ApPathData be acceptable?

Actions #2

Updated by Scott Stagg about 12 years ago

Gabe's solution sounds good to me. This fits with a feature that I've been wanting for a while. I've been wanting to recenter picks based on an alignment run. That way the particles in the corresponding stack could be as centered as possible.

Actions

Also available in: Atom PDF