[apollo] [Apollo 2.0.4] Annotations reportedly disappearing from postgresql database

[Juhana Kammonen]

Dear all,

We are having a manual annotation project of an eykaryote organism. Circa 30
people are participating and I am coordinating the effort. We are running
Apollo 2.0.4 and One person reported that her annotation had disappeared from
the annotation database and another person had later annotated the same
feature. She gave very good details on the location and times of editing.

I looked on the Apollo admin view (Report::Changes) and saw no other edits for
the reported location than those done by the later person. Thus, it seemed
that the location had actually not been annotated before.

I proceeded to dig the information in the postgresql database which we use for
our annotation database. Sure enough, the previous edits in the location were
visible in the database (table feature_events). These were followed by events
of the later person doing annotations in the location. There was no indication
of anybody deleting any information in that location or other errors of the
database.

I suspected that this "vanishing" of the annotation may have had something to
do with the length of the comments inserted into the comment field of the
annotation information. The person who reported this vanishing has written
very long comments and I was wondering whether the comment field has some kind
of limitation for length (varchar) and whether this would have somehow
invalidated the previous annotation.

The annotation postgresql database is extremely abundant in information, so I
would like to know what is the proper place where to trace any errors or
deletion events for features.

Based on the total number of annotations and those of individual annotators,
these events appear as highly unusual for our effort.

Kind regards,
Juhana K
University of Helsinki
Finland

[Nathan Dunn]

We did have a bug where this was a problem:

https://github.com/GMOD/Apollo/issues/963 <https://github.com/GMOD/Apollo/issues/963>

https://github.com/GMOD/Apollo/pull/964/files <https://github.com/GMOD/Apollo/pull/964/files>

Its possible that for some reason the database isn’t reflecting these changes, but you should be able to patch this manually and hopefully this will re-appear.

A couple of questions:

1 - in postgreql, can you list the schema on your table \d+ feature_property
2 - if the table is varchar(256), you should be able to alter it directly to text:
ALTER TABLE feature_property ALTER COLUMN value TYPE text;
If it was already text and this doesn’t fix it can you check your tomcat logs (often called catalina.out) and post the errors as well as any javascript console errors?

Thanks,

Nathan Dunn, PhD
Berkeley Bioinformatics Open-source Projects (BBOP)
Environmental Genomics and Systems Biology, Lawrence Berkeley National Laboratory

Post by Juhana Kammonen
Dear all,
We are having a manual annotation project of an eykaryote organism. Circa 30
people are participating and I am coordinating the effort. We are running
Apollo 2.0.4 and One person reported that her annotation had disappeared from
the annotation database and another person had later annotated the same
feature. She gave very good details on the location and times of editing.
I looked on the Apollo admin view (Report::Changes) and saw no other edits for
the reported location than those done by the later person. Thus, it seemed
that the location had actually not been annotated before.
I proceeded to dig the information in the postgresql database which we use for
our annotation database. Sure enough, the previous edits in the location were
visible in the database (table feature_events). These were followed by events
of the later person doing annotations in the location. There was no indication
of anybody deleting any information in that location or other errors of the
database.
I suspected that this "vanishing" of the annotation may have had something to
do with the length of the comments inserted into the comment field of the
annotation information. The person who reported this vanishing has written
very long comments and I was wondering whether the comment field has some kind
of limitation for length (varchar) and whether this would have somehow
invalidated the previous annotation.
The annotation postgresql database is extremely abundant in information, so I
would like to know what is the proper place where to trace any errors or
deletion events for features.
Based on the total number of annotations and those of individual annotators,
these events appear as highly unusual for our effort.
Kind regards,
Juhana K
University of Helsinki
Finland
This list is for the Apollo Annotation Editing Tool. Info at http://genomearchitect.org/