Databanks management
Directory structure housing the banks
As indicated in the section Directory structure, all databanks installed by BeeDeeM are located in the directory ${biobaseRootDir}.
Lets take the example bank descriptors PDB_proteins, Uniprot_Swissprot and Refseq_Viruses; these are '.dsc' suffixed files located in ${installDir}/conf/descriptors
directory of BeeDeeM.
These banks will be installed into the directories described by their respective bank descriptor. If you look at these descriptors (e.g. ${installDir}/conf/descriptors/PDB_proteins.dsc) and you look directive db.ldir, this is what you will see:
Descriptor
Installation instruction
Type
Bank name
PDB_proteins.dsc
db.ldir=${mirrordir}
p
PDB_proteins
Uniprot_SwissProt.dsc
db.ldir=${mirrordir}
p
Uniprot_SwissProt
Refseq_Viruses.dsc
db.ldir=${mirrordir}
n
Refseq_Viruses
All of these 'db.ldir', 'type' and 'bank name' directives translate into the following directory tree:
It is worth noting that reserved descriptor keyword ${mirrordir} translates to ${biobaseRootDir} at runtime.
This example shows that in NO CASE the various databases are installed in the same directory. If this is not done, BeeDeeM may malfunction, particularly during updates. In effect, if two databases are installed in the same directory, you cannot update them individually.
To install a new database, it is best to comply with the tree structure given above.
For example, if you wan to install the BLAST nucleic database 'MyBank', install it in its own directory under:
In the same way, if you want to install the flat file proteic database 'MyOtherBank', install it in its own directory under:
What to do after processing
When the installation of a sequence database was successful, it is automatically installed into production for your users.
The “old” copy of the database is always deleted, but this behaviour can be controlled using parameter 'history' of a databank descriptor.
Let's take the example from the database descriptor "PDB_proteins.dsc" located in ${installDir}/conf/descriptors
. Here, we are interested in two lines of this file:
the first line gives the name of the database to install and the second line specifies where it should be installed on your system. Therefore, these two lines tell BeeDeeM that the database PDB (sequences only) will be installed in the directory ${mirrordir}|p|PDB_proteins under the name PDB_proteins.
When the processing of this database has succeeded, PDB_proteins is installed into the following directory:
Note the presence of the term current in the path: it identifies the PDB_proteins database currently in production, i.e. the version of the database used by your users.
Now, let us suppose that you ask BeeDeeM to update this database as of 26/02/2022. Throughout the processing of the database PDB, BeeDeeM will work in a temporary directory named as follows:
Note that the sub-directory download
is at the same level as the sub-directory current
(see below). This means that the database PDB_proteins in production remains available (and cohabits with) the version being installed.
If the installation procedure ends successfully, the directories are renamed as follows ONLY if descriptor parameter 'history' is set to '1':
(the version that was in production migrates out of production). If 'history' is set to '0', then the 'old' release of PDB_proteins is automatically deleted.
If the installation procedure fails, these two directories are unchanged:
In this case, you should consult the BeeDeeM log files and search for the lines containing "WARN" (refer to Control of execution for more details).
Each such "WARN" line contains the description of an error that occurred during processing. Once the problem has been identified, BeeDeeM can be restarted: see section Restart after failure.
Last updated
Was this helpful?