Hi Joachim,
We are looking into this, for now i have a solution that can help you
move forward with your work. Once you have created your mart using
MBuilder (I am assuming you are using ensembl or vega schemas) therefore
your mart tables would follow same naming conventions as Ensembl or
Vega. Lets assume you have used vega databases to create a mart.
Now Connect MartEditor to vega_mart_53 on martdb.ensembl.org:5316 and do
'Save All'.
This will give you 3 XMLs with full fledge configuration.
Now connect MartEditor to your mart, eg mini_mart_53. do 'Upload All'
and multiSelect all the xml of vega_mart_53 you downloaded in previous
step. Export.
now your perl bin/configure...... should work just fine.
You can add more species using MartEditor's Alt+N option to existing
template and you may also slim down your xmls (using MartEditor) to suit
your needs.
Hope this is helpful,
Syed
Joachim Baran wrote:
Hello,
2009/8/17 Syed Haider <[email protected]>:
I see that that your xmls are created by lots of manual editing and
somewhere during the process the links between template and configs are
lost, usually MartEditor is supposed to take care of this if the XMLs are
created automatically by the tool. Anyways, since you only have 4 datasets,
could you please share with us the mart dump if its of reasonable size or
can we access your mart db otherwise ?
This week the sys-admins do maintenance on the database server, so I
cannot really provide you a dump right now. However, I have tried to
build a very small mart on my local machine without any manual editing
at all and I am getting the same problem.
From homo_sapiens_vega_53_36o and mus_musculus_vega_53_37f I have
derived the two smaller schemas homo_sapiens_mini_53_36o and
mus_musculus_mini_53_37f, respectively, with only the gene table
copied from the vega databases. Then I created a dataset for the gene
table in MartBuilder, contructed and executed the SQL, and run through
the following steps in MartEditor:
1. Naive
2. Export
3. Alt+N, pick the other species, click on 'No', select the previously added set
4. Export
Just to verify that there are really two datasets in the mart, I did
another import in MartEditor and made sure that this is the case.
After that I ran perl bin/configure.pl --clean -r conf/miniMart.xml
which gives me the following output:
onnection parameters of [Mini Mart 53 (Office)] [ OK ]
[NEW CONFIGURATION] .... WITH MEMORY [default]
default ... Mini Mart 53 (Office) .... 001/002 ... mmusculus_gene
............................ (RDBMS) localhost:3306 ................
OK
default ... Mini Mart 53 (Office) .... 002/002 ... hsapiens_gene
............................. (RDBMS) localhost:3306 ................
OK
Setting possible links between datasets
....(scanning) 100%
....(linking) 100%
....(sorting) 100%
....(clustering) 2/2 - 0 remain
....(resolving) 100%
Building templates for visible datasets
.... 100%
Compiling templates for visible datasets
[1/1] Attribute Panel of Dataset.. :hsapiens_gene
[1/1] Filter Panel of Dataset..... :hsapiens_gene
In the web-interface, I still only see hsapiens_gene. It also says
'ERROR: requested template missing', but that seems to be a different
problem that only occurs once I have selected hsapiens_gene.
I am sending you the XML and ZIPed SQL-files as an attachment, because
I am not quite sure which of these steps is wrong or what I am missing
out. As far as I understand it, I should be able to see two datasets
in the web-interface now.
Thanks,
Joachim
