I'm following the instructions at http://genome.ucsc.edu/admin/cvs.html since my Linux server has IBM Power PC chips.
However, I received a compiler error when running the "make cgi" command.
Could some one help me fix this bug?
itialized -I../inc -I../../inc -I../../../inc -I../../../../inc -I../../../../../inc -I/usr/include/mysql -o customFactory.o -c customFactory.c
cc1: warnings being treated as errors
customFactory.c: In function ‘wigLoaderEncoding’:
customFactory.c:1325: warning: comparison is always true due to limited range of data type
make: *** [customFactory.o] Error 1
make: Leaving directory `/home/dmccully/kent/src/hg/lib'
make: *** [hgLib] Error 2
Contractor, McCully Systems, NIAMS, NIH
i had little difficulty setting up the balanced proxies and ldap auth,
thanks to your wiki and mailing list. but now i'm running into an odd
problem i'm hope you fellas can help shine some light on...
i set up multiple galaxy sites running as different users (in the same
group), with each site submitting to a different sge cluster (based on
users sge env vars), but all use the same postgresql db and a common
i get an error when i try to add a dataset in the library admin view.
however i can add folders, no problem.
Error - <class 'sqlalchemy.orm.exc.UnmappedInstanceError'>: Class
'__builtin__.NoneType' is not mapped
line 143 in __call__
line 80 in __call__
line 102 in __call__
line 632 in __call__
line 126 in __call__
line 803 in upload_library_dataset
line 127 in do
line 925 in refresh
UnmappedInstanceError: Class '__builtin__.NoneType' is not mapped
when i use the upload tool, it submits without error ("Your upload has
been queued..."), the files appears in the database/files folder
(permissions okay), but no history item is created and no error
message appears in the logs.
any hints or suggestions would be greatly appreciated!
do you foresee any problems with the setup i'm trying? what about
collisions of job ids since all sites use a common db and each site
uses a different sge cluster?
does this part of my universe_wsgi.ini.main.job look right? i'm not
sure what the implications of these settings are.
use = egg:PasteDeploy#prefix
prefix = /main
filter-with = proxy-prefix
cookie_path = /main
cookie_prefix = main
i have similar files for other sites besides 'main' (e.g. test, etc.)
It's difficult to tell what is going on because I can't access any of the URLs, and the error message is from a fairly low-level function with many callers (udcMustRead). It is trying to read bytes from the local sparse file cache, and if it fails, you get that error message. Most of its callers are in header-parsing functions. But if you can read the bigWig file that you fetch with wget, then the header is not corrupted. The first read that we attempt when opening a big* file is a 4-byte magic number, so it could be that our udc/net code is not able to get any bytes from your internal URL.
If you can run hgTracks in gdb and send us a backtrace, that could help. The process is something like this:
1. Cause the failure using an hgTracks GET url and copy the params that follow the ?. E.g. if the failing URL is http://host..../hgTracks?hgsid=123&db=etc , copy "hgsid=123&db=etc".
2. cd /usr/local/apache/cgi-bin [or wherever hgTracks is installed]
3. export HGDB_CONF=hg.conf
4. gdb hgTracks
If your hgTracks URL doesn't have any CGI params following a ?, let us know how you reach the failing page.
Hope that helps,
----- "Davide Cittaro" <davide.cittaro(a)ifom-ieo-campus.it> wrote:
> From: "Davide Cittaro" <davide.cittaro(a)ifom-ieo-campus.it>
> To: genome-mirror(a)soe.ucsc.edu, galaxy-dev(a)bx.psu.edu
> Sent: Friday, May 28, 2010 2:43:39 AM GMT -08:00 US/Canada Pacific
> Subject: [Genome-mirror] Enabling bigWig for galaxy
> Dear all,
> I'm trying to enable bigwig support in Galaxy. Besides the sniff
> section which is probably not working, I'm trying to enable at least
> the visualization on a UCSC mirror. I see that galaxy link to display
> includes this hgt.customText
> which contains the track definition:
> track type=bigWig name="Pu1_100219.bw"
> Unfortunately the genome browser (a local mirror) says:
> Error udc couldn't read 4 bytes from
> did read 0
> I've tried to fetch the bigwig file in bigDataUrl link with wget and
> I'm able to reach it.
> Needless to say, the bigwig file is a proper one, I've tried to load
> it from another location and it can be displayed.
> At this point I don't know how to debug it...
> @UCSC people: when am I supposed to get that error? Which conditions
> raise it?
> @Galaxy people: since the link is valid, why I cannot present it to
> Davide Cittaro
> Cogentech - Consortium for Genomic Technologies
> via adamello, 16
> 20139 Milano
> tel.: +39(02)574303007
> e-mail: davide.cittaro(a)ifom-ieo-campus.it
> Genome-mirror mailing list
I've done with plugging our local UCSC mirror into galaxy. I'm able to view most of the data except for BAM and bigWig files.
I'm adding bigWig as new datatype as we mostly work with that. Besides the fact I'm not able to sniff properly the bigWig data type (but I still can define the type at import time), I'm not able to view in UCSC.
Essentially I've copied the ucsc/bam.xml file and edited to enable bigWig tracks.
The problem now is that BAM (or bigWig) can't be viewed by the UCSC because of authentication issues :-(
The UCSC mirrors says that this address
generated by galaxy cannot be viewed (auth required)... Indeed this URL format differs from the one defined for bed files which is in the form of HOST:PORT/root/display_as?PARAMS. The latter has a special configuration in my apache conf file (as suggested by Nate and others), but the same conf for
doesn't work (I believe this depends on /lib/galaxy/web/framework/middleware/remoteuser.py implementation).
Cogentech - Consortium for Genomic Technologies
via adamello, 16
I'm trying to enable bigwig support in Galaxy. Besides the sniff section which is probably not working, I'm trying to enable at least the visualization on a UCSC mirror. I see that galaxy link to display includes this hgt.customText
which contains the track definition:
track type=bigWig name="Pu1_100219.bw" bigDataUrl=http://host001.instruments.ifom-ieo-campus.it:8080/display_application/e00d4fc8bba8b6c5/ucsc_bigWig/campus/a3e0990f3de110b5/data/galaxy_e00d4fc8bba8b6c5.bw db=hg18
Unfortunately the genome browser (a local mirror) says:
Error udc couldn't read 4 bytes from http://host001.instruments.ifom-ieo-campus.it:8080/display_application/e0..., did read 0
I've tried to fetch the bigwig file in bigDataUrl link with wget and I'm able to reach it.
Needless to say, the bigwig file is a proper one, I've tried to load it from another location and it can be displayed.
At this point I don't know how to debug it...
@UCSC people: when am I supposed to get that error? Which conditions raise it?
@Galaxy people: since the link is valid, why I cannot present it to UCSC?
Cogentech - Consortium for Genomic Technologies
via adamello, 16
We do not currently have a best way to directly execute JARs in Galaxy. It does not appear that setting $classpath or $path, etc, will allow the direct execution of jars without providing the path or when not found in the current working directory.
One simple possibility is for Galaxy to allow a configurable directory where required JARs can dropped / linked to; this variable would then be available for use in tool configurations. Do you have any suggestions/ideas/thoughts?
Thanks for using Galaxy,
On May 6, 2010, at 3:59 AM, Nils Homer wrote:
> What is the best way to include wrap a JAR file in galaxy? Also, what is
> the best way to submit tools and their associated wrappers?
> Thank-you for your help in advance,
> Nils Homer
> galaxy-user mailing list
On May 26, 2010, at 10:39 AM, Davide Cittaro wrote:
> On May 26, 2010, at 4:30 PM, Greg Von Kuster wrote:
> So, to summarize:
> 1- a bai file is created when adding a bam to a dataset into data libraries
> 2- the same bai is referenced when importing from a data library to a history
> 3- if I create/modify/update a bam file already in history a bai file is automagically created as metadata of the new one (right?)
> what if something went wrong at step 1 and I don't have bai file? how can I recreate it?
Click the pencil icon in the history item and click "Auto-detect" in the Edit Attributes form to re-generate the metadata.
Greg Von Kuster
Galaxy Development Team
I am trying to write a tool whose input and output types are defined from an
option selected by the user.
The first input is defined from a select list using a loc file - no probs
<param name="source_select" type="select" label="Option 1">
<column name="name" index="0"/>
<column name="value" index="1"/>
<column name="inputType" index="2"/>
<column name="outType" index="3"/>
Based on the selection I want to define the allowed input types (above in
column 3) and specify the output type (above in column 4).
Are the values of the select columns available for use in
conditionals/change_format operations? If not, how can I get around it? What
other methods of attack are there for the problem.
On a related issue, if th user specifies their input file, can i access the
metadata for its type and filter the select list based on the input file?
Two minor issues with user registration:
1. If a user registers with an invalid public name (e.g. "A. Gordon 12345") a confusing red error message is shown:
"User name must contain only lower-case letters, numbers and '-'"
It's confusing (IMHO) because a "user name" is usually associated with the name you're going to login with (which is the email in this case). The message should say "Public name must contain only...".
It'll also be nice to remove the word "user" from the "Public User name" label. "User name" is one thing, "Public name" is another. "Public User Name" is confusing.
2. After seeing the red warning about a bad public name AND fixing the name, clicking "submit" still gives the same red error message about a bad public user name.
Step to reproduce:
1. Go to main galaxy web site, select "Register" from the "user" menu.
2. Fill the fields with:
confirm password: 123456
public user name: This.Is.an%Invalid-12345-Public-Name
3. Click "Submit"
you'll get a red warning about an invalid user name.
4. change the public name to: ThisIsAValidPubLicName
5. Click "submit"
you'll still get a red warning about an invalid user name.
The only way around this is to clear the public name field, and leave it empty when registering.