Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
Vahid
@VJalili
@jmchilton raising exceptions such as RequestParameterMissingException in places other than api/controller causes 500 error. Any thoughts?
Marius van den Beek
@mvdbeek
do you have an example ?
I think MessageExceptions are caught in the ErrorMiddleware, so if they don't leave galaxy that makes sense
Vahid
@VJalili
e.g.,: https://github.com/VJalili/galaxy/blob/e1d014939fe752cfcf99debc60e5756efe211fbf/lib/galaxy/authnz/managers.py#L141
but when raised in api, the api caller receives a correct error message and code: https://github.com/VJalili/galaxy/blob/e1d014939fe752cfcf99debc60e5756efe211fbf/lib/galaxy/webapps/galaxy/api/cloudauthz.py#L106
Marius van den Beek
@mvdbeek
Sure, that makes sense, that's what the ErrorMiddleware does
but if you call this from somewhere where it doesn't pass through the middleware it won't do that
at least that's my understanding
Vahid
@VJalili
Sounds reasonable. Any suggestion for throwing exceptions behind the middleware? (e.g., in the authnz I linked)
Marius van den Beek
@mvdbeek
huh, that depends on who needs to react to the exception I guess
Vahid
@VJalili
I used to return errors to the caller api/controller. However, I think we're moving from that practice.
its the caller of API that needs to react
Marius van den Beek
@mvdbeek
like this is still going through the HTTP API ?
Vahid
@VJalili
correct
Marius van den Beek
@mvdbeek
then it should actually be caught
Vahid
@VJalili
caught in api/controller?
Marius van den Beek
@mvdbeek
https://github.com/galaxyproject/galaxy/blob/9b4de72ca68afc2353beff333775422039baae25/lib/galaxy/web/framework/middleware/error.py#L160
I think
wrong line, https://github.com/galaxyproject/galaxy/blob/9b4de72ca68afc2353beff333775422039baae25/lib/galaxy/web/framework/middleware/error.py#L166
I'm not entirely sure on this, I just recently looked into this for the py3 porting
sorry, gotta run
Vahid
@VJalili
good hint anyway, thanks.
Slugger70
@Slugger70

Hey there! I'm running Galaxy 18.05 on a server, I have a tool that creates some javascript html output which I have whitelisted. I still get

javascript is needed to display data.
If you try to view this data on galaxy please contact your administrator to authorise javascript or download the file to view.

If I donwload the file it displays no problem.. I checked the whitelist config file and the tool that produced it is listed there. FastQC works fine.

Nicola Soranzo
@nsoranzo
Does the tool id have "strange" characters?
Slugger70
@Slugger70
It has Capital letters and an underscore. "-" is in the name
And the tool id doesn't match the xml name.
Marius van den Beek
@mvdbeek
if you guys are ever debugging a bug that happens occasionally and you can't reproduce it by looping over the test it might be worth setting PYTHONHASHSEED to a couple of different values, and see if the error becomes reproducible. Definitely helped narrow down 2 bugs with random dictionary order
wish I'd known that months ago ...
scholtalbers
@scholtalbers
hey all I'm writing a test (as requested by @bgruening .. ;)) for bismark. The first tool of the bunch generates a qname_input_sorted.bam and the subsequent tools will require that as input. However, it seems galaxy will not accept this format for upload(?) - i.e. i don't know how to write a test that has this param: <param name="mapping_output" value="mapped_reads.bam" ftype="qname_input_sorted.bam"/> - the error message I get is: `An invalid option was selected for file_type qname_input_sorted please verify'
Marius van den Beek
@mvdbeek
yeah -- that's on purpose. set the uploaded filetype to qnamed_sorted.bam. qname_input_sorted.bam is for aligners that were run in galaxy and that output the reads in the order that they were read in. This isn't the official "SO:queryname", but qname_sorted.bam is compatible with qname_input_sorted.bam , so there will be no additional conversion
scholtalbers
@scholtalbers
so that would be the tool input parameter to accept format="qname_sorted.bam,qname_input_sorted.bam" or just format="qname_sorted.bam"? and for the test the input file will also be typed with format "qname_sorted.bam"?
Marius van den Beek
@mvdbeek
just qname_input_sorted.bam as parameter type (qname_sorted.bam subclasses qname_input_sorted.bam, so that would be the same as qname_sorted.bam,qname_input_sorted.bam). Correct, the test file should be qname_sorted.bam
it's a little tricky unfortunately, but that's the best that avoids unnecessary conversions while still being correct :/
scholtalbers
@scholtalbers
it is indeed, but thanks for the explanation - lets see if i can get the tests right
no luck: https://github.com/bgruening/galaxytools/blob/a8a3c74fcb07bbd57f3026048341ba01b8cbea67/tools/bismark/bismark_deduplicate/bismark_deduplicate_wrapper.xml
Marius van den Beek
@mvdbeek
in what sense ?
scholtalbers
@scholtalbers
ah sorry, still get the invalid input - just a sec and I can paste the error message
Marius van den Beek
@mvdbeek
https://travis-ci.org/bgruening/galaxytools/builds/418236067?utm_source=github_status&utm_medium=notification seems to have passed ?
scholtalbers
@scholtalbers
u"An invalid option was selected for file_type, u'qname_sorted.bam', please verify."
Marius van den Beek
@mvdbeek
is that locally ?
(I think the PR might be missing a .shed.yml file, nothing got tested)
scholtalbers
@scholtalbers
yeah indeed, cause the other tests are not finished yet and should be throwing errors in any case
its not a new set of tools, they are existing, the shed.yml should not be updated?
Marius van den Beek
@mvdbeek
is ~/.planemo/gx_repo up to date ?
sometimes it get's stuck on an old galaxy -- just rm it and try again ?
scholtalbers
@scholtalbers
might be will try
Marius van den Beek
@mvdbeek
for me it fails like this:
---------------------- >> begin tool stdout << -----------------------
Deduplicating with: 'deduplicate_bismark -s submitted_reads.bam --bam'
Processing single-end Bismark output file(s) (SAM format):
submitted_reads.bam


If there are several alignments to a single position in the genome the first alignment will be chosen. Since the input files are not in any way sorted this is a near-enough random selection of reads.

Checking file >>submitted_reads.bam<< for signs of file truncation...

samtools view: writing to standard output failed: Broken pipe
samtools view: error closing standard output: -1

Output file is: submitted_reads.deduplicated.bam


Total number of alignments analysed in submitted_reads.bam:    554
Total number duplicated alignments removed:    50 (9.03%)
Duplicated alignments were found at:    44 different position(s)

Total count of deduplicated leftover sequences: 504 (90.97% of total)

skipping header line:    @HD    VN:1.0    SO:queryname
skipping header line:    @SQ    SN:chrY_JH584300_random    LN:182347
skipping header line:    @SQ    SN:chrY_JH584301_random    LN:259875
skipping header line:    @SQ    SN:chrY_JH584302_random    LN:155838
skipping header line:    @SQ    SN:chrY_JH584303_random    LN:158099
skipping header line:    @PG    ID:Bismark    VN:v0.19.1    CL:"bismark --bam --gzip --temp_dir /var/folders/cf/k7c5rjhj1sggk6x8z1jnw4b80000gp/T/tmpn4m609_i -o /var/folders/cf/k7c5rjhj1sggk6x8z1jnw4b80000gp/T/tmpn4m609_i/results --quiet --fastq -L 20 -D 15 -R 2 --un --ambiguous /var/folders/cf/k7c5rjhj1sggk6x8z1jnw4b80000gp/T/tmpe8nkca8w /private/var/folders/cf/k7c5rjhj1sggk6x8z1jnw4b80000gp/T/tmpDiq_J8/files/000/dataset_2.dat"

----------------------- >> end tool stdout << ------------------------

---------------------- >> begin tool stderr << -----------------------

----------------------- >> end tool stderr << ------------------------

FAIL
which from the datatype perspective looks OK. might even just be an issue with how samtools view is used inside bismark
scholtalbers
@scholtalbers
yes thanks @mvdbeek it was the galaxy version
Marius van den Beek
@mvdbeek
cool, good luck!