Homer Tag Directory Galaxy Composite Dataset

# HG changeset patch # User kevyin # Date 1354001673 18000 # Node ID 74c1fc7bb164d3028f73792e810f67af78c9b6f0 # Parent 4ea55669f00ec4938c73a247a66c33c0bee2006d display tag directories with html diff -r 4ea55669f00e -r 74c1fc7bb164 README --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/README Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,13 @@ +Homer wrapper for Galaxy + +Code repo: https://bitbucket.org/gvl/homer + +=========================================: +LICENSE for this wrapper: +=========================================: +Kevin Ying +Garvan Institute: http://www.garvan.org.au +GVL: https://genome.edu.au/wiki/GVL + +http://opensource.org/licenses/mit-license.php + diff -r 4ea55669f00e -r 74c1fc7bb164 annotatePeaks.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/annotatePeaks.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,41 @@ + + + homer + + + + + annotatePeaks.pl $input_bed $genome_selector 1> $out_annotated + 2> $out_log || echo "Error running annotatePeaks." >&2 + + + + + + + + + + + + + + + + + + + + + + + .. class:: infomark + + **Homer annoatePeaks** + More information on accepted formats + http://biowhat.ucsd.edu/homer/ngs/annotation.html + + + + + diff -r 4ea55669f00e -r 74c1fc7bb164 bed2pos.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/bed2pos.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,34 @@ + + + homer + + + + + bed2pos.pl $input_bed 1> $out_pos + 2> $out_log || echo "Error running bed2pos." >&2 + + + + + + + + + + + + + + + + + + + .. class:: infomark + + **Homer bed2pos.pl** + http://biowhat.ucsd.edu/homer/ngs/miscellaneous.html + + + diff -r 4ea55669f00e -r 74c1fc7bb164 datatypes_conf.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/datatypes_conf.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,9 @@ + + + + + + + + + diff -r 4ea55669f00e -r 74c1fc7bb164 findPeaks.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/findPeaks.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,52 @@ + + + homer + + Homer's peakcaller. Requires tag directories (see makeTagDirectory) + + + findPeaks $tagDir.extra_files_path $options -o $outputPeakFile + + #if $control_tagDir: + -i $control_tagDir.extra_files_path + #end if + + 2> $out_log || echo "Error running findPeaks." >&2 + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + + .. class:: infomark + + **Homer findPeaks** + http://biowhat.ucsd.edu/homer/ngs/peaks.html + + + + diff -r 4ea55669f00e -r 74c1fc7bb164 homerDatatypes.py --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/homerDatatypes.py Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,122 @@ +""" +homer composite datatypes +NOTE: makeTagDirectory doesn't actually use this!!!! It uses html as the format + I've never got it to properly display the contents of the datatype for after + clicking the "eye" icon, so using html instead + +Use at your peril +Kevin Ying +for the homer wrapper + +Dec 2012 +""" + +import logging, os, sys, time, tempfile, shutil, string, glob +import data +from galaxy import util +from cgi import escape +import urllib, binascii +from galaxy.web import url_for +from galaxy.datatypes import metadata +from galaxy.datatypes.metadata import MetadataElement +from galaxy.datatypes.data import Text +from galaxy.datatypes.tabular import Tabular +from galaxy.datatypes.images import Html +from galaxy.datatypes.interval import Interval +from galaxy.util.hash_util import * + +gal_Log = logging.getLogger(__name__) +verbose = False + +class HomerTagDirectory(Html): + """ + base class to use for homer tag directory datatypes + composite datatype elements + stored in extra files path + """ + + #MetadataElement( name="base_name", desc="base name for all transformed versions of this genetic dataset", default='HomerTagDirectory', + #readonly=True, set_in_upload=False) + + composite_type = 'auto_primary_file' + #allow_datatype_change = False + file_ext = 'homerTagDirectory' + + def __init__(self, **kwd): + Html.__init__(self, **kwd) + self.add_composite_file('tagInfo.txt', description="tagInfo") + self.add_composite_file('tagAutocorrelation.txt', description="tagAutocorrelation") + self.add_composite_file('tagCountDistribution.txt', description="tagCountDistribution") + self.add_composite_file('tagLengthDistribution.txt', description="tagLengthDistribution") + + def generate_primary_file(self, dataset=None): + rval = ['Homer Tag Directory Galaxy Composite Dataset

'] + rval.append('

This composite dataset is composed of the following files:

%s (%s)%s
%s%s

' ) + return "\n".join( rval ) + + + def regenerate_primary_file(self,dataset): + """ + cannot do this until we are setting metadata + """ + #bn = dataset.metadata.base_name + efp = dataset.extra_files_path + flist = os.listdir(efp) + rval = ['Files for Composite Dataset %s

Composite %s contains:

%s

' ) + f = file(dataset.file_name,'w') + f.write("\n".join( rval )) + f.write('\n') + f.close() + + def get_mime(self): + """Returns the mime type of the datatype""" + return 'text/html' + + def set_meta( self, dataset, **kwd ): + + """ + for homer maketagdirectory eg + + """ + Html.set_meta( self, dataset, **kwd ) + if kwd.get('overwrite') == False: + if verbose: + gal_Log.debug('@@@ HomerTagDirectory set_meta called with overwrite = False') + return True + try: + efp = dataset.extra_files_path + except: + if verbose: + gal_Log.debug('@@@ HomerTagDirectory set_meta failed %s - dataset %s has no efp ?' % (sys.exc_info()[0], dataset.name)) + return False + try: + flist = os.listdir(efp) + except: + if verbose: gal_Log.debug('@@@ HomerTagDirectory set_meta failed %s - dataset %s has no efp ?' % (sys.exc_info()[0],dataset.name)) + return False + if len(flist) == 0: + if verbose: + gal_Log.debug('@@@ HomerTagDirectory set_meta failed - %s efp %s is empty?' % (dataset.name,efp)) + return False + self.regenerate_primary_file(dataset) + if not dataset.info: + dataset.info = 'Galaxy HomerTagDirectory datatype object' + if not dataset.blurb: + dataset.blurb = 'Composite file - Homer Galaxy toolkit' + return True + diff -r 4ea55669f00e -r 74c1fc7bb164 makeTagDirectory.py --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/makeTagDirectory.py Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,94 @@ +""" + + +""" +import re +import os +import sys +import subprocess +import optparse +import shutil +import tempfile + +def getFileString(fpath, outpath): + """ + format a nice file size string + """ + size = '' + fp = os.path.join(outpath, fpath) + s = '? ?' + if os.path.isfile(fp): + n = float(os.path.getsize(fp)) + if n > 2**20: + size = ' (%1.1f MB)' % (n/2**20) + elif n > 2**10: + size = ' (%1.1f KB)' % (n/2**10) + elif n > 0: + size = ' (%d B)' % (int(n)) + s = '%s %s' % (fpath, size) + return s + +class makeTagDirectory(): + """wrapper + """ + + def __init__(self,opts=None, args=None): + self.opts = opts + self.args = args + + def run_makeTagDirectory(self): + """ + makeTagDirectory [options] [alignment file 2] + + """ + if self.opts.format != "bam": + cl = [self.opts.executable] + args + ["-format" , self.opts.format] + else: + cl = [self.opts.executable] + args + print cl + p = subprocess.Popen(cl) + retval = p.wait() + + + html = self.gen_html(args[0]) + #html = self.gen_html() + return html,retval + + def gen_html(self, dr=os.getcwd()): + flist = os.listdir(dr) + print flist + """ add a list of all files in the tagdirectory + """ + res = ['

Files created by makeTagDirectory

\n'] + + flist.sort() + for i,f in enumerate(flist): + if not(os.path.isdir(f)): + fn = os.path.split(f)[-1] + res.append('\n' % (fn,getFileString(fn, dr))) + + res.append('

\n') + + return res + +if __name__ == '__main__': + op = optparse.OptionParser() + op.add_option('-e', '--executable', default='makeTagDirectory') + op.add_option('-o', '--htmloutput', default=None) + op.add_option('-f', '--format', default="sam") + opts, args = op.parse_args() + #assert os.path.isfile(opts.executable),'## makeTagDirectory.py error - cannot find executable %s' % opts.executable + + #if not os.path.exists(opts.outputdir): + #os.makedirs(opts.outputdir) + f = makeTagDirectory(opts, args) + + html,retval = f.run_makeTagDirectory() + f = open(opts.htmloutput, 'w') + f.write(''.join(html)) + f.close() + if retval <> 0: + print >> sys.stderr, serr # indicate failure + + + diff -r 4ea55669f00e -r 74c1fc7bb164 makeTagDirectory.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/makeTagDirectory.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,50 @@ + + + homer + + Simple wrapper for makeTagDirectory. Used by findPeaks + + makeTagDirectory.py ${tagDir.files_path} + #for $alignF in $alignmentFiles + $alignF.file -f $alignF.file.ext + #end for + -o $tagDir + 2> $out_log || echo "Error running homer_makeTagDirectory." >&2 + + + + + + + + + + + + + + + + + + + + + + + + + + + .. class:: infomark + + After the job finishes, save the zip file, unzip and open "index.html" + + The actual meme-chip command is displayed under "Command line summary" + + Defaults: + http://meme.nbcr.net/meme/doc/meme-chip.html + + + + diff -r 4ea55669f00e -r 74c1fc7bb164 pos2bed.xml --- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/pos2bed.xml Tue Nov 27 02:34:33 2012 -0500 @@ -0,0 +1,34 @@ + + + homer + + + + + pos2bed.pl $input_peak 1> $out_bed + 2> $out_log || echo "Error running pos2bed." >&2 + + + + + + + + + + + + + + + + + + + .. class:: infomark + + **Homer pos2bed.pl** + http://biowhat.ucsd.edu/homer/ngs/miscellaneous.html + + +