Bed file generation is here!

 * Added a new makebed function which returns a generator rather
   than writing to outstream.
 * Updated the make_description function in makebed.py to be able
   to handle cases like '<flowcell_id> (deleted)'.
 * Added a bedfile generation view which uses the makebed generator
   function to prevent memory issues on the webserver.
 * Visit <url>/library/ for trying out the new features!

diff --git a/gaworkflow/frontend/fctracker/views.py b/gaworkflow/frontend/fctracker/views.py
index 7f117c9c3fc2988df0c172c2f7b6a73cbdaa2a49..05358e969adb34335e9259a0e03fbc0756308f46 100644 (file)
--- a/gaworkflow/frontend/fctracker/views.py
+++ b/gaworkflow/frontend/fctracker/views.py
@@ -1,9 +1,13 @@
 # Create your views here.
 from gaworkflow.frontend.fctracker.models import Library
 from gaworkflow.frontend.fctracker.results import get_flowcell_result_dict, flowcellIdStrip
+from gaworkflow.frontend import settings
+from gaworkflow.util import makebed
 from gaworkflow.util import opener
 from django.http import HttpResponse
 
+import StringIO
+
 #from django.db.models import base 
 
 def library(request):
@@ -80,8 +84,43 @@ def result_fc_cnm_eland_lane(request, fc_id, cnm, lane):
     f = opener.autoopen(filepath, 'r')
     
     return HttpResponse(f)
+
+
+def bedfile_fc_cnm_eland_lane(request, fc_id, cnm, lane):
+    """
+    returns a bed file for a given flowcell, CN-M (i.e. C1-33), and lane
+    """
+    fc_id = flowcellIdStrip(fc_id)
+    d = get_flowcell_result_dict(fc_id)
+    
+    if d is None:
+        return HttpResponse('<b>Results for Flowcell %s not found.' % (fc_id))
+    
+    if cnm not in d:
+        return HttpResponse('<b>Results for Flowcell %s; %s not found.' % (fc_id, cnm))
     
+    erd = d[cnm]['eland_results']
+    lane = int(lane)
+    
+    if lane not in erd:
+        return HttpResponse('<b>Results for Flowcell %s; %s; lane %s not found.' % (fc_id, cnm, lane))
+    
+    filepath = erd[lane]
+    
+    # Eland result file
+    fi = open(filepath, 'r')
+    # output memory file
+    
+    
+    name, description = makebed.make_description(settings.DATABASE_NAME,
+                                                 fc_id,
+                                                 lane)
+    
+    bedgen = makebed.make_bed_from_eland_stream_generator(fi, name, description)
+    
+    return HttpResponse(bedgen)
 
+    
 def _files(flowcell_id, lane):
     """
     Sets up available files for download
@@ -100,6 +139,7 @@ def _files(flowcell_id, lane):
         
         if int(lane) in erd:
             output.append('<a href="/results/%s/%s/eland_result/%s">eland_result(%s)</a>' % (flowcell_id, c_name, lane, c_name))
+            output.append('<a href="/results/%s/%s/bedfile/%s">bedfile(%s)</a>' % (flowcell_id, c_name, lane, c_name))
     
     if len(output) == 0:
         return ''

diff --git a/gaworkflow/frontend/urls.py b/gaworkflow/frontend/urls.py
index b359c33e4d97d9a0b8b3ed7b2972e62ba1a1237c..add92790f2e24a31759f0de514077451ffcfb7cb 100644 (file)
--- a/gaworkflow/frontend/urls.py
+++ b/gaworkflow/frontend/urls.py
@@ -15,5 +15,6 @@ urlpatterns = patterns('',
      #(r'^databrowse/(.*)', databrowse.site.root),
      (r'^library/$', 'gaworkflow.frontend.fctracker.views.library'),
      (r'^library/(?P<lib_id>\w+)/$', 'gaworkflow.frontend.fctracker.views.library_to_flowcells'),
-     (r'^results/(?P<fc_id>\w+)/(?P<cnm>C[1-9]-[0-9]+)/eland_result/(?P<lane>[1-8])','gaworkflow.frontend.fctracker.views.result_fc_cnm_eland_lane')
+     (r'^results/(?P<fc_id>\w+)/(?P<cnm>C[1-9]-[0-9]+)/eland_result/(?P<lane>[1-8])','gaworkflow.frontend.fctracker.views.result_fc_cnm_eland_lane'),
+     (r'^results/(?P<fc_id>\w+)/(?P<cnm>C[1-9]-[0-9]+)/bedfile/(?P<lane>[1-8])','gaworkflow.frontend.fctracker.views.bedfile_fc_cnm_eland_lane'),
 )

diff --git a/gaworkflow/util/makebed.py b/gaworkflow/util/makebed.py
index 738b815b270d2d8352d6157a7b5e1948b437e3a3..4d73cb6cbcad46b73481529a311ae86dd926ec3d 100755 (executable)
--- a/gaworkflow/util/makebed.py
+++ b/gaworkflow/util/makebed.py
@@ -41,6 +41,45 @@ def make_bed_from_eland_stream(instream, outstream, name, description, chromosom
       sense_color[fields[SENSE]],
       os.linesep  
     ))
+    
+def make_bed_from_eland_stream_generator(instream, name, description, chromosome_prefix='chr'):
+  """
+  read an eland result file from instream and output it as a generator (iterator)
+  """
+  # indexes into fields in eland_result.txt file
+  SEQ = 1
+  CHR = 6
+  START = 7
+  SENSE = 8
+  # map eland_result.txt sense 
+  sense_map = { 'F': '+', 'R': '-'}
+  sense_color = { 'F': '0,0,255', 'R': '255,255,0' }
+  # provide default track names
+  if name is None: name = "track"
+  if description is None: description = "eland result file"
+  bed_header = 'track name="%s" description="%s" visibility=4 itemRgb="ON"'
+  bed_header += os.linesep
+  yield bed_header % (name, description)
+
+  for line in instream:
+    fields = line.split()
+    # we need more than the CHR field, and it needs to match a chromosome
+    if len(fields) <= CHR or \
+          (chromosome_prefix is not None and \
+             fields[CHR][:3] != chromosome_prefix):
+      continue
+    start = fields[START]
+    stop = int(start) + len(fields[SEQ])
+    chromosome, extension = fields[CHR].split('.')
+    assert extension == "fa"
+    yield '%s %s %d read 0 %s - - %s%s' % (
+      chromosome,
+      start,
+      stop,
+      sense_map[fields[SENSE]], 
+      sense_color[fields[SENSE]],
+      os.linesep  
+    )
 
 def make_description(database, flowcell_id, lane):
     """
@@ -49,7 +88,7 @@ def make_description(database, flowcell_id, lane):
     from gaworkflow.util.fctracker import fctracker
 
     fc = fctracker(database)
-    cells = fc._get_flowcells("where flowcell_id='%s'" % (flowcell_id))
+    cells = fc._get_flowcells("where flowcell_id LIKE '%s%%'" % (flowcell_id))
     if len(cells) != 1:
       raise RuntimeError("couldn't find flowcell id %s" % (flowcell_id))
     lane = int(lane)
@@ -59,7 +98,11 @@ def make_description(database, flowcell_id, lane):
     name = "%s-%s" % (flowcell_id, lane)
 
     cell_id, cell = cells.items()[0]
-    assert cell_id == flowcell_id
+    
+    #The assertion is no longer true after I changed
+    # the where statement to include the LIKE command.
+    # because flowcells are being renamed to 'FC12269 (deleted)'
+    #assert cell_id == flowcell_id
 
     cell_library_id = cell['lane_%d_library_id' %(lane,)]
     cell_library = cell['lane_%d_library' %(lane,)]