IPAR detection is more reliable than firecrest so do it first, and then

[htsworkflow.git] / htsworkflow / pipelines / runfolder.py

diff --git a/htsworkflow/pipelines/runfolder.py b/htsworkflow/pipelines/runfolder.py
index f327b7891868034ca38c8327bdeed5ed2efdd135..5d663b3d9924890097cd4219b254fdf8339f0f9c 100644 (file)
--- a/htsworkflow/pipelines/runfolder.py
+++ b/htsworkflow/pipelines/runfolder.py
@@ -136,6 +136,19 @@ class PipelineRun(object):
         tree = ElementTree.parse(filename).getroot()
         self.set_elements(tree)
 
+def load_pipeline_run_xml(pathname):
+    """
+    Load and instantiate a Pipeline run from a run xml file
+
+    :Parameters: 
+      - `pathname` : location of an run xml file
+
+    :Returns: initialized PipelineRun object
+    """
+    tree = ElementTree.parse(pathname).getroot()
+    run = PipelineRun(xml=tree)
+    return run
+
 def get_runs(runfolder):
     """
     Search through a run folder for all the various sub component runs
@@ -168,7 +181,7 @@ def get_runs(runfolder):
                     p.gerald = g
                     runs.append(p)
                 except IOError, e:
-                    print "Ignoring", str(e)
+                    logging.error("Ignoring " + str(e))
 
     datadir = os.path.join(runfolder, 'Data')
 
@@ -178,15 +191,81 @@ def get_runs(runfolder):
     for firecrest_pathname in glob(os.path.join(datadir,"*Firecrest*")):
         logging.info('Found firecrest in ' + datadir)
         image_analysis = firecrest.firecrest(firecrest_pathname)
-        scan_post_image_analysis(runs, runfolder, image_analysis, firecrest_pathname)
+        if image_analysis is None:
+           logging.warn(
+                "%s is an empty or invalid firecrest directory" % (firecrest_pathname,)
+            )
+       else:
+            scan_post_image_analysis(
+                runs, runfolder, image_analysis, firecrest_pathname
+            )
     # scan for IPAR directories
     for ipar_pathname in glob(os.path.join(datadir,"IPAR_*")):
         logging.info('Found ipar directories in ' + datadir)
         image_analysis = ipar.ipar(ipar_pathname)
-        scan_post_image_analysis(runs, runfolder, image_analysis, ipar_pathname)
+        if image_analysis is None:
+           logging.warn(
+                "%s is an empty or invalid IPAR directory" %(ipar_pathname,)
+            )
+       else:
+            scan_post_image_analysis(
+                runs, runfolder, image_analysis, ipar_pathname
+            )
 
     return runs
 
+def get_specific_run(gerald_dir):
+    """
+    Given a gerald directory, construct a PipelineRun out of its parents
+
+    Basically this allows specifying a particular run instead of the previous
+    get_runs which scans a runfolder for various combinations of
+    firecrest/ipar/bustard/gerald runs.
+    """
+    from htsworkflow.pipelines import firecrest
+    from htsworkflow.pipelines import ipar
+    from htsworkflow.pipelines import bustard
+    from htsworkflow.pipelines import gerald
+
+    bustard_dir = os.path.abspath(os.path.join(gerald_dir, '..'))
+    image_dir = os.path.abspath(os.path.join(gerald_dir, '..', '..'))
+
+    runfolder_dir = os.path.abspath(os.path.join(image_dir, '..','..'))
+   
+    logging.debug('--- use-run detected options ---')
+    logging.debug('runfolder: %s' % (runfolder_dir,))
+    logging.debug('image_dir: %s' % (image_dir,))
+    logging.debug('bustard_dir: %s' % (bustard_dir,))
+    logging.debug('gerald_dir: %s' % (gerald_dir,))
+
+    # find our processed image dir
+    image_run = ipar.ipar(image_dir)
+    if image_run is None:
+        image_run = firecrest.firecrest(image_dir)
+    if image_run is None:
+        msg = '%s does not contain an image processing step' % (image_dir,)
+        logging.error(msg)
+        return None
+
+    # find our base calling
+    base_calling_run = bustard.bustard(bustard_dir)
+    if base_calling_run is None:
+        logging.error('%s does not contain a bustard run' % (bustard_dir,))
+        return None
+
+    # find alignments
+    gerald_run = gerald.gerald(gerald_dir)
+    if gerald_run is None:
+        logging.error('%s does not contain a gerald run' % (gerald_dir,))
+        return None
+
+    p = PipelineRun(runfolder_dir)
+    p.image_analysis = image_run
+    p.bustard = base_calling_run
+    p.gerald = gerald_run
+    
+    logging.info('Constructed PipelineRun from %s' % (gerald_dir,))
+    return p
 
 def extract_run_parameters(runs):
     """
@@ -195,7 +274,7 @@ def extract_run_parameters(runs):
     for run in runs:
       run.save()
 
-def summarize_mapped_reads(mapped_reads):
+def summarize_mapped_reads(genome_map, mapped_reads):
     """
     Summarize per chromosome reads into a genome count
     But handle spike-in/contamination symlinks seperately.
@@ -205,7 +284,7 @@ def summarize_mapped_reads(mapped_reads):
     genome = 'unknown'
     for k, v in mapped_reads.items():
         path, k = os.path.split(k)
-        if len(path) > 0:
+        if len(path) > 0 and not genome_map.has_key(path):
             genome = path
             genome_reads += v
         else:
@@ -236,7 +315,7 @@ def summarize_lane(gerald, lane_id):
       report.append('Repeat (0,1,2 mismatches) %d %d %d' % \
                     (mc['R0'], mc['R1'], mc['R2']))
       report.append("Mapped Reads")
-      mapped_reads = summarize_mapped_reads(eland_result.mapped_reads)
+      mapped_reads = summarize_mapped_reads(eland_result.genome_map, eland_result.mapped_reads)
       for name, counts in mapped_reads.items():
         report.append("  %s: %d" % (name, counts))
       report.append('')
@@ -294,6 +373,8 @@ def extract_results(runs, output_base_dir=None):
       # save run file
       r.save(cycle_dir)
 
+      return
+
       # Copy Summary.htm
       summary_path = os.path.join(r.gerald.pathname, 'Summary.htm')
       if os.path.exists(summary_path):