Mercurial > repos > peterjc > mira_assembler

changeset 1:c947750f82fb draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
Uploaded v0.0.5
author peterjc
date Wed, 24 Apr 2013 11:51:50 -0400
parents d0a5acdf1638
children 8bddbb4b2575
files tools/sr_assembly/mira.py tools/sr_assembly/mira.txt tools/sr_assembly/mira.xml
diffstat 3 files changed, 127 insertions(+), 57 deletions(-) [+]
line wrap: on
line diff
--- a/tools/sr_assembly/mira.py	Tue Jun 07 15:47:17 2011 -0400
+++ b/tools/sr_assembly/mira.py	Wed Apr 24 11:51:50 2013 -0400
@@ -7,40 +7,57 @@
 import shutil
 import time
 
+WRAPPER_VER = "0.0.5" #Keep in sync with the XML file
+
 def stop_err(msg, err=1):
     sys.stderr.write(msg+"\n")
     sys.exit(err)
 
-def tcs_to_tabular(old, new):
-    in_handle = open(old, "rU")
-    out_handle = open(new, "w")
-    assert in_handle.readline() == "#TCS V1.0\n"
-    assert in_handle.readline() == "#\n"
-    assert in_handle.readline() == "# contig name          padPos  upadPos | B  Q | tcov covA covC covG covT cov* | qA qC qG qT q* |  S | Tags\n"
-    assert in_handle.readline() == "#\n"
-    out_handle.write("#%s\n" % "\t".join(["contig", "pasPos", "upadPos", "B", "Q",
-                                         "tcov", "covA", "covC", "covG", "covT", "cov*",
-                                         "qA", "qC", "qG", "qT", "q*", "S", "Tags"]))
-    for line in in_handle:
-        parts = line.rstrip("\n").split(None,22)
-        assert parts[3] == parts[6] == parts[13] == parts[19] == parts[21] == "|"
-        wanted = parts[:3] + parts[4:6]+parts[7:13]+parts[14:19]+parts[20:21]+parts[22:]
-        out_handle.write("%s\n" % "\t".join(wanted))
-    out_handle.close()
-    in_handle.close()
+
+def get_version():
+    """Run MIRA to find its version number"""
+    # At the commend line I would use: mira -v | head -n 1
+    # however there is some pipe error when doing that here.
+    try:
+        child = subprocess.Popen(["mira", "-v"],
+                                 stdout=subprocess.PIPE,
+                                 stderr=subprocess.STDOUT)
+    except Exception, err:
+        sys.stderr.write("Error invoking command:\n%s\n\n%s\n" % (" ".join(cmd), err))
+        sys.exit(1)
+    ver, tmp = child.communicate()
+    del child
+    return ver.split("\n", 1)[0]
+
+
+mira_ver = get_version()
+if "V3.4." not in mira_ver:
+    stop_err("This wrapper is for MIRA V3.4, not %s" % mira_ver)
+if "-v" in sys.argv:
+    print "MIRA wrapper version %s," % WRAPPER_VER
+    print mira_ver
+    sys.exit(0)
+
 
 def collect_output(temp, name):
     n3 = (temp, name, name, name)
-    tcs_to_tabular("%s/%s_assembly/%s_d_results/%s_out.tcs" % n3, out_tcs)
-    for old, new in [("%s/%s_assembly/%s_d_results/%s_out.unpadded.fasta" % n3, out_fasta),
-                     ("%s/%s_assembly/%s_d_results/%s_out.unpadded.fasta.qual" % n3, out_qual),
-                     ("%s/%s_assembly/%s_d_results/%s_out.wig" % n3, out_wig),
-                     ("%s/%s_assembly/%s_d_results/%s_out.caf" % n3, out_caf),
-                     ("%s/%s_assembly/%s_d_results/%s_out.ace" % n3, out_ace)]:
+    f = "%s/%s_assembly/%s_d_results" % (temp, name, name)
+    if not os.path.isdir(f):
+        stop_err("Missing output folder")
+    if not os.listdir(f):
+        stop_err("Empty output folder")
+    missing = []
+    for old, new in [("%s/%s_out.unpadded.fasta" % (f, name), out_fasta),
+                     ("%s/%s_out.unpadded.fasta.qual" % (f, name), out_qual),
+                     ("%s/%s_out.wig" % (f, name), out_wig),
+                     ("%s/%s_out.caf" % (f, name), out_caf),
+                     ("%s/%s_out.ace" % (f, name), out_ace)]:
         if not os.path.isfile(old):
-            stop_err("Missing %s output file" % os.path.splitext(old)[-1])
+            missing.append(os.path.splitext(old)[-1])
         else:
             shutil.move(old, new)
+    if missing:
+        stop_err("Missing output files: %s" % ", ".join(missing))
 
 def clean_up(temp, name):
     folder = "%s/%s_assembly" % (temp, name)
@@ -48,22 +65,45 @@
         shutil.rmtree(folder)
 
 #TODO - Run MIRA in /tmp or a configurable directory?
+#Currently Galaxy puts us somewhere safe like:
+#/opt/galaxy-dist/database/job_working_directory/846/
 temp = "."
-name, out_fasta, out_qual, out_tcs, out_ace, out_caf, out_wig, out_log = sys.argv[1:9]
+name, out_fasta, out_qual, out_ace, out_caf, out_wig, out_log = sys.argv[1:8]
+
 start_time = time.time()
+cmd_list =sys.argv[8:]
+cmd = " ".join(cmd_list)
+
+assert os.path.isdir(temp)
+d = "%s_assembly" % name
+assert not os.path.isdir(d), "Path %s already exists" % d
 try:
-    cmd = " ".join(sys.argv[9:])
-    child = subprocess.Popen(sys.argv[9:],
-                             stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+    #Check path access
+    os.mkdir(d)
+except Exception, err:
+    sys.stderr.write("Error making directory %s\n%s" % (d, err))
+    sys.exit(1)
+
+#print os.path.abspath(".")
+#print cmd
+
+handle = open(out_log, "w")
+try:
+    #Run MIRA
+    child = subprocess.Popen(cmd_list,
+                             stdout=handle,
+                             stderr=subprocess.STDOUT)
 except Exception, err:
     sys.stderr.write("Error invoking command:\n%s\n\n%s\n" % (cmd, err))
+    #TODO - call clean up?
+    handle.write("Error invoking command:\n%s\n\n%s\n" % (cmd, err))
+    handle.close()
     sys.exit(1)
 #Use .communicate as can get deadlocks with .wait(),
 stdout, stderr = child.communicate()
+assert not stdout and not stderr #Should be empty as sent to handle
 run_time = time.time() - start_time
 return_code = child.returncode
-handle = open(out_log, "w")
-handle.write(stdout)
 handle.write("\n\nMIRA took %0.2f minutes\n" % (run_time / 60.0))
 print "MIRA took %0.2f minutes" % (run_time / 60.0)
 if return_code:
@@ -75,6 +115,10 @@
              return_code)
 handle.close()
 
+#print "Collecting output..."
 collect_output(temp, name)
+
+#print "Cleaning up..."
 clean_up(temp, name)
+
 print "Done"
--- a/tools/sr_assembly/mira.txt	Tue Jun 07 15:47:17 2011 -0400
+++ b/tools/sr_assembly/mira.txt	Wed Apr 24 11:51:50 2013 -0400
@@ -7,7 +7,7 @@
 
 This tool is a short Python script (to collect the MIRA output and move it
 to where Galaxy expects the files, and convert MIRA's TCS file into a tab
-separate file for use in Galaxy). There are just two files to install:
+separated file for use in Galaxy). There are just two files to install:
 
 * mira.py (the Python script)
 * mira.xml (the Galaxy tool definition)
@@ -16,9 +16,9 @@
 modify the tools_conf.xml file to tell Galaxy to offer the tool and also do
 this to tools_conf.xml.sample in order to run any tests:
 
-<tool file="sr_assembly/seq_primer_clip.xml" />
+<tool file="sr_assembly/mira.xml" />
 
-You will also need to install MIRA, we used version 3.2.1. See:
+You will also need to install MIRA, we used version 3.4.0. See:
 
 http://chevreux.org/projects_mira.html
 http://sourceforge.net/projects/mira-assembler/
@@ -33,7 +33,17 @@
 History
 =======
 
-v0.0.1 - Initial version (working prototype)
+v0.0.1 - Initial version (working prototype, using MIRA 3.2.1)
+v0.0.2 - Improve capture of stdout/stderr (should see it as it runs)
+v0.0.3 - Support Ion Torrent reads, now requires MIRA 3.4.0 or later
+         (some other switches changed, e.g. -OUT rrol to rrot, which
+         means the wrapper no longer works with MIRA 3.2.x)
+       - The contig summary file (TCS file) was removed in MIRA 3.4
+       - Report all missing output files (not just first missing one)
+v0.0.4 - Fix problem with backbone arguments inroduced in v0.0.3
+v0.0.5 - Implement the <version_command> tag to record the wrapper
+         version and the MIRA version being used.
+       - Check using MIRA 3.4 (later versions have a different API)
 
 
 Developers
--- a/tools/sr_assembly/mira.xml	Tue Jun 07 15:47:17 2011 -0400
+++ b/tools/sr_assembly/mira.xml	Wed Apr 24 11:51:50 2013 -0400
@@ -1,40 +1,47 @@
-<tool id="mira_assembler" name="Assemble with MIRA" version="0.0.1">
-    <description>Takes Sanger, Roche, and Illumina data</description>
-	<command interpreter="python">mira.py mira $out_fasta $out_qual $out_tcs $out_ace $out_caf $out_wig $out_log
+<tool id="mira_assembler" name="Assemble with MIRA" version="0.0.5">
+    <description>Takes Sanger, Roche, Illumina, and Ion Torrent data</description>
+	<version_command interpreter="python">mira.py -v</version_command>
+	<command interpreter="python">mira.py mira $out_fasta $out_qual $out_ace $out_caf $out_wig $out_log
 ##Give the wrapper script list of output filenames, then the mira command...
 mira --job=$job_method,$job_type,$job_quality
 
 ##Input files
 #if $condBackbone.use == "true":
     ## Can this be linked to job_method as well? If mapping we need the backbone...
-    -SB:lb=yes -SB:bft=fasta -FN:bbin=${condBackbone.filename}
+    -SB:lb=1:bft=fasta -FN:bbin=${condBackbone.filename}
 #end if
 #if $condSanger.use == "true":
-    Sanger_SETTINGS
-    ## Not easy to add sanger to --job, so use load_sequence_data(lsd) instead
-    -LR:lsd=yes
+    SANGER_SETTINGS
+    ## Not easy in Galaxy to add sanger to --job, so use load_sequence_data(lsd) instead
     ## I expect hard trimmed FASTQ files with no NCBI traceinfo XML file
-    -LR:mxti=no -LR:ft=fastq -FN:fqi=${condSanger.filename}
+    -LR:lsd=1:mxti=0:ft=fastq -FN:fqi=${condSanger.filename}
 #end if
 #if $condRoche.use == "true":
     454_SETTINGS
-    ## Not easy to add 454 to --job, so use load_sequence_data(lsd) instead
-    -LR:lsd=yes
+    ## Not easy in Galaxy to add 454 to --job, so use load_sequence_data(lsd) instead
     ## I expect hard trimmed FASTQ files with no NCBI traceinfo XML file
-    -LR:mxti=no -LR:ft=fastq -FN:fqi=${condRoche.filename}
+    -LR:lsd=1:mxti=0:ft=fastq -FN:fqi=${condRoche.filename}
 #end if
 #if $condIllumina.use == "true":
     SOLEXA_SETTINGS
-    ## Not easy to add solexa to --job, so use load_sequence_data(lsd) instead
-    -LR:lsd=yes -LR:ft=fastq -FN:fgi=${condIllumina.filename}
+    ## Not easy in Galaxy to add solexa to --job, so use load_sequence_data(lsd) instead
+    -LR:lsd=1:ft=fastq -FN:fqi=${condIllumina.filename}
     ##TODO - Look at -LR FASTQ qual offset (fqqo)
 #end if
-
+#if $condIonTorrent.use == "true":
+    IONTOR_SETTINGS
+    ## Not easy in Galaxy to add iontor to --job, so use load_sequence_data(lsd) instead
+    ## I expect hard trimmed FASTQ files with no NCBI traceinfo XML file
+    -LR:lsd=1:mxti=0:ft=fastq -FN:fqi=${condIonTorrent.filename}
+#end if
 
 ##Output files
 COMMON_SETTINGS
-##remove_rollover_logs, remove_log_directory
--OUT:rrol=yes -OUT:rld=yes
+##Explicitly request the FASTA (+QUAL), CAF, ACE, WIG output
+##Explicitly disable formats we won't use like MAF (reduce IO)
+-OUT:orf=1:orc=1:ora=1:orw=1:orm=0:org=0:ors=0
+##remove_rollover_tmps, remove_tmp_directory
+-OUT:rrot=1:rtd=1
 
     </command>
 	<inputs>
@@ -47,9 +54,9 @@
             <option value="est">EST (transcriptome)</option>
         </param>
         <param name="job_quality" type="select" label="Assembly quality grade">
-            <option value="normal">Normal</option>
+            <option value="accurate">Accurate</option>
+            <option value="normal">Normal (deprecated)</option>
             <option value="draft">Draft</option>
-            <option value="accurate">Accurate</option>
         </param>
         <!-- Backbone -->
         <conditional name="condBackbone">
@@ -82,7 +89,8 @@
            </param>
            <when value="false" />
            <when value="true">
-              <param name="filename" type="data" format="fastq,sff" label="Roche 454 reads file" help="FASTQ format" />
+              <!-- TODO? Support SFF files directly, e.g. with sff_extract, but will need linker sequences -->
+              <param name="filename" type="data" format="fastq" label="Roche 454 reads file" help="FASTQ format" />
            </when>
         </conditional>
         <!-- Illumina -->
@@ -96,11 +104,22 @@
               <param name="filename" type="data" format="fastq" label="Solexa/Illumina reads file" help="FASTQ format" />
            </when>
         </conditional>
+        <!-- Ion Torrent -->
+        <conditional name="condIonTorrent">
+           <param name="use" type="select" label="Ion Torrent reads?">
+               <option value="false">No</option>
+               <option value="true">Yes</option>
+           </param>
+           <when value="false" />
+           <when value="true">
+              <!-- TODO? Support SFF files directly, e.g. with sff_extract -->
+              <param name="filename" type="data" format="fastq" label="Ion Torrent reads file" help="FASTQ format" />
+           </when>
+        </conditional>
 	</inputs>
 	<outputs>
 	    <data name="out_fasta" format="fasta" label="MIRA contigs (FASTA)" />
 	    <data name="out_qual" format="qual454" label="MIRA contigs (QUAL)" />
-	    <data name="out_tcs" format="tabular" label="MIRA contigs summary" />
 	    <data name="out_caf" format="txt" label="MIRA contigs (CAF)" />
 	    <data name="out_ace" format="txt" label="MIRA contigs (ACE)" />
 	    <data name="out_wig" format="wig" label="MIRA coverage (Wiggle)" />
@@ -117,9 +136,6 @@
 
 Runs MIRA v3, collects the output, and throws away all the temporary files.
 
-The MIRA transposed contig summary (TCS) file is converted into a tabular file for use within Galaxy.
-This records one line per base per contig, and including things like the base, quality, coverage and any tags.
-
 **Citation**
 
 This tool uses MIRA. If you use this tool in scientific work leading to a