Mercurial > repos > iuc > khmer_abundance_distribution_single

--- a/filter-below-abund.py	Sat Jan 21 14:42:04 2017 -0500
+++ b/filter-below-abund.py	Fri Sep 07 10:59:39 2018 -0400
@@ -38,11 +38,10 @@
 import os
 import sys

-import khmer
-from khmer.thread_utils import ThreadedSequenceProcessor, verbose_fasta_iter
+import screed
+from khmer import Countgraph, ReadParser
+from khmer.utils import (broken_paired_reader, write_record)

-WORKER_THREADS = 8
-GROUPSIZE = 100
 CUTOFF = 50


@@ -51,12 +50,9 @@
     infiles = sys.argv[2:]

     print('file with ht: %s' % counting_ht)
-    print('-- settings:')
-    print('N THREADS', WORKER_THREADS)
-    print('--')

     print('making hashtable')
-    ht = khmer.load_countgraph(counting_ht)
+    ht = Countgraph.load(counting_ht)
     K = ht.ksize()

     for infile in infiles:
@@ -65,22 +61,18 @@

         outfp = open(outfile, 'w')

-        def process_fn(record, ht=ht):
-            name = record['name']
-            seq = record['sequence']
+        paired_iter = broken_paired_reader(ReadParser(infile), min_length=K,
+                                           force_single=True)
+        for n, is_pair, read1, read2 in paired_iter:
+            name = read1.name
+            seq = read1.sequence
             if 'N' in seq:
                 return None, None

             trim_seq, trim_at = ht.trim_below_abundance(seq, CUTOFF)

             if trim_at >= K:
-                return name, trim_seq
-
-            return None, None
-
-        tsp = ThreadedSequenceProcessor(process_fn, WORKER_THREADS, GROUPSIZE)
-
-        tsp.start(verbose_fasta_iter(infile), outfp)
+                write_record(screed.Record(name=name, sequence=trim_seq), outfp)


 if __name__ == '__main__':
--- a/macros.xml	Sat Jan 21 14:42:04 2017 -0500
+++ b/macros.xml	Fri Sep 07 10:59:39 2018 -0400
@@ -1,9 +1,7 @@
 <macros>
-    <token name="@WRAPPER_VERSION@">2.0</token>
+    <token name="@WRAPPER_VERSION@">3.0.0a1</token>
     <xml name="requirements">
         <requirements>
-            <requirement type="package" version="0.98">bz2file</requirement>
-            <requirement type="package" version="0.9">screed</requirement>
             <requirement type="package" version="@WRAPPER_VERSION@">khmer</requirement>
         </requirements>
     </xml>
@@ -47,7 +45,7 @@
             <when value="specific">
                 <param name="ksize" type="integer" value="20" label="k-mer size" help="k-mer size to use" />
                 <param name="n_tables" type="integer" min="1" value="4" label="n_tables" help="number of tables to use" />
-                <param name="tablesize_specific" type="text" value="1000000.0"
+                <param name="tablesize_specific" type="text" value="1000000.0"
                     label="tablesize" help="(--max-tablesize) upper bound on the tablesize to use" />
             </when>
         </conditional>
@@ -93,31 +91,7 @@
     <token name="@HELP_FOOTER@"><![CDATA[
 (from the khmer project: http://khmer.readthedocs.org/en/v2.0/ )]]></token>
     <xml name="software-citation">
-        <citation type="bibtex">@article{khmer2015,
-     author = "Crusoe, Michael R. and Alameldin, Hussien F. and Awad, Sherine
-  and Bucher, Elmar and Caldwell, Adam and Cartwright, Reed and Charbonneau,
-  Amanda and Constantinides, Bede and Edvenson, Greg and Fay, Scott and Fenton,
-  Jacob and Fenzl, Thomas and Fish, Jordan and Garcia-Gutierrez, Leonor and
-  Garland, Phillip and Gluck, Jonathan and González, Iván and Guermond, Sarah
-  and Guo, Jiarong and Gupta, Aditi and Herr, Joshua R. and Howe, Adina and
-  Hyer, Alex and Härpfer, Andreas and Irber, Luiz and Kidd, Rhys and Lin, David
-  and Lippi, Justin and Mansour, Tamer and McA'Nulty, Pamela and McDonald, Eric
-  and Mizzi, Jessica and Murray, Kevin D. and Nahum, Joshua R. and Nanlohy,
-  Kaben and Nederbragt, Alexander Johan and Ortiz-Zuazaga, Humberto and Ory,
-  Jeramia and Pell, Jason and Pepe-Ranney, Charles and Russ, Zachary N and
-  Schwarz, Erich and Scott, Camille and Seaman, Josiah and Sievert, Scott and
-  Simpson, Jared and Skennerton, Connor T. and Spencer, James and Srinivasan,
-  Ramakrishnan and Standage, Daniel and Stapleton, James A. and Stein, Joe and
-  Steinman, Susan R and Taylor, Benjamin and Trimble, Will and Wiencko, Heather
-  L. and Wright, Michael and Wyss, Brian and Zhang, Qingpeng and zyme, en and
-  Brown, C. Titus"
-     title = "The khmer software package: enabling efficient nucleotide
-  sequence analysis",
-     year = "2015",
-     month = "08",
-     publisher = "F1000",
-     url = "http://dx.doi.org/10.12688/f1000research.6924.1"
-  }</citation>
+        <citation type="doi">10.12688/f1000research.6924.1</citation>
     </xml>
     <xml name="diginorm-citation">
         <citation type="bibtex">@unpublished{diginorm,
@@ -130,19 +104,7 @@
     url = "http://arxiv.org/abs/1203.4802",
 }</citation></xml>
     <xml name="graph-citation">
-        <citation type="bibtex">@article{Pell2012,
-  doi = {10.1073/pnas.1121464109},
-  url = {http://dx.doi.org/10.1073/pnas.1121464109},
-  year  = {2012},
-  month = {jul},
-  publisher = {Proceedings of the National Academy of Sciences},
-  volume = {109},
-  number = {33},
-  pages = {13272--13277},
-  author = {J. Pell and A. Hintze and R. Canino-Koning and A. Howe and J. M. Tiedje and C. T. Brown},
-  title = {Scaling metagenome sequence assembly with probabilistic de Bruijn graphs},
-  journal = {Proceedings of the National Academy of Sciences}
-  }</citation>
+        <citation type="doi">10.1073/pnas.1121464109</citation>
     </xml>
     <xml name="counting-citation">
         <citation type="doi">10.1371/journal.pone.0101271</citation>
--- a/tool_dependencies.xml	Sat Jan 21 14:42:04 2017 -0500
+++ /dev/null	Thu Jan 01 00:00:00 1970 +0000
@@ -1,12 +0,0 @@
-<?xml version="1.0"?>
-<tool_dependency>
-    <package name="bz2file" version="0.98">
-        <repository changeset_revision="73382c8fa8f3" name="package_bz2file_0_98" owner="iuc" toolshed="https://toolshed.g2.bx.psu.edu" />
-    </package>
-    <package name="screed" version="0.9">
-        <repository changeset_revision="639e5e37404d" name="package_screed_0_9" owner="iuc" toolshed="https://toolshed.g2.bx.psu.edu" />
-    </package>
-    <package name="khmer" version="2.0">
-        <repository changeset_revision="c02c1e1abd48" name="package_khmer_2_0" owner="iuc" toolshed="https://toolshed.g2.bx.psu.edu" />
-    </package>
-</tool_dependency>