Implement a functionnal test with metaquast as final step.

.travis.yml

@@ -2,6 +2,7 @@ dist: trusty
 sudo: required
 language: cpp
 
+
 #gcc 4.9
 matrix:
   include:
@@ -17,21 +18,27 @@ matrix:
          - MATRIX_EVAL="CC=gcc-4.9 && CXX=g++-4.9"
 
 before_install:
-    - eval "${MATRIX_EVAL}"
+  - ./travis_download_db.sh & pid=$!
+  - free -m
+  - eval "${MATRIX_EVAL}"
+
 
 install:
   - sudo apt-get -qq update
+  - sudo apt-get install -y xz-utils
   - wget https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh -O miniconda.sh
   - bash miniconda.sh -b -p $HOME/miniconda
   - export PATH="$HOME/miniconda/bin:$PATH"
   - hash -r
   - conda config --set always_yes yes --set changeps1 no
   # Useful for debugging any issues with conda
   - conda info -a
-  - sudo apt-get install -y curl default-jdk ant libsparsehash-dev zlib1g-dev bzip2
-  - conda install --yes -c bioconda samtools
-  - conda install --yes pytest
-  - ./build.py
+  - conda env create -n matam -f environment.yml
+  - source activate matam
+  - ./travis_build.sh
+  - cd $HOME && wget https://sourceforge.net/projects/quast/files/quast-4.6.1.tar.gz && tar xvf quast-4.6.1.tar.gz  && export PATH="$HOME/quast-4.6.1":$PATH && cd -
+  # Wait until the db is dowloaded
+  - while kill -0 $pid 2> /dev/null;  do sleep 1; done;
 
 script:
-   - pytest
+  - travis_wait 60 pytest -rs

environment.yml

@@ -0,0 +1,36 @@
+name: matam
+
+channels:
+- bioconda
+- conda-forge
+- salford_systems
+- defaults
+
+dependencies:
+  - sed
+  - git
+  - curl
+  #- gcc-5
+  #- libgcc-5
+  - python >=3.5
+  - numpy
+#  - autoconf
+  - make
+  - cmake >=3.1
+  - wget
+#  - automake
+  - sparsehash
+  - zlib
+  - bzip2
+  - coreutils # Centos6 (otherwise sort --parallel is unrecognized)
+  - samtools ==1.4.1 # Centos7 (otherwise need libbz2.so.1.0)
+  - vsearch
+  - sga
+  - sortmerna >=2.1*
+  - rdptools
+  - krona
+  - exonerate=2.2
+  - pytest
+
+prefix: /home/lcouderc/miniconda3
+

tests/test_binaries_utils.py

@@ -16,7 +16,6 @@
                                             'sortmerna',
                                             'vsearch',
                                             'java', #rdp
-                                            'classifier.jar', #rdp
                                             'ktImportText', #krona
                                             ])
 

tests/test_functional.py

@@ -0,0 +1,137 @@
+import os
+import sys
+import tempfile
+import subprocess
+import shutil
+import multiprocessing
+import pytest
+import collections
+
+CURRENT_DIR = os.path.dirname(os.path.abspath(__file__))
+SCRIPTS_DIR = os.path.join(CURRENT_DIR, '..', 'scripts')
+EXAMPLES_DIR = os.path.join(CURRENT_DIR, '..', 'examples')
+DB_DIR = os.path.join(CURRENT_DIR, '..', 'db')
+
+sys.path.append(SCRIPTS_DIR)
+
+from binary_utils import Binary
+
+# skip all module tests if needed
+pytestmark = pytest.mark.skipif(
+    not os.path.isdir(DB_DIR),
+    reason='DB_DIR is missing:%s' % DB_DIR
+)
+
+
+@pytest.fixture(scope='module')
+def matam_results():
+    out = tempfile.mkdtemp(dir='/tmp/', prefix='matam_functionnal_test_')
+    p = {
+        'bin': os.path.join(SCRIPTS_DIR, 'matam_assembly.py'),
+        'reads': os.path.join(
+            EXAMPLES_DIR,
+            '16sp_simulated_dataset/16sp.art_HS25_pe_100bp_50x.fq'
+        ),
+        'db': os.path.join(DB_DIR, 'SILVA_128_SSURef_NR95'),
+        'out': out,
+        'cpu': multiprocessing.cpu_count()
+    }
+    cmd = '{bin} -i {reads} -d {db} -o {out} --cpu {cpu} --max_memory 3000 \
+        --debug --coverage_threshold 2000 \
+        --perform_taxonomic_assignment'.format(**p)
+
+    completed_process = subprocess.run(cmd, shell=True)
+    return_code = completed_process.returncode
+    fasta = os.path.join(out, 'final_assembly.fa')
+    krona_html = os.path.join(out, 'krona.html')
+    krona_tab = os.path.join(out, 'krona.tab')
+    rdp_tab = os.path.join(out, 'rdp.tab')
+
+    MatamResults = collections.namedtuple(
+        "MatamResults",
+        "return_code fasta krona_html krona_tab rdp_tab"
+    )
+    results = MatamResults(
+        return_code=return_code,
+        fasta=fasta,
+        krona_html=krona_html,
+        krona_tab=krona_tab,
+        rdp_tab=rdp_tab
+    )
+
+    yield results
+
+    if os.path.isdir(out):
+        shutil.rmtree(out)
+
+
+def exists_and_not_empty(fpath):
+    return os.path.isfile(fpath) and os.stat(fpath).st_size != 0
+
+
+def test_return_code(matam_results):
+    assert matam_results.return_code == 0
+
+
+def test_final_fasta_file(matam_results):
+    assert exists_and_not_empty(matam_results.fasta)
+
+
+def test_krona_html(matam_results):
+    assert exists_and_not_empty(matam_results.krona_html)
+
+
+def test_krona_tab(matam_results):
+    assert exists_and_not_empty(matam_results.krona_tab)
+
+
+def test_rdp_tab(matam_results):
+    assert exists_and_not_empty(matam_results.rdp_tab)
+
+
+def extract_metaquast_val(tsv):
+    with open(tsv, 'r') as tsv_handler:
+        lines = tsv_handler.readlines()
+        return float(lines[1].split('\t')[1].strip())
+
+
+@pytest.mark.skipif(
+    not Binary.which('metaquast.py'),
+    reason="requires metaquast.py to be in PATH"
+)
+def test_metaquast(matam_results):
+    data_directory = tempfile.mkdtemp(dir='/tmp/', prefix='metaquast_')
+    fasta = matam_results.fasta
+    true_ref = os.path.join(EXAMPLES_DIR, '16sp_simulated_dataset/16sp.fasta')
+    cmd = "metaquast.py -a all --ambiguity-score 1 --min-identity 97 -x 500 \
+        --unaligned-part-size 200 -R %s %s" % (true_ref, fasta)
+    subprocess.run(cmd, shell=True, cwd=data_directory)
+
+    genome_fraction_file = os.path.join(
+        data_directory,
+        'quast_results/latest/summary/TSV/Genome_fraction_(%).tsv'
+    )
+    mismatches_file = os.path.join(
+        data_directory,
+        'quast_results/latest/summary/TSV/#_mismatches_per_100_kbp.tsv'
+    )
+    indels_file = os.path.join(
+        data_directory,
+        'quast_results/latest/summary/TSV/#_indels_per_100_kbp.tsv'
+    )
+    ns_file = os.path.join(
+        data_directory,
+        'quast_results/latest/summary/TSV/#_N\'s_per_100_kbp.tsv'
+    )
+
+    genome_fraction = extract_metaquast_val(genome_fraction_file)
+    mismatches = extract_metaquast_val(mismatches_file)
+    indels = extract_metaquast_val(indels_file)
+    ns = extract_metaquast_val(ns_file)
+    error_rate = (mismatches + indels + ns) / 1000  # 100000bp * 100
+
+    assert genome_fraction > 86.4
+    assert error_rate < 0.06
+
+    if os.path.isdir(data_directory):
+        shutil.rmtree(data_directory)

tests/test_rdp.py

@@ -11,26 +11,32 @@
 from rdp import run_rdp_classifier, read_rpd_file, get_lineage, filter_rdp_file
 from binary_utils import Binary
 
+
 def test_run_rdp_classifier_ok():
     fasta = os.path.join(SAMPLE_DIR, 'scaffolds.fa')
-    bin = Binary.which('java')
     jar = Binary.which('classifier.jar')
-    rdp_exe = '{java} -Xmx1g -jar {jar}'.format(java=bin, jar=jar)
+
+    if jar is not None:
+        java = Binary.assert_which('java')
+        rdp_exe = '{java} -Xmx1g -jar {jar}'.format(java=java, jar=jar)
+    else:
+        rdp_exe = Binary.assert_which('classifier')
+
     result_file = tempfile.NamedTemporaryFile()
     run_rdp_classifier(rdp_exe, fasta, result_file.name)
 
 
 def test_read_rdp_results():
     rdp_file = os.path.join(SAMPLE_DIR, 'rdp.txt')
     lines = list(read_rpd_file(rdp_file))
-    assert len(lines) == 23 # 23 scaffolds
+    assert len(lines) == 23  # 23 scaffolds
 
 
 def test_filter_rdp_file():
     rdp_file = os.path.join(SAMPLE_DIR, 'rdp.txt')
     result_file = tempfile.NamedTemporaryFile()
     filter_rdp_file(rdp_file, result_file.name)
     for line in read_rpd_file(result_file.name):
-        assert len(line) == 19 # seqid + 6 taxonomic levels * 3
+        assert len(line) == 19  # seqid + 6 taxonomic levels * 3
         if line[0] == "87":
             assert get_lineage(line) == ['unclassified'] * 6

travis_build.sh

@@ -0,0 +1,12 @@
+#!/bin/bash
+
+# An activated env is needed for $CONDA_PREFIX
+#CC=${CONDA_PREFIX}/bin/gcc
+#CXX=${CONDA_PREFIX}/bin/g++
+
+# Build componentsearch
+cd componentsearch && make && cd -
+
+# Build ovgraphbuild
+build_dir=ovgraphbuild/build
+mkdir -p  $build_dir && cd $build_dir && cmake .. -G"CodeBlocks - Unix Makefiles" && make && cd -

travis_download_db.sh

@@ -0,0 +1,5 @@
+#!/bin/bash
+
+wget --quiet http://bioinfo.lifl.fr/matam/SILVA_128_SSURef_NR95_indexed_max_mem_3G.tar.xz{.md5,}
+md5sum -c SILVA_128_SSURef_NR95_indexed_max_mem_3G.tar.xz.md5
+tar Jxvf SILVA_128_SSURef_NR95_indexed_max_mem_3G.tar.xz