view test/test_library_lookup.py @ 60:35f506f30ae4

fixed small rule in pdfread, and other small enhancements

'''
Created on Mar 6, 2012

@author: marcelk
'''
from GCMS import library_lookup, match_library
from pkg_resources import resource_filename  # @UnresolvedImport # pylint: disable=E0611
import os
import shutil
import tempfile
import unittest


class Test(unittest.TestCase):
    '''
    Tests the 'library_lookup' Galaxy tool
    '''

    def setUp(self):
        self.ri_database = resource_filename(__name__, "data/RIDB_subset.txt")
        self.nist_output = resource_filename(__name__, "data/NIST_tabular.txt")
        self.ridb_poly_regress = resource_filename(__name__, "data/ridb_poly_regression.txt")
        self.ridb_linear_regress = resource_filename(__name__, "data/ridb_linear_regression.txt")

    def test_create_lookup_table(self):
        '''
        Tests the 'create_lookup_table' function
        '''
        column_type = 'Capillary'
        polarity = 'Semi-standard non-polar6'
        lookup_dict = library_lookup.create_lookup_table(self.ri_database, column_type, polarity)
        self.assertFalse(False in [res[4] == 'Capillary' for res in lookup_dict['4177166']])
        self.assertEqual(['C51276336', '2,6-Dimethyl-octa-1,7-dien-3,6-diol', 'C10H18O2',
                          '1277', 'Capillary', 'Semi-standard non-polar', 'DB-5MS', '1',
                          'C51276336_DB-5MS', '', '', ''], lookup_dict['51276336'][1])

    def test_read_model(self):
        '''
        Tests reading the regression model data containing the parameters required for converting
        retention indices between GC-columns
        '''
        model, _ = library_lookup._read_model(self.ridb_poly_regress)
        # Order of values: coefficient 1 through 4, left limit, right limit
        # Polynomial model
        self.assertEqual([20.6155874639486, 0.945187096379008, 3.96480787567566e-05, -9.04377237159287e-09,
                          628.0, 2944.0, 405.0, 0, 0.998685262365514], model['HP-5']['SE-54'])
        self.assertEqual([-92.3963391356951, 1.26116176393346, -0.000191991657547972, 4.15387371263164e-08,
                          494.0, 2198.0, 407.0, 0, 0.996665023122993], model['Apiezon L']['Squalane'])
        # Linear model
        model, _ = library_lookup._read_model(self.ridb_linear_regress)
        self.assertEqual([2.81208738561543, 0.99482475526584, 628.0, 2944.0, 405.0, 0, 0.998643883946458],
                         model['HP-5']['SE-54'])
        self.assertEqual([19.979922768462, 0.993741869298272, 494.0, 2198.0, 407.0, 0, 0.99636062891041],
                         model['Apiezon L']['Squalane'])

    def test_apply_regression(self):
        '''
        Tests the regression model on some arbitrary retention indices
        '''
        poly_model, _ = library_lookup._read_model(self.ridb_poly_regress)
        linear_model, _ = library_lookup._read_model(self.ridb_linear_regress)
        retention_indices = [1000, 1010, 1020, 1030, 1040, 1050]
        converted_poly = []
        converted_linear = []
        for ri in retention_indices:
            converted_poly.append(library_lookup._apply_poly_regression('HP-5', 'DB-5', ri, poly_model))
            converted_linear.append(library_lookup._apply_linear_regression('HP-5', 'DB-5', ri, linear_model))

        self.assertEqual([1003.0566541860778, 1013.0979459524663, 1023.1358645806529, 1033.170466241159,
                          1043.2018071045052, 1053.2299433412131], converted_poly)
        self.assertEqual([1001.8127584915925, 1011.830140783027, 1021.8475230744615, 1031.864905365896,
                          1041.8822876573306, 1051.899669948765], converted_linear)
        
        # Test polynomial limit detection, the following RI falls outside of the possible limits
        ri = 3400
        converted_poly = library_lookup._apply_poly_regression('HP-5', 'DB-5', ri, poly_model)
        self.assertEqual(False, converted_poly)

    def test_preferred_hit(self):
        ''' Tests the matching of the hits with the preferred column, including regression '''
        model, method = library_lookup._read_model(self.ridb_poly_regress)
        column_type = 'Capillary'
        polarity = 'Semi-standard non-polar'
        lookup_dict = library_lookup.create_lookup_table(self.ri_database, column_type, polarity)
        hits = lookup_dict['150867']
        # No regression, should however consider order of given preference
        match = library_lookup._preferred(hits, ['SE-52', 'DB-5', 'HP-5'], column_type, polarity, model, method)
        expected = (['C150867', '(E)-phytol', 'C20H40O', '2110', 'Capillary',
                    'Semi-standard non-polar', 'SE-52', '', 'C150867_SE-52', '', '', ''], False)
        self.assertEqual(expected, match)

        # Perform regression by looking for 'OV-101' which isn't there. 'SE-52' has the best regression model
        # of the available columns
        match = library_lookup._preferred(hits, ['OV-101'], column_type, polarity, model, method)
        expected = (['C150867', '(E)-phytol', 'C20H40O', 2158.5769891569125, 'Capillary',
                     'Semi-standard non-polar', 'SE-52', '', 'C150867_SE-52', '', '', ''], 'SE-52')
        self.assertEqual(expected, match)

    def test_format_result(self):
        '''
        Tests the 'format_result' function
        '''
        column_type = 'Capillary'
        polarity = 'Semi-standard non-polar'

        # Look for DB-5
        pref_column = ['DB-5']
        model, method = library_lookup._read_model(self.ridb_poly_regress)
        lookup_dict = library_lookup.create_lookup_table(self.ri_database, column_type, polarity)
        data = library_lookup.format_result(lookup_dict, self.nist_output, pref_column, column_type,
                                            polarity, model, method)#False, None)

        # remove non-hits from set:
        data = _get_hits_only(data)
        self.assertEqual(['C544354', 'Ethyl linoleate', 'C20H36O2', '2155', 'Capillary', 'Semi-standard non-polar',
                           'DB-5', '1', 'C544354_DB-5', '1810', 'None', '', '', '0'], data[20])
        self.assertEqual(111, len(data))

        # Look for both DB-5 and HP-5
        pref_column = ['DB-5', 'HP-5']
        data = library_lookup.format_result(lookup_dict, self.nist_output, pref_column, column_type,
                                            polarity, False, None)
        # remove non-hits from set:
        data = _get_hits_only(data)
        self.assertEqual(['C502614', '.beta.-(E)-Farnesene', 'C15H24', '1508', 'Capillary', 'Semi-standard non-polar',
                          'DB-5', '1', 'C502614_DB-5', '942', 'None', '1482', '1522', '22'], data[50])
        self.assertEqual(106, len(data))


    def test_save_data(self):
        '''
        Tests the creation of the output tabular file
        '''
        temp_folder = tempfile.mkdtemp(prefix='gcms_combine_output_')
        saved_data = '{0}/{1}'.format(temp_folder, 'output.tsv')
        column_type = 'Capillary'
        polarity = 'Semi-standard non-polar'
        pref_column = ['DB-5']
        lookup_dict = library_lookup.create_lookup_table(self.ri_database, column_type, polarity)
        data = library_lookup.format_result(lookup_dict, self.nist_output, pref_column, column_type, polarity, False, None)
        library_lookup._save_data(data, saved_data)
        self.failUnless(os.path.exists(saved_data))
        shutil.rmtree(temp_folder)
        
        
    def test_match_library_get_lib_files(self):
        '''
        Tests the match_library.py functionality
        '''
        riqc_libs_dir = resource_filename(__name__, "../repositories/PRIMS-metabolomics/RI_DB_libraries")
        get_library_files_output = match_library.get_directory_files(riqc_libs_dir)
        self.assertEqual(2, len(get_library_files_output))
        self.assertEqual("Library_RI_DB_capillary_columns-noDuplicates", get_library_files_output[0][0])
        #TODO change assert below to assert that the result is a file, so the test can run on other dirs as well:
        #self.assertEqual("E:\\workspace\\PRIMS-metabolomics\\python-tools\\tools\\GCMS\\test\\data\\riqc_libs\\RI DB library (capillary columns) Dec.2012.txt", get_library_files_output[0][1])
        #self.assertEqual("RI DB library (capillary columns) Jan.2013", get_library_files_output[1][0])  
        try:
            get_library_files_output = match_library.get_directory_files("/blah")
            # should not come here
            self.assertTrue(False)
        except:
            # should come here
            self.assertTrue(True)

def _get_hits_only(data):
    '''
    removes items that have RI == 0.0 and Name == '' (these are dummy lines just for the output
    '''
    result = []
    for item_idx in xrange(len(data)):
        item = data[item_idx]
        if item[1] != '' and item[3] > 0.0 :
            result.append(item) 

    return result 


if __name__ == "__main__":
    #import sys;sys.argv = ['', 'Test.testName']
    unittest.main()