def setUp(self):
"""Create MetadataMap objects that will be used in the tests."""
# Create a map using the overview tutorial mapping file.
self.overview_map_str = [
"#SampleID\tBarcodeSequence\tTreatment\tDOB\tDescription",
"PC.354\tAGCACGAGCCTA\tControl\t20061218\t354",
"PC.355\tAACTCGTCGATG\tControl\t20061218\t355",
"PC.356\tACAGACCACTCA\tControl\t20061126\t356",
"PC.481\tACCAGCGACTAG\tControl\t20070314\t481",
"PC.593\tAGCAGCACTTGT\tControl\t20071210\t593",
"PC.607\tAACTGTGCGTAC\tFast\t20071112\t607",
"PC.634\tACAGAGTCGGCT\tFast\t20080116\t634",
"PC.635\tACCGCAGAGTCA\tFast\t20080116\t635",
"PC.636\tACGGTGAGTGTC\tFast\t20080116\t636"]
self.overview_map = MetadataMap(
*parse_mapping_file_to_dict(self.overview_map_str))
# Create the same overview tutorial map, but this time with some
# comments.
self.comment = "# Some comments about this mapping file"
self.map_with_comments_str = self.overview_map_str[:]
self.map_with_comments_str.insert(1, self.comment)
self.map_with_comments = MetadataMap(*parse_mapping_file_to_dict(
self.map_with_comments_str))
# Create a MetadataMap object that has no metadata (i.e. no sample IDs,
# so no metadata about samples).
self.empty_map = MetadataMap({}, [])
# Create a MetadataMap object that has samples (i.e. sample IDs) but
# not associated metadata (i.e. no columns other than SampleID).
self.no_metadata_str = ["#SampleID",
"PC.354",
"PC.355",
"PC.356",
"PC.481",
"PC.593",
"PC.607",
"PC.634",
"PC.635",
"PC.636"]
self.no_metadata = MetadataMap(*parse_mapping_file_to_dict(
self.no_metadata_str))
# Create a MetadataMap object that has a category with only one value
# throughout the entire column.
self.single_value_str = ["#SampleID\tFoo",
"PC.354\tfoo",
"PC.355\tfoo",
"PC.356\tfoo",
"PC.481\tfoo",
"PC.593\tfoo",
"PC.607\tfoo",
"PC.634\tfoo",
"PC.635\tfoo",
"PC.636\tfoo"]
self.single_value = MetadataMap(*parse_mapping_file_to_dict(
self.single_value_str))
def test_ne(self):
"""Test whether two MetadataMaps are not equal."""
self.assertTrue(self.empty_map != MetadataMap({}, ["foo"]))
self.assertTrue(self.overview_map != MetadataMap(
self.overview_map._metadata, ["foo"]))
self.assertTrue(
self.overview_map != MetadataMap({}, self.overview_map.Comments))
self.assertTrue(self.overview_map != self.empty_map)
self.assertTrue(self.overview_map != self.map_with_comments)
self.assertTrue(self.overview_map != self.no_metadata)
def format_mapping_file_to_js(mapping_file_data, mapping_file_headers,
columns):
"""Write a javascript representation of the mapping file
Parameters
----------
mapping_file_data : list of str
contents of the mapping file
mapping_file_headers : list of list of str
headers of the mapping file
columns : list of str
valid columns to use, usually a subset of mapping_file_headers
Returns
-------
str
JavaScript representation of the mapping file
"""
map_str = []
mapping_file_dict = mapping_file_to_dict(mapping_file_data,
mapping_file_headers)
map_values = []
for k, v in mapping_file_dict.items():
if 'SampleID' in columns:
vals = ["'%s'" % k
] + ["'%s'" % v[col] for col in mapping_file_headers[1:]]
else:
vals = ["'%s'" % v[col] for col in mapping_file_headers[1:]]
map_values.append("'%s': [%s]" % (k, ','.join(vals)))
if 'SampleID' not in columns:
mapping_file_headers = mapping_file_headers[1:]
# format the mapping file as javascript objects
map_str.append('var g_mappingFileHeaders = [%s];\n' %
','.join(["'%s'" % col for col in mapping_file_headers]))
map_str.append('var g_mappingFileData = { %s };\n' % ','.join(map_values))
map_object = MetadataMap(mapping_file_dict, [])
# make sure the comparison for SampleID is made first because otherwise
# if the metadata map tries to check 'SampleID' it will raise an exception
animatable_categories = [
category for category in columns
if category != 'SampleID' and map_object.isNumericCategory(category)
]
map_str.append('var g_animatableMappingFileHeaders = [%s];\n' %
','.join(["'%s'" % col for col in animatable_categories]))
return ''.join(map_str)
def format_mapping_file_to_js(mapping_file_data, mapping_file_headers,
columns):
"""Write a javascript representation of the mapping file
Parameters
----------
mapping_file_data : list of str
contents of the mapping file
mapping_file_headers : list of list of str
headers of the mapping file
columns : list of str
valid columns to use, usually a subset of mapping_file_headers
Returns
-------
str
JavaScript representation of the mapping file
"""
map_str = []
mapping_file_dict = mapping_file_to_dict(mapping_file_data,
mapping_file_headers)
map_values = []
for k, v in mapping_file_dict.items():
if 'SampleID' in columns:
vals = ["'%s'" % k] + ["'%s'" % v[col]
for col in mapping_file_headers[1:]]
else:
vals = ["'%s'" % v[col] for col in mapping_file_headers[1:]]
map_values.append("'%s': [%s]" % (k, ','.join(vals)))
if 'SampleID' not in columns:
mapping_file_headers = mapping_file_headers[1:]
# format the mapping file as javascript objects
map_str.append('var g_mappingFileHeaders = [%s];\n' %
','.join(["'%s'" % col for col in mapping_file_headers]))
map_str.append('var g_mappingFileData = { %s };\n' % ','.join(map_values))
map_object = MetadataMap(mapping_file_dict, [])
# make sure the comparison for SampleID is made first because otherwise
# if the metadata map tries to check 'SampleID' it will raise an exception
animatable_categories = [category for category in columns
if category != 'SampleID' and
map_object.isNumericCategory(category)]
map_str.append('var g_animatableMappingFileHeaders = [%s];\n' %
','.join(["'%s'" % col for col in animatable_categories]))
return ''.join(map_str)
def setUp(self):
"""Create MetadataMap objects that will be used in the tests."""
# Create a map using the overview tutorial mapping file.
self.overview_map_str = [
"#SampleID\tBarcodeSequence\tTreatment\tDOB\tDescription",
"PC.354\tAGCACGAGCCTA\tControl\t20061218\t354",
"PC.355\tAACTCGTCGATG\tControl\t20061218\t355",
"PC.356\tACAGACCACTCA\tControl\t20061126\t356",
"PC.481\tACCAGCGACTAG\tControl\t20070314\t481",
"PC.593\tAGCAGCACTTGT\tControl\t20071210\t593",
"PC.607\tAACTGTGCGTAC\tFast\t20071112\t607",
"PC.634\tACAGAGTCGGCT\tFast\t20080116\t634",
"PC.635\tACCGCAGAGTCA\tFast\t20080116\t635",
"PC.636\tACGGTGAGTGTC\tFast\t20080116\t636"
]
self.overview_map = MetadataMap(
*parse_mapping_file_to_dict(self.overview_map_str))
# Create the same overview tutorial map, but this time with some
# comments.
self.comment = "# Some comments about this mapping file"
self.map_with_comments_str = self.overview_map_str[:]
self.map_with_comments_str.insert(1, self.comment)
self.map_with_comments = MetadataMap(
*parse_mapping_file_to_dict(self.map_with_comments_str))
# Create a MetadataMap object that has no metadata (i.e. no sample IDs,
# so no metadata about samples).
self.empty_map = MetadataMap({}, [])
# Create a MetadataMap object that has samples (i.e. sample IDs) but
# not associated metadata (i.e. no columns other than SampleID).
self.no_metadata_str = [
"#SampleID", "PC.354", "PC.355", "PC.356", "PC.481", "PC.593",
"PC.607", "PC.634", "PC.635", "PC.636"
]
self.no_metadata = MetadataMap(
*parse_mapping_file_to_dict(self.no_metadata_str))
# Create a MetadataMap object that has a category with only one value
# throughout the entire column.
self.single_value_str = [
"#SampleID\tFoo", "PC.354\tfoo", "PC.355\tfoo", "PC.356\tfoo",
"PC.481\tfoo", "PC.593\tfoo", "PC.607\tfoo", "PC.634\tfoo",
"PC.635\tfoo", "PC.636\tfoo"
]
self.single_value = MetadataMap(
*parse_mapping_file_to_dict(self.single_value_str))
def test_eq(self):
"""Test whether two MetadataMaps are equal."""
self.assertTrue(self.empty_map == MetadataMap({}, []))
self.assertTrue(self.overview_map == MetadataMap(
self.overview_map._metadata, self.overview_map.Comments))
def test_parseMetadataMap(self):
"""Test parsing a mapping file into a MetadataMap instance."""
obs = MetadataMap.parseMetadataMap(self.overview_map_str)
self.assertEqual(obs, self.overview_map)
class MetadataMapTests(TestCase):
"""Tests for the MetadataMap class."""
def setUp(self):
"""Create MetadataMap objects that will be used in the tests."""
# Create a map using the overview tutorial mapping file.
self.overview_map_str = [
"#SampleID\tBarcodeSequence\tTreatment\tDOB\tDescription",
"PC.354\tAGCACGAGCCTA\tControl\t20061218\t354",
"PC.355\tAACTCGTCGATG\tControl\t20061218\t355",
"PC.356\tACAGACCACTCA\tControl\t20061126\t356",
"PC.481\tACCAGCGACTAG\tControl\t20070314\t481",
"PC.593\tAGCAGCACTTGT\tControl\t20071210\t593",
"PC.607\tAACTGTGCGTAC\tFast\t20071112\t607",
"PC.634\tACAGAGTCGGCT\tFast\t20080116\t634",
"PC.635\tACCGCAGAGTCA\tFast\t20080116\t635",
"PC.636\tACGGTGAGTGTC\tFast\t20080116\t636"
]
self.overview_map = MetadataMap(
*parse_mapping_file_to_dict(self.overview_map_str))
# Create the same overview tutorial map, but this time with some
# comments.
self.comment = "# Some comments about this mapping file"
self.map_with_comments_str = self.overview_map_str[:]
self.map_with_comments_str.insert(1, self.comment)
self.map_with_comments = MetadataMap(
*parse_mapping_file_to_dict(self.map_with_comments_str))
# Create a MetadataMap object that has no metadata (i.e. no sample IDs,
# so no metadata about samples).
self.empty_map = MetadataMap({}, [])
# Create a MetadataMap object that has samples (i.e. sample IDs) but
# not associated metadata (i.e. no columns other than SampleID).
self.no_metadata_str = [
"#SampleID", "PC.354", "PC.355", "PC.356", "PC.481", "PC.593",
"PC.607", "PC.634", "PC.635", "PC.636"
]
self.no_metadata = MetadataMap(
*parse_mapping_file_to_dict(self.no_metadata_str))
# Create a MetadataMap object that has a category with only one value
# throughout the entire column.
self.single_value_str = [
"#SampleID\tFoo", "PC.354\tfoo", "PC.355\tfoo", "PC.356\tfoo",
"PC.481\tfoo", "PC.593\tfoo", "PC.607\tfoo", "PC.634\tfoo",
"PC.635\tfoo", "PC.636\tfoo"
]
self.single_value = MetadataMap(
*parse_mapping_file_to_dict(self.single_value_str))
def test_parseMetadataMap(self):
"""Test parsing a mapping file into a MetadataMap instance."""
obs = MetadataMap.parseMetadataMap(self.overview_map_str)
self.assertEqual(obs, self.overview_map)
def test_parseMetadataMap_empty(self):
"""Test parsing empty mapping file contents."""
self.assertRaises(QiimeParseError, MetadataMap.parseMetadataMap, [])
def test_eq(self):
"""Test whether two MetadataMaps are equal."""
self.assertTrue(self.empty_map == MetadataMap({}, []))
self.assertTrue(self.overview_map == MetadataMap(
self.overview_map._metadata, self.overview_map.Comments))
def test_ne(self):
"""Test whether two MetadataMaps are not equal."""
self.assertTrue(self.empty_map != MetadataMap({}, ["foo"]))
self.assertTrue(self.overview_map != MetadataMap(
self.overview_map._metadata, ["foo"]))
self.assertTrue(
self.overview_map != MetadataMap({}, self.overview_map.Comments))
self.assertTrue(self.overview_map != self.empty_map)
self.assertTrue(self.overview_map != self.map_with_comments)
self.assertTrue(self.overview_map != self.no_metadata)
def test_getSampleMetadata(self):
"""Test metadata by sample ID accessor with valid sample IDs."""
exp = {
'BarcodeSequence': 'AGCACGAGCCTA',
'Treatment': 'Control',
'DOB': '20061218',
'Description': '354'
}
obs = self.overview_map.getSampleMetadata('PC.354')
self.assertEqual(obs, exp)
exp = {
'BarcodeSequence': 'ACCAGCGACTAG',
'Treatment': 'Control',
'DOB': '20070314',
'Description': '481'
}
obs = self.map_with_comments.getSampleMetadata('PC.481')
self.assertEqual(obs, exp)
exp = {
'BarcodeSequence': 'ACGGTGAGTGTC',
'Treatment': 'Fast',
'DOB': '20080116',
'Description': '636'
}
obs = self.map_with_comments.getSampleMetadata('PC.636')
self.assertEqual(obs, exp)
exp = {}
obs = self.no_metadata.getSampleMetadata('PC.636')
self.assertEqual(obs, exp)
def test_getSampleMetadata_bad_sample_id(self):
"""Test metadata by sample ID accessor with invalid sample IDs."""
# Nonexistent sample ID.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata,
'PC.000')
self.assertRaises(KeyError, self.no_metadata.getSampleMetadata,
'PC.000')
# Integer sample ID.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata, 42)
# Sample ID of type None.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata, None)
# Sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, 's1')
# Integer sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, 1)
# Sample ID of None on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, None)
def test_getCategoryValue(self):
"""Test category value by sample ID/category name accessor."""
exp = "Fast"
obs = self.overview_map.getCategoryValue('PC.634', 'Treatment')
self.assertEqual(obs, exp)
exp = "20070314"
obs = self.overview_map.getCategoryValue('PC.481', 'DOB')
self.assertEqual(obs, exp)
exp = "ACGGTGAGTGTC"
obs = self.map_with_comments.getCategoryValue('PC.636',
'BarcodeSequence')
self.assertEqual(obs, exp)
def test_getCategoryValues(self):
"""Test category value list by sample ID/category name accessor."""
smpl_ids = [
'PC.354', 'PC.355', 'PC.356', 'PC.481', 'PC.593', 'PC.607',
'PC.634', 'PC.635', 'PC.636'
]
exp = [
'Control', 'Control', 'Control', 'Control', 'Control', 'Fast',
'Fast', 'Fast', 'Fast'
]
obs = self.overview_map.getCategoryValues(smpl_ids, 'Treatment')
self.assertEqual(obs, exp)
def test_isNumericCategory(self):
"""Test checking if a category is numeric."""
obs = self.overview_map.isNumericCategory('Treatment')
self.assertEqual(obs, False)
obs = self.overview_map.isNumericCategory('DOB')
self.assertEqual(obs, True)
def test_hasUniqueCategoryValues(self):
"""Test checking if a category has unique values."""
obs = self.overview_map.hasUniqueCategoryValues('Treatment')
self.assertEqual(obs, False)
obs = self.overview_map.hasUniqueCategoryValues('DOB')
self.assertEqual(obs, False)
obs = self.overview_map.hasUniqueCategoryValues('Description')
self.assertEqual(obs, True)
def test_hasSingleCategoryValue(self):
"""Test checking if a category has only a single value."""
obs = self.overview_map.hasSingleCategoryValue('Treatment')
self.assertEqual(obs, False)
obs = self.single_value.hasSingleCategoryValue('Foo')
self.assertEqual(obs, True)
def test_getCategoryValue_bad_sample_id(self):
"""Test category value by sample ID accessor with bad sample IDs."""
# Nonexistent sample ID.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.000', 'Treatment')
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.000', 'Treatment')
# Integer sample ID.
self.assertRaises(KeyError, self.overview_map.getCategoryValue, 42,
'DOB')
# Sample ID of type None.
self.assertRaises(KeyError, self.overview_map.getCategoryValue, None,
'Treatment')
# Sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, 's1',
'foo')
# Integer sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, 1, 'bar')
# Sample ID of None on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, None,
'baz')
def test_getCategoryValue_bad_category(self):
"""Test category value by sample ID accessor with bad categories."""
# Nonexistent category.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', 'foo')
# Integer category.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', 42)
# Category of type None.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', None)
# Category on map with no metadata, but that has sample IDs.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', 'Treatment')
# Integer category on map with no metadata.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', 34)
# Category of type None on map with no metadata.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', None)
def test_SampleIds(self):
"""Test sample IDs accessor."""
exp = [
"PC.354", "PC.355", "PC.356", "PC.481", "PC.593", "PC.607",
"PC.634", "PC.635", "PC.636"
]
obs = self.overview_map.SampleIds
self.assertEqual(obs, exp)
obs = self.no_metadata.SampleIds
self.assertEqual(obs, exp)
obs = self.empty_map.SampleIds
self.assertEqual(obs, [])
def test_CategoryNames(self):
"""Test category names accessor."""
exp = ["BarcodeSequence", "DOB", "Description", "Treatment"]
obs = self.overview_map.CategoryNames
self.assertEqual(obs, exp)
obs = self.no_metadata.CategoryNames
self.assertEqual(obs, [])
obs = self.empty_map.CategoryNames
self.assertEqual(obs, [])
def test_filterSamples(self):
"""Test filtering out samples from metadata map."""
exp = ['PC.356', 'PC.593']
self.overview_map.filterSamples(['PC.593', 'PC.356'])
obs = self.overview_map.SampleIds
self.assertEqual(obs, exp)
self.overview_map.filterSamples([])
self.assertEqual(self.overview_map.SampleIds, [])
def test_filterSamples_strict(self):
"""Test strict checking of sample prescence when filtering."""
with self.assertRaises(ValueError):
self.overview_map.filterSamples(['PC.356', 'abc123'])
with self.assertRaises(ValueError):
self.empty_map.filterSamples(['foo'])
def test_filterSamples_no_strict(self):
"""Test missing samples does not raise error."""
self.overview_map.filterSamples(['PC.356', 'abc123'], strict=False)
self.assertEqual(self.overview_map.SampleIds, ['PC.356'])
self.empty_map.filterSamples(['foo'], strict=False)
self.assertEqual(self.empty_map.SampleIds, [])
def test_is_valid_git_refname(self):
"""Test correct validation of refnames"""
# valid branchnames
self.assertTrue(is_valid_git_refname('master'))
self.assertTrue(is_valid_git_refname('debuggatron_2000'))
self.assertTrue(is_valid_git_refname('refname/bar'))
self.assertTrue(is_valid_git_refname('ref.nameslu/_eggs_/spam'))
self.assertTrue(is_valid_git_refname('valid{0}char'.format(
unichr(40))))
self.assertTrue(is_valid_git_refname('master@head'))
self.assertTrue(is_valid_git_refname('bar{thing}foo'))
# case happening with git < 1.6.6
self.assertFalse(
is_valid_git_refname(
'--abbrev-ref\nbaa350d7b7063d585ca293fc16ef15e0765dc9ee'))
# different invalid refnames, for a description of each group see the
# man page of git check-ref-format
self.assertFalse(is_valid_git_refname('bar/.spam/eggs'))
self.assertFalse(is_valid_git_refname('bar.lock/spam/eggs'))
self.assertFalse(is_valid_git_refname('bar.lock'))
self.assertFalse(is_valid_git_refname('.foobar'))
self.assertFalse(is_valid_git_refname('ref..name'))
self.assertFalse(
is_valid_git_refname(u'invalid{0}char'.format(unichr(177))))
self.assertFalse(
is_valid_git_refname('invalid{0}char'.format(unichr(39))))
self.assertFalse(is_valid_git_refname('ref~name/bar'))
self.assertFalse(is_valid_git_refname('refname spam'))
self.assertFalse(is_valid_git_refname('bar/foo/eggs~spam'))
self.assertFalse(is_valid_git_refname('bar:_spam_'))
self.assertFalse(is_valid_git_refname('eggtastic^2'))
self.assertFalse(is_valid_git_refname('areyourandy?'))
self.assertFalse(is_valid_git_refname('bar/*/spam'))
self.assertFalse(is_valid_git_refname('bar[spam]/eggs'))
self.assertFalse(is_valid_git_refname('/barfooeggs'))
self.assertFalse(is_valid_git_refname('barfooeggs/'))
self.assertFalse(is_valid_git_refname('bar/foo//////eggs'))
self.assertFalse(is_valid_git_refname('dotEnding.'))
self.assertFalse(is_valid_git_refname('@{branch'))
self.assertFalse(is_valid_git_refname('contains\\slash'))
self.assertFalse(is_valid_git_refname('$newbranch'))
def test_is_valid_git_sha1(self):
""" """
# valid sha1 strings
self.assertTrue(
is_valid_git_sha1('65a9ba2ef4b126fb5b054ea6b89b457463db4ec6'))
self.assertTrue(
is_valid_git_sha1('a29a9911e41253405494c43889925a6d79ca26db'))
self.assertTrue(
is_valid_git_sha1('e099cd5fdea89eba929d6051fbd26cc9e7a0c961'))
self.assertTrue(
is_valid_git_sha1('44235d322c3386bd5ce872d9d7ea2e10d27c86cb'))
self.assertTrue(
is_valid_git_sha1('7d2fc23E04540EE92c742948cca9ed5bc54d08d1'))
self.assertTrue(
is_valid_git_sha1('fb5dc0285a8b11f199c4f3a7547a2da38138373f'))
self.assertTrue(
is_valid_git_sha1('0b2abAEb195ba7ebc5cfdb53213a66fbaddefdb8'))
# invalid length
self.assertFalse(is_valid_git_sha1('cca9ed5bc54d08d1'))
self.assertFalse(is_valid_git_sha1(''))
# invalid characters
self.assertFalse(
is_valid_git_sha1('fb5dy0f85a8b11f199c4f3a75474a2das8138373'))
self.assertFalse(
is_valid_git_sha1('0x5dcc816fbc1c2e8eX087d7d2ed8d2950a7c16b'))
def preprocess_mapping_file(data, headers, columns, unique=False, single=False,
clones=0):
"""Process a mapping file to expand the data or remove unuseful fields
Inputs:
data: mapping file data
headers: mapping file headers
columns: list of headers to keep, if one of these headers includes two
ampersands, this function will create a new column by merging the delimited
columns.
unique: keep columns where all values are unique
single: keep columns where all values are the same
clones: number of times to replicate the metadata
Outputs:
data: processed mapping file data
headers: processed mapping file headers
"""
# The sample ID must always be there, else it's meaningless data
if 'SampleID' != columns[0]:
columns = ['SampleID'] + columns
# process concatenated columns if needed
merge = []
for column in columns:
# the list can contain None so check "if column" before treating as str
if column and '&&' in column:
merge.append(column)
# each element needs several columns to be merged
for new_column in merge:
indices = [headers.index(header_name) for header_name in
new_column.split('&&')]
# join all the fields of the metadata that are listed in indices
for line in data:
line.append(''.join([line[index] for index in indices]))
headers.append(new_column)
# remove all unique or singled valued columns that are not included in
# the list of categories that should be kept i. e. columns
if unique or single:
columns_to_remove = []
metadata = MetadataMap(mapping_file_to_dict(data, headers), [])
# the --coloy_by option in the script interface allows the user to
# specify the categories you want to use in the generated plot, this
# the default behaviour is to color by all categories that are not
# unique. If the user specifies a category with with the --color_by
# option and this category contains a unique values, this category must
# still be added thus the structure of the next few lines that
# form the structure for the two different routes. (1) where no value
# is specified in the CLI (the value of columns will be [None, x1, x2,
# x3] where x{1,2,3} are categories requested in other CLI options) and
# (2) where a value is specified in the CLI.
#
# TL;DR
# see https://github.com/biocore/emperor/issues/271
if None in columns:
columns = headers[:]
f_unique = metadata.hasUniqueCategoryValues
f_single = metadata.hasSingleCategoryValue
else:
f_unique = lambda x: metadata.hasUniqueCategoryValues(x) and\
x not in columns
f_single = lambda x: metadata.hasSingleCategoryValue(x) and\
x not in columns
# find columns that have values that are all unique
if unique:
for c in headers[1::]:
if f_unique(c):
columns_to_remove.append(c)
# remove categories where there is only one value
if single:
for c in headers[1::]:
if f_single(c):
columns_to_remove.append(c)
columns_to_remove = list(set(columns_to_remove))
# remove the single or unique columns
data, headers = keep_columns_from_mapping_file(data, headers,
columns_to_remove, negate=True)
else:
# when a None is contained in columns, we imply we want to use all the
# available categories in the mapping file, thus just overwrite the
# value
if None in columns:
columns = headers[:]
# remove anything not specified in the input
data, headers = keep_columns_from_mapping_file(data, headers, columns)
# sanitize the mapping file data and headers
data, headers = sanitize_mapping_file(data, headers)
# clones mean: replicate the metadata retagging the sample ids with a suffix
if clones:
out_data = []
for index in range(0, clones):
out_data.extend([[element[0]+'_%d' % index]+element[1::]
for element in data])
data = out_data
return data, headers
def test_parseMetadataMap(self):
"""Test parsing a mapping file into a MetadataMap instance."""
obs = MetadataMap.parseMetadataMap(self.overview_map_str)
self.assertEqual(obs, self.overview_map)
class MetadataMapTests(TestCase):
"""Tests for the MetadataMap class."""
def setUp(self):
"""Create MetadataMap objects that will be used in the tests."""
# Create a map using the overview tutorial mapping file.
self.overview_map_str = [
"#SampleID\tBarcodeSequence\tTreatment\tDOB\tDescription",
"PC.354\tAGCACGAGCCTA\tControl\t20061218\t354",
"PC.355\tAACTCGTCGATG\tControl\t20061218\t355",
"PC.356\tACAGACCACTCA\tControl\t20061126\t356",
"PC.481\tACCAGCGACTAG\tControl\t20070314\t481",
"PC.593\tAGCAGCACTTGT\tControl\t20071210\t593",
"PC.607\tAACTGTGCGTAC\tFast\t20071112\t607",
"PC.634\tACAGAGTCGGCT\tFast\t20080116\t634",
"PC.635\tACCGCAGAGTCA\tFast\t20080116\t635",
"PC.636\tACGGTGAGTGTC\tFast\t20080116\t636"]
self.overview_map = MetadataMap(
*parse_mapping_file_to_dict(self.overview_map_str))
# Create the same overview tutorial map, but this time with some
# comments.
self.comment = "# Some comments about this mapping file"
self.map_with_comments_str = self.overview_map_str[:]
self.map_with_comments_str.insert(1, self.comment)
self.map_with_comments = MetadataMap(*parse_mapping_file_to_dict(
self.map_with_comments_str))
# Create a MetadataMap object that has no metadata (i.e. no sample IDs,
# so no metadata about samples).
self.empty_map = MetadataMap({}, [])
# Create a MetadataMap object that has samples (i.e. sample IDs) but
# not associated metadata (i.e. no columns other than SampleID).
self.no_metadata_str = ["#SampleID",
"PC.354",
"PC.355",
"PC.356",
"PC.481",
"PC.593",
"PC.607",
"PC.634",
"PC.635",
"PC.636"]
self.no_metadata = MetadataMap(*parse_mapping_file_to_dict(
self.no_metadata_str))
# Create a MetadataMap object that has a category with only one value
# throughout the entire column.
self.single_value_str = ["#SampleID\tFoo",
"PC.354\tfoo",
"PC.355\tfoo",
"PC.356\tfoo",
"PC.481\tfoo",
"PC.593\tfoo",
"PC.607\tfoo",
"PC.634\tfoo",
"PC.635\tfoo",
"PC.636\tfoo"]
self.single_value = MetadataMap(*parse_mapping_file_to_dict(
self.single_value_str))
def test_parseMetadataMap(self):
"""Test parsing a mapping file into a MetadataMap instance."""
obs = MetadataMap.parseMetadataMap(self.overview_map_str)
self.assertEqual(obs, self.overview_map)
def test_parseMetadataMap_empty(self):
"""Test parsing empty mapping file contents."""
self.assertRaises(QiimeParseError, MetadataMap.parseMetadataMap, [])
def test_eq(self):
"""Test whether two MetadataMaps are equal."""
self.assertTrue(self.empty_map == MetadataMap({}, []))
self.assertTrue(self.overview_map == MetadataMap(
self.overview_map._metadata, self.overview_map.Comments))
def test_ne(self):
"""Test whether two MetadataMaps are not equal."""
self.assertTrue(self.empty_map != MetadataMap({}, ["foo"]))
self.assertTrue(self.overview_map != MetadataMap(
self.overview_map._metadata, ["foo"]))
self.assertTrue(self.overview_map != MetadataMap({},
self.overview_map.Comments))
self.assertTrue(self.overview_map != self.empty_map)
self.assertTrue(self.overview_map != self.map_with_comments)
self.assertTrue(self.overview_map != self.no_metadata)
def test_getSampleMetadata(self):
"""Test metadata by sample ID accessor with valid sample IDs."""
exp = {'BarcodeSequence': 'AGCACGAGCCTA', 'Treatment': 'Control',
'DOB': '20061218', 'Description': '354'}
obs = self.overview_map.getSampleMetadata('PC.354')
self.assertEqual(obs, exp)
exp = {'BarcodeSequence': 'ACCAGCGACTAG', 'Treatment': 'Control',
'DOB': '20070314', 'Description': '481'}
obs = self.map_with_comments.getSampleMetadata('PC.481')
self.assertEqual(obs, exp)
exp = {'BarcodeSequence': 'ACGGTGAGTGTC', 'Treatment': 'Fast',
'DOB': '20080116', 'Description': '636'}
obs = self.map_with_comments.getSampleMetadata('PC.636')
self.assertEqual(obs, exp)
exp = {}
obs = self.no_metadata.getSampleMetadata('PC.636')
self.assertEqual(obs, exp)
def test_getSampleMetadata_bad_sample_id(self):
"""Test metadata by sample ID accessor with invalid sample IDs."""
# Nonexistent sample ID.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata,
'PC.000')
self.assertRaises(KeyError, self.no_metadata.getSampleMetadata,
'PC.000')
# Integer sample ID.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata, 42)
# Sample ID of type None.
self.assertRaises(KeyError, self.overview_map.getSampleMetadata, None)
# Sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, 's1')
# Integer sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, 1)
# Sample ID of None on empty map.
self.assertRaises(KeyError, self.empty_map.getSampleMetadata, None)
def test_getCategoryValue(self):
"""Test category value by sample ID/category name accessor."""
exp = "Fast"
obs = self.overview_map.getCategoryValue('PC.634', 'Treatment')
self.assertEqual(obs, exp)
exp = "20070314"
obs = self.overview_map.getCategoryValue('PC.481', 'DOB')
self.assertEqual(obs, exp)
exp = "ACGGTGAGTGTC"
obs = self.map_with_comments.getCategoryValue(
'PC.636', 'BarcodeSequence')
self.assertEqual(obs, exp)
def test_getCategoryValues(self):
"""Test category value list by sample ID/category name accessor."""
smpl_ids = ['PC.354', 'PC.355', 'PC.356', 'PC.481', 'PC.593', 'PC.607',
'PC.634', 'PC.635', 'PC.636']
exp = ['Control','Control','Control','Control','Control','Fast'
,'Fast','Fast','Fast']
obs = self.overview_map.getCategoryValues(smpl_ids, 'Treatment')
self.assertEqual(obs, exp)
def test_isNumericCategory(self):
"""Test checking if a category is numeric."""
obs = self.overview_map.isNumericCategory('Treatment')
self.assertEqual(obs, False)
obs = self.overview_map.isNumericCategory('DOB')
self.assertEqual(obs, True)
def test_hasUniqueCategoryValues(self):
"""Test checking if a category has unique values."""
obs = self.overview_map.hasUniqueCategoryValues('Treatment')
self.assertEqual(obs, False)
obs = self.overview_map.hasUniqueCategoryValues('DOB')
self.assertEqual(obs, False)
obs = self.overview_map.hasUniqueCategoryValues('Description')
self.assertEqual(obs, True)
def test_hasSingleCategoryValue(self):
"""Test checking if a category has only a single value."""
obs = self.overview_map.hasSingleCategoryValue('Treatment')
self.assertEqual(obs, False)
obs = self.single_value.hasSingleCategoryValue('Foo')
self.assertEqual(obs, True)
def test_getCategoryValue_bad_sample_id(self):
"""Test category value by sample ID accessor with bad sample IDs."""
# Nonexistent sample ID.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.000', 'Treatment')
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.000', 'Treatment')
# Integer sample ID.
self.assertRaises(KeyError, self.overview_map.getCategoryValue, 42,
'DOB')
# Sample ID of type None.
self.assertRaises(KeyError, self.overview_map.getCategoryValue, None,
'Treatment')
# Sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, 's1',
'foo')
# Integer sample ID on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, 1,
'bar')
# Sample ID of None on empty map.
self.assertRaises(KeyError, self.empty_map.getCategoryValue, None,
'baz')
def test_getCategoryValue_bad_category(self):
"""Test category value by sample ID accessor with bad categories."""
# Nonexistent category.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', 'foo')
# Integer category.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', 42)
# Category of type None.
self.assertRaises(KeyError, self.overview_map.getCategoryValue,
'PC.354', None)
# Category on map with no metadata, but that has sample IDs.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', 'Treatment')
# Integer category on map with no metadata.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', 34)
# Category of type None on map with no metadata.
self.assertRaises(KeyError, self.no_metadata.getCategoryValue,
'PC.354', None)
def test_SampleIds(self):
"""Test sample IDs accessor."""
exp = ["PC.354", "PC.355", "PC.356", "PC.481", "PC.593", "PC.607",
"PC.634", "PC.635", "PC.636"]
obs = self.overview_map.SampleIds
self.assertEqual(obs, exp)
obs = self.no_metadata.SampleIds
self.assertEqual(obs, exp)
obs = self.empty_map.SampleIds
self.assertEqual(obs, [])
def test_CategoryNames(self):
"""Test category names accessor."""
exp = ["BarcodeSequence", "DOB", "Description", "Treatment"]
obs = self.overview_map.CategoryNames
self.assertEqual(obs, exp)
obs = self.no_metadata.CategoryNames
self.assertEqual(obs, [])
obs = self.empty_map.CategoryNames
self.assertEqual(obs, [])
def test_filterSamples(self):
"""Test filtering out samples from metadata map."""
exp = ['PC.356', 'PC.593']
self.overview_map.filterSamples(['PC.593', 'PC.356'])
obs = self.overview_map.SampleIds
self.assertEqual(obs, exp)
self.overview_map.filterSamples([])
self.assertEqual(self.overview_map.SampleIds, [])
def test_filterSamples_strict(self):
"""Test strict checking of sample prescence when filtering."""
with self.assertRaises(ValueError):
self.overview_map.filterSamples(['PC.356', 'abc123'])
with self.assertRaises(ValueError):
self.empty_map.filterSamples(['foo'])
def test_filterSamples_no_strict(self):
"""Test missing samples does not raise error."""
self.overview_map.filterSamples(['PC.356', 'abc123'], strict=False)
self.assertEqual(self.overview_map.SampleIds, ['PC.356'])
self.empty_map.filterSamples(['foo'], strict=False)
self.assertEqual(self.empty_map.SampleIds, [])
def test_is_valid_git_refname(self):
"""Test correct validation of refnames"""
# valid branchnames
self.assertTrue(is_valid_git_refname('master'))
self.assertTrue(is_valid_git_refname('debuggatron_2000'))
self.assertTrue(is_valid_git_refname('refname/bar'))
self.assertTrue(is_valid_git_refname('ref.nameslu/_eggs_/spam'))
self.assertTrue(is_valid_git_refname('valid{0}char'.format(
unichr(40))))
self.assertTrue(is_valid_git_refname('master@head'))
self.assertTrue(is_valid_git_refname('bar{thing}foo'))
# case happening with git < 1.6.6
self.assertFalse(is_valid_git_refname(
'--abbrev-ref\nbaa350d7b7063d585ca293fc16ef15e0765dc9ee'))
# different invalid refnames, for a description of each group see the
# man page of git check-ref-format
self.assertFalse(is_valid_git_refname('bar/.spam/eggs'))
self.assertFalse(is_valid_git_refname('bar.lock/spam/eggs'))
self.assertFalse(is_valid_git_refname('bar.lock'))
self.assertFalse(is_valid_git_refname('.foobar'))
self.assertFalse(is_valid_git_refname('ref..name'))
self.assertFalse(is_valid_git_refname(u'invalid{0}char'.format(
unichr(177))))
self.assertFalse(is_valid_git_refname('invalid{0}char'.format(
unichr(39))))
self.assertFalse(is_valid_git_refname('ref~name/bar'))
self.assertFalse(is_valid_git_refname('refname spam'))
self.assertFalse(is_valid_git_refname('bar/foo/eggs~spam'))
self.assertFalse(is_valid_git_refname('bar:_spam_'))
self.assertFalse(is_valid_git_refname('eggtastic^2'))
self.assertFalse(is_valid_git_refname('areyourandy?'))
self.assertFalse(is_valid_git_refname('bar/*/spam'))
self.assertFalse(is_valid_git_refname('bar[spam]/eggs'))
self.assertFalse(is_valid_git_refname('/barfooeggs'))
self.assertFalse(is_valid_git_refname('barfooeggs/'))
self.assertFalse(is_valid_git_refname('bar/foo//////eggs'))
self.assertFalse(is_valid_git_refname('dotEnding.'))
self.assertFalse(is_valid_git_refname('@{branch'))
self.assertFalse(is_valid_git_refname('contains\\slash'))
self.assertFalse(is_valid_git_refname('$newbranch'))
def test_is_valid_git_sha1(self):
""" """
# valid sha1 strings
self.assertTrue(is_valid_git_sha1(
'65a9ba2ef4b126fb5b054ea6b89b457463db4ec6'))
self.assertTrue(is_valid_git_sha1(
'a29a9911e41253405494c43889925a6d79ca26db'))
self.assertTrue(is_valid_git_sha1(
'e099cd5fdea89eba929d6051fbd26cc9e7a0c961'))
self.assertTrue(is_valid_git_sha1(
'44235d322c3386bd5ce872d9d7ea2e10d27c86cb'))
self.assertTrue(is_valid_git_sha1(
'7d2fc23E04540EE92c742948cca9ed5bc54d08d1'))
self.assertTrue(is_valid_git_sha1(
'fb5dc0285a8b11f199c4f3a7547a2da38138373f'))
self.assertTrue(is_valid_git_sha1(
'0b2abAEb195ba7ebc5cfdb53213a66fbaddefdb8'))
# invalid length
self.assertFalse(is_valid_git_sha1('cca9ed5bc54d08d1'))
self.assertFalse(is_valid_git_sha1(''))
# invalid characters
self.assertFalse(is_valid_git_sha1(
'fb5dy0f85a8b11f199c4f3a75474a2das8138373'))
self.assertFalse(is_valid_git_sha1(
'0x5dcc816fbc1c2e8eX087d7d2ed8d2950a7c16b'))
def preprocess_mapping_file(data,
headers,
columns,
unique=False,
single=False,
clones=0):
"""Process a mapping file to expand the data or remove unuseful fields
Parameters
----------
data: list of list of str
mapping file data
headers: list of str
mapping file headers
columns: list of str
headers to keep, if one of these headers includes two ampersands,
this function will create a new column by merging the delimited columns
unique: bool, optional
keep columns where all values are unique. Default: false.
single: bool, optional
keep columns where all values are the same. Default: false
clones: int, optional
number of times to replicate the metadata. Default: 0
Returns
-------
data: list of list of str
processed mapping file data
headers: list of str
processed mapping file headers
"""
# The sample ID must always be there, else it's meaningless data
if 'SampleID' != columns[0]:
columns = ['SampleID'] + columns
# process concatenated columns if needed
merge = []
for column in columns:
# the list can contain None so check "if column" before treating as str
if column and '&&' in column:
merge.append(column)
# each element needs several columns to be merged
for new_column in merge:
indices = [
headers.index(header_name)
for header_name in new_column.split('&&')
]
# join all the fields of the metadata that are listed in indices
for line in data:
line.append(''.join([line[index] for index in indices]))
headers.append(new_column)
# remove all unique or singled valued columns that are not included in
# the list of categories that should be kept i. e. columns
if unique or single:
columns_to_remove = []
metadata = MetadataMap(mapping_file_to_dict(data, headers), [])
# the --coloy_by option in the script interface allows the user to
# specify the categories you want to use in the generated plot, this
# the default behaviour is to color by all categories that are not
# unique. If the user specifies a category with with the --color_by
# option and this category contains a unique values, this category must
# still be added thus the structure of the next few lines that
# form the structure for the two different routes. (1) where no value
# is specified in the CLI (the value of columns will be [None, x1, x2,
# x3] where x{1,2,3} are categories requested in other CLI options) and
# (2) where a value is specified in the CLI.
#
# TL;DR
# see https://github.com/biocore/emperor/issues/271
if None in columns:
columns = headers[:]
f_unique = metadata.hasUniqueCategoryValues
f_single = metadata.hasSingleCategoryValue
else:
def f_unique(x):
return metadata.hasUniqueCategoryValues(x) and x not in columns
def f_single(x):
return metadata.hasSingleCategoryValue(x) and x not in columns
# find columns that have values that are all unique
if unique:
for c in headers[1::]:
if f_unique(c):
columns_to_remove.append(c)
# remove categories where there is only one value
if single:
for c in headers[1::]:
if f_single(c):
columns_to_remove.append(c)
columns_to_remove = list(set(columns_to_remove))
# remove the single or unique columns
data, headers = keep_columns_from_mapping_file(data,
headers,
columns_to_remove,
negate=True)
else:
# when a None is contained in columns, we imply we want to use all the
# available categories in the mapping file, thus just overwrite the
# value
if None in columns:
columns = headers[:]
# remove anything not specified in the input
data, headers = keep_columns_from_mapping_file(data, headers, columns)
# sanitize the mapping file data and headers
data, headers = sanitize_mapping_file(data, headers)
# clones mean: replicate the metadata retagging the sample ids with a
# suffix
if clones:
out_data = []
for index in range(0, clones):
out_data.extend([[element[0] + '_%d' % index] + element[1::]
for element in data])
data = out_data
return data, headers
# check for all the mapping fields
for col in color_by_column_names:
# for concatenated columns check each individual field
parts = col.split('&&')
offending_fields.extend(p for p in parts if p not in lookup_header)
else:
# if the user didn't specify the header names display everything
color_by_column_names = [None]
# extract a list of the custom axes provided and each element is numeric
if custom_axes:
custom_axes = custom_axes.strip().strip("'").strip('"').split(',')
# the MetadataMap object makes some checks easier
map_object = MetadataMap(mapping_file_to_dict(mapping_data, header),
[])
for axis in custom_axes:
# append the field to the error queue that it belongs to
if axis not in lookup_header:
offending_fields.append(axis)
break
# make sure this value is in the mapping file
elif axis not in color_by_column_names:
color_by_column_names.append(axis)
# perform only if the for loop does not call break
else:
# make sure all these axes are numeric
for axis in custom_axes:
if not map_object.isNumericCategory(axis):
non_numeric_categories.append(axis)
# check for all the mapping fields
for col in color_by_column_names:
# for concatenated columns check each individual field
parts = col.split('&&')
offending_fields.extend(p for p in parts if p not in lookup_header)
else:
# if the user didn't specify the header names display everything
color_by_column_names = [None]
# extract a list of the custom axes provided and each element is numeric
if custom_axes:
custom_axes = custom_axes.strip().strip("'").strip('"').split(',')
# the MetadataMap object makes some checks easier
map_object = MetadataMap(mapping_file_to_dict(mapping_data, header),
[])
for axis in custom_axes:
# append the field to the error queue that it belongs to
if axis not in lookup_header:
offending_fields.append(axis)
break
# make sure this value is in the mapping file
elif axis not in color_by_column_names:
color_by_column_names.append(axis)
# perform only if the for loop does not call break
else:
# make sure all these axes are numeric
for axis in custom_axes:
if not map_object.isNumericCategory(axis):
non_numeric_categories.append(axis)
def preprocess_mapping_file(data, headers, columns, unique=False, single=False,
clones=0):
"""Process a mapping file to expand the data or remove unuseful fields
Inputs:
data: mapping file data
headers: mapping file headers
columns: list of headers to keep, if one of these headers includes two
ampersands, this function will create a new column by merging the delimited
columns.
unique: keep columns where all values are unique
single: keep columns where all values are the same
clones: number of times to replicate the metadata
Outputs:
data: processed mapping file data
headers: processed mapping file headers
"""
# The sample ID must always be there, else it's meaningless data
if 'SampleID' != columns[0]:
columns = ['SampleID'] + columns
# process concatenated columns if needed
merge = []
for column in columns:
if '&&' in column:
merge.append(column)
# each element needs several columns to be merged
for new_column in merge:
indices = [headers.index(header_name) for header_name in
new_column.split('&&')]
# join all the fields of the metadata that are listed in indices
for line in data:
line.append(''.join([line[index] for index in indices]))
headers.append(new_column)
# remove all unique or singled valued columns
if unique or single:
columns_to_remove = []
metadata = MetadataMap(mapping_file_to_dict(data, headers), [])
# find columns that have values that are all unique
if unique == True:
columns_to_remove += [column_name for column_name in headers[1::]
if metadata.hasUniqueCategoryValues(column_name)]
# remove categories where there is only one value
if single == True:
columns_to_remove += [column_name for column_name in headers[1::]
if metadata.hasSingleCategoryValue(column_name)]
columns_to_remove = list(set(columns_to_remove))
# remove the single or unique columns
data, headers = keep_columns_from_mapping_file(data, headers,
columns_to_remove, negate=True)
# remove anything not specified in the input
data, headers = keep_columns_from_mapping_file(data, headers, columns)
# sanitize the mapping file data and headers
data, headers = sanitize_mapping_file(data, headers)
# clones mean: replicate the metadata retagging the sample ids with a suffix
if clones:
out_data = []
for index in range(0, clones):
out_data.extend([[element[0]+'_%d' % index]+element[1::]
for element in data])
data = out_data
return data, headers
def preprocess_mapping_file(data,
headers,
columns,
unique=False,
single=False,
clones=0):
"""Process a mapping file to expand the data or remove unuseful fields
Inputs:
data: mapping file data
headers: mapping file headers
columns: list of headers to keep, if one of these headers includes two
ampersands, this function will create a new column by merging the delimited
columns.
unique: keep columns where all values are unique
single: keep columns where all values are the same
clones: number of times to replicate the metadata
Outputs:
data: processed mapping file data
headers: processed mapping file headers
"""
# The sample ID must always be there, else it's meaningless data
if 'SampleID' != columns[0]:
columns = ['SampleID'] + columns
# process concatenated columns if needed
merge = []
for column in columns:
if '&&' in column:
merge.append(column)
# each element needs several columns to be merged
for new_column in merge:
indices = [
headers.index(header_name)
for header_name in new_column.split('&&')
]
# join all the fields of the metadata that are listed in indices
for line in data:
line.append(''.join([line[index] for index in indices]))
headers.append(new_column)
# remove all unique or singled valued columns
if unique or single:
columns_to_remove = []
metadata = MetadataMap(mapping_file_to_dict(data, headers), [])
# find columns that have values that are all unique
if unique == True:
columns_to_remove += [
column_name for column_name in headers[1::]
if metadata.hasUniqueCategoryValues(column_name)
]
# remove categories where there is only one value
if single == True:
columns_to_remove += [
column_name for column_name in headers[1::]
if metadata.hasSingleCategoryValue(column_name)
]
columns_to_remove = list(set(columns_to_remove))
# remove the single or unique columns
data, headers = keep_columns_from_mapping_file(data,
headers,
columns_to_remove,
negate=True)
# remove anything not specified in the input
data, headers = keep_columns_from_mapping_file(data, headers, columns)
# sanitize the mapping file data and headers
data, headers = sanitize_mapping_file(data, headers)
# clones mean: replicate the metadata retagging the sample ids with a suffix
if clones:
out_data = []
for index in range(0, clones):
out_data.extend([[element[0] + '_%d' % index] + element[1::]
for element in data])
data = out_data
return data, headers
if '&&' in col:
for _col in col.split('&&'):
if _col not in lookup_header:
offending_fields.append(col)
elif col not in lookup_header:
offending_fields.append(col)
else:
# if the user didn't specify the header names display everything
color_by_column_names = [None]
# extract a list of the custom axes provided and each element is numeric
if custom_axes:
custom_axes = custom_axes.strip().strip("'").strip('"').split(',')
# the MetadataMap object makes some checks easier
map_object = MetadataMap(mapping_file_to_dict(mapping_data, header), [])
for axis in custom_axes:
# append the field to the error queue that it belongs to
if axis not in lookup_header:
offending_fields.append(axis)
break
# make sure this value is in the mapping file
elif axis not in color_by_column_names:
color_by_column_names.append(axis)
# perform only if the for loop does not call break
else:
# make sure all these axes are numeric
for axis in custom_axes:
if map_object.isNumericCategory(axis) == False:
non_numeric_categories.append(axis)
def main():
#option_parser, opts, args = parse_command_line_parameters(**script_info)
input_coords = args.input_coords
map_fp = args.map_fp
output_dir = args.output_dir
color_by_column_names = None #opts.color_by
add_unique_columns = False #opts.add_unique_columns
custom_axes = None #opts.custom_axes
ignore_missing_samples = False #opts.ignore_missing_samples
missing_custom_axes_values = None #opts.missing_custom_axes_values
jackknifing_method = 'IQR' #opts.ellipsoid_method
master_pcoa = None #opts.master_pcoa
taxa_fp = None #opts.taxa_fp
n_taxa_to_keep = False #opts.n_taxa_to_keep
biplot_fp = None #opts.biplot_fp
add_vectors = [None, None] #opts.add_vectors
verbose_output = False #opts.verbose
number_of_axes = 10 #opts.number_of_axes
compare_plots = False #opts.compare_plots
number_of_segments = 8 #opts.number_of_segments
pct_variation_below_one = True #opts.pct_variation_below_one
# add some metadata to the output
emperor_autograph = format_emperor_autograph(map_fp, input_coords, 'HTML')
# verifying that the number of axes requested is greater than 3
if number_of_axes < 3:
print(('You need to plot at least 3 axes.'))
# verifying that the number of segments is between the desired range
if not (4 <= number_of_segments <= 14):
print(('number_of_segments should be between 4 and 14.'))
# append headernames that the script didn't find in the mapping file
# according to different criteria to the following variables
offending_fields = []
non_numeric_categories = []
serial_comparison = True
# can't do averaged pcoa plots _and_ custom axes in the same plot
if custom_axes is not None and isdir(input_coords):
if custom_axes.count(',') > 0:
print(('Jackknifed plots are limited to one custom '
'axis, currently trying to use: %s. Make '
'sure you use only one.' % custom_axes))
# make sure the flag is not misunderstood from the command line interface
if not isdir(input_coords) and compare_plots:
print("Cannot use the '--compare_plots' flag unless the "
"input path is a directory.")
# before creating any output, check correct parsing of the main input files
#try:
mapping_data, header, comments = parse_mapping_file(open(map_fp, 'U'))
try:
pass
except:
sys.exit(("The metadata mapping file '%s' does not seem "
"to be formatted correctly, verify the "
"formatting is QIIME compliant by using "
"validate_mapping_file.py") % map_fp)
else:
# use this set variable to make presence/absensce checks faster
lookup_header = set(header)
mapping_ids = {row[0] for row in mapping_data}
# dir means jackknifing or coordinate comparison type of processing
if isdir(input_coords):
offending_coords_fp = []
coords_headers = []
coords_data = []
coords_eigenvalues = []
coords_pct = []
coord_fps = guess_coordinates_files(input_coords)
# QIIME generates folders of transformed coordinates for the specific
# purpose of connecting all coordinates to a set of origin coordinates.
# The name of this file is suffixed as _transformed_reference.txt
trans_suf = '_transformed_reference.txt'
transformed = [f for f in coord_fps if f.endswith(trans_suf)]
# this could happen and we rather avoid this problem
if len(coord_fps) == 0:
print('Could not use any of the files in the input '
'directory.')
# the master pcoa must be the first in the list of coordinates; however
# if the visualization is not a jackknifed plot this gets ignored
if master_pcoa and not compare_plots:
if master_pcoa in coord_fps: # remove it if duplicated
coord_fps.remove(master_pcoa)
coord_fps = [master_pcoa] + coord_fps # prepend it to the list
# passing a master file means that the comparison is not serial
elif master_pcoa and compare_plots:
serial_comparison = False
# guarantee that the master is the first and is not repeated
if master_pcoa in coord_fps:
coord_fps.remove(master_pcoa)
sorted_filenames = sort_comparison_filenames(coord_fps)
coord_fps = [master_pcoa] + sorted_filenames
elif master_pcoa is None and len(transformed):
master_pcoa = transformed[0]
serial_comparison = False
# Note: the following steps are to guarantee consistency.
# remove the master from the list and re-add it as a first element
# the rest of the files must be sorted alphabetically so the result
# will be: ['unifrac_transformed_reference.txt',
# 'unifrac_transformed_q1.txt', 'unifrac_transformed_q2.txt'] etc
coord_fps.remove(master_pcoa)
coord_fps = [master_pcoa] + sort_comparison_filenames(coord_fps)
for fp in coord_fps:
try:
parsed = parse_coords(open(fp, 'U'))
except (ValueError, QiimeParseError):
offending_coords_fp.append(fp)
# do not add any of the data and move along
continue
else:
# pack all the data correspondingly only if it was correctly
# parsed
coords_headers.append(parsed[0])
coords_data.append(parsed[1])
coords_eigenvalues.append(parsed[2])
coords_pct.append(parsed[3])
# in case there were files that couldn't be parsed
if offending_coords_fp:
errout = ', '.join(offending_coords_fp)
sys.exit(("The following file(s): '%s' could not be "
"parsed properly. Make sure the input folder "
"only contains coordinates files.") % errout)
# check all files contain the same sample identifiers by flattening the
# list of available sample ids and returning the sample ids that are
# in one of the sets of sample ids but not in the globablly shared ids
_coords_headers = set(flatten(coords_headers))
_per_file_missing = [_coords_headers - set(e) for e in coords_headers]
non_shared_ids = set(flatten(_per_file_missing))
if non_shared_ids:
errout = ', '.join(non_shared_ids)
sys.exit(("The following sample identifier(s): '%s' "
"are not shared between all the files. The "
"files used to make a jackknifed PCoA plot "
"or coordinate comparison plot (procustes "
"plot) must share all the same sample "
"identifiers between each other.") % errout)
# number of samples ids that are shared between coords and mapping
# files
sids_intersection = mapping_ids.intersection(_coords_headers)
# sample ids that are not mapped but are in the coords
sids_difference = _coords_headers.difference(mapping_ids)
# used to perform different validations in the script, very similar for
# the case where the input is not a directory
number_intersected_sids = len(sids_intersection)
required_number_of_sids = len(coords_headers[0])
else:
try:
parsed = parse_coords(open(input_coords, 'U'))
# this exception was noticed when there were letters in the coords file
# other exeptions should be catched here; code will be updated then
except (ValueError, QiimeParseError):
sys.exit(("The PCoA file '%s' does not seem to be a "
"coordinates formatted file, verify by "
"manually inspecting the contents.") %
input_coords)
else:
coords_headers = parsed[0]
coords_data = parsed[1]
coords_eigenvalues = parsed[2]
coords_pct = parsed[3]
# number of samples ids that are shared between coords and mapping
# files
sids_intersection = mapping_ids.intersection(coords_headers)
# sample ids that are not mapped but are in the coords
sids_difference = set(coords_headers).difference(mapping_ids)
number_intersected_sids = len(sids_intersection)
required_number_of_sids = len(coords_headers)
if taxa_fp:
try:
# This should really use BIOM's Table.from_tsv
# for summarized tables the "otu_ids" are really the "lineages"
parsed = parse_otu_table(open(taxa_fp, 'U'), count_map_f=float,
remove_empty_rows=True)
except ValueError(e):
sys.exit(("There was a problem parsing the --taxa_fp: "
"%s" % e.message))
else:
otu_sample_ids = parsed[0]
lineages = parsed[1]
otu_table = parsed[2]
# make sure there are matching sample ids with the otu table
if not sids_intersection.issuperset(otu_sample_ids):
sys.exit("The sample identifiers in the OTU table must "
"have at least one match with the data in the "
"mapping file and with the coordinates file. "
"Verify you are using input files that belong "
"to the same dataset.")
if len(lineages) <= 1:
sys.exit("Contingency tables with one or fewer rows "
"are not supported, please try passing a "
"contingency table with more than one row.")
else:
# empty lists indicate that there was no taxa file passed in
otu_sample_ids, lineages, otu_table = [], [], []
# sample ids must be shared between files
if number_intersected_sids <= 0:
sys.exit('None of your sample identifiers match between the'
' mapping file and the coordinates file. Verify '
'you are using a coordinates file and a mapping '
'file that belong to the same dataset.')
# the intersection of the sample ids in the coords and the sample ids in
# the mapping file must at the very least include all ids in the coords
# file Otherwise it isn't valid; unless --ignore_missing_samples is set
# True
if number_intersected_sids != required_number_of_sids:
if ignore_missing_samples:
# keep only the samples that are mapped in the mapping file
coords_headers, coords_data = keep_samples_from_pcoa_data(
coords_headers, coords_data, sids_intersection)
else:
message = ("The metadata mapping file has fewer sample "
"identifiers than the coordinates file. Verify you are "
"using a mapping file that contains at least all the "
"samples contained in the coordinates file(s). You can "
"force the script to ignore these samples by passing "
"the '--ignore_missing_samples' flag.")
if verbose_output:
missing_ids = ', '.join(sids_difference)
message += ' Offending sample identifier(s): %s.' % missing_ids
sys.exit(message)
# ignore samples that exist in the coords but not in the mapping file,
# note: we're using sids_intersection so if --ignore_missing_samples is
# enabled we account for unmapped coords, else the program will exit before
# this point
header, mapping_data = filter_mapping_file(mapping_data, header,
sids_intersection,
include_repeat_cols=True)
# catch the errors that could occur when filling the mapping file values
if missing_custom_axes_values:
# the fact that this uses parse_metadata_state_descriptions makes the
# following option '-x Category:7;PH:12' to work as well as the
# script-interface-documented '-x Category:7 -x PH:12' option
for val in missing_custom_axes_values:
if ':' not in val:
sys.exit("Not valid missing value for custom "
"axes: %s" % val)
_mcav = ';'.join(missing_custom_axes_values)
try:
mapping_data = fill_mapping_field_from_mapping_file(mapping_data,
header, _mcav)
except AssertionError(e):
print(e.message)
except EmperorInputFilesError(e):
print(e.message)
# check that all the required columns exist in the metadata mapping file
if color_by_column_names:
color_by_column_names = color_by_column_names.split(',')
# check for all the mapping fields
for col in color_by_column_names:
# for concatenated columns check each individual field
parts = col.split('&&')
offending_fields.extend(p for p in parts if p not in lookup_header)
else:
# if the user didn't specify the header names display everything
color_by_column_names = [None]
# extract a list of the custom axes provided and each element is numeric
if custom_axes:
custom_axes = custom_axes.strip().strip("'").strip('"').split(',')
# the MetadataMap object makes some checks easier
map_object = MetadataMap(mapping_file_to_dict(mapping_data, header),
[])
for axis in custom_axes:
# append the field to the error queue that it belongs to
if axis not in lookup_header:
offending_fields.append(axis)
break
# make sure this value is in the mapping file
elif axis not in color_by_column_names:
color_by_column_names.append(axis)
# perform only if the for loop does not call break
else:
# make sure all these axes are numeric
for axis in custom_axes:
if not map_object.isNumericCategory(axis):
non_numeric_categories.append(axis)
# make multiple checks for the add_vectors option
if add_vectors != [None, None]:
add_vectors = add_vectors.split(',')
# check there are at the most two categories specified for this option
if len(add_vectors) > 2:
print("The '--add_vectors' option can accept up to "
"two different fields from the mapping file; "
"currently trying to use %d (%s)." %
(len(add_vectors), ', '.join(add_vectors)))
# make sure the field(s) exist
for col in add_vectors:
# concatenated fields are allowed now so check for each field
if '&&' in col:
for _col in col.split('&&'):
if _col not in lookup_header:
offending_fields.append(col)
break
# only execute this block of code if all checked fields exist
else:
# make sure that if it's going to be used for vector
# creation it gets used for coloring and map postprocessing
if col not in color_by_column_names:
color_by_column_names.append(col)
# if it's a column without concatenations
elif col not in lookup_header:
offending_fields.append(col)
break
else:
# check this vector value is in the color by category
if col not in color_by_column_names:
color_by_column_names.append(col)
# perform only if the for loop does not call break
else:
# check that the second category is all with numeric values
if len(add_vectors) == 2:
map_object = MetadataMap(mapping_file_to_dict(mapping_data,
header),
[])
# if it has non-numeric values add it to the list of offenders
if not map_object.isNumericCategory(add_vectors[1]):
msg = add_vectors[1] + '(used in --add_vectors)'
non_numeric_categories.append(msg)
else:
add_vectors.append(None)
# terminate the program for the cases where a mapping field was not found
# or when a mapping field didn't meet the criteria of being numeric
if offending_fields:
sys.exit("Invalid field(s) '%s'; the valid field(s) are: "
"'%s'" % (', '.join(offending_fields),
', '.join(header)))
if non_numeric_categories:
sys.exit(("The following field(s): '%s' contain values "
"that are not numeric, hence not suitable for "
"'--custom_axes' nor for '--add_vectors'. Try "
"the '--missing_custom_axes_values' option to "
"fix these values." %
', '.join(non_numeric_categories)))
# process the coordinates file first, preventing the case where the custom
# axes is not in the coloring categories i. e. in the --colory_by
# categories
preprocessed_coords = preprocess_coords_file(coords_headers, coords_data,
coords_eigenvalues,
coords_pct, header,
mapping_data, custom_axes,
jackknifing_method,
compare_plots,
pct_variation_below_one)
coords_headers = preprocessed_coords[0]
coords_data = preprocessed_coords[1]
coords_eigenvalues = preprocessed_coords[2]
coords_pct = preprocessed_coords[3]
coords_low = preprocessed_coords[4]
coords_high = preprocessed_coords[5]
clones = preprocessed_coords[6]
# process the otu table after processing the coordinates to get custom axes
# (when available) or any other change that occurred to the coordinates
preprocessed_otu_table = preprocess_otu_table(otu_sample_ids, otu_table,
lineages, coords_data,
coords_headers,
n_taxa_to_keep)
otu_coords = preprocessed_otu_table[0]
otu_table = preprocessed_otu_table[1]
otu_lineages = preprocessed_otu_table[2]
otu_prevalence = preprocessed_otu_table[3]
lines = preprocessed_otu_table[4]
# remove the columns in the mapping file that are not informative taking
# into account the header names that were already authorized to be used
# and take care of concatenating the fields for the && merged columns
mapping_data, header = preprocess_mapping_file(mapping_data, header,
color_by_column_names,
not add_unique_columns,
clones=clones)
# create the output directory before creating any other output
if not isdir(opts.output_dir):
makedirs(opts.output_dir)
fp_out = open(join(output_dir, 'index.html'), 'w')
fp_out.write(emperor_autograph+'\n')
fp_out.write(EMPEROR_HEADER_HTML_STRING)
# write the html file
fp_out.write(format_mapping_file_to_js(mapping_data, header, header))
# certain percents being explained cannot be displayed in the GUI
try:
fp_out.write(format_pcoa_to_js(coords_headers, coords_data,
coords_eigenvalues, coords_pct,
custom_axes, coords_low,
coords_high,
number_of_axes=number_of_axes,
number_of_segments=number_of_segments))
except EmperorLogicError(e):
sys.exit(e.message)
fp_out.write(format_taxa_to_js(otu_coords, otu_lineages, otu_prevalence))
fp_out.write(format_vectors_to_js(mapping_data, header, coords_data,
coords_headers, add_vectors[0],
add_vectors[1]))
fp_out.write(format_comparison_bars_to_js(coords_data, coords_headers,
clones, serial_comparison))
has_taxa = taxa_fp is not None
has_input_coords = isdir(input_coords) and not compare_plots
has_add_vectors = add_vectors != [None, None]
has_clones = clones > 0
fp_out.write(format_emperor_html_footer_string(has_taxa, has_input_coords,
has_add_vectors,
has_clones))
fp_out.close()
copy_support_files(output_dir)
# write the biplot coords in the output file if a path is passed
if biplot_fp and taxa_fp:
if biplot_fp.endswith('/') or isdir(biplot_fp):
print("Do not specify a path to a new (path ending "
"in a slash) or existing directory for "
"biplot_fp. The output file will be a "
"tab-delimited text file.")
# make sure this file can be created
try:
fd = open(biplot_fp, 'w')
except IOError:
sys.exit("There was a problem creating the file with "
"the coordinates for the biplots (%s)."
% biplot_fp)
else:
fd.writelines(lines)
fd.close()