def test_load_config_id_name_attr(self):
row = self.csv_feat_row_dict.copy()
row['Name'] = 'ID'
csv = self.csv("features.csv", [row.values()])
with patch('tiny.rna.configuration.open', mock_open(read_data=csv)):
dummy_file = '/dev/null'
_, gff_files = counter.load_config(dummy_file, False)
# Expect {file: [empty Name Attribute list]}
from_dummy = from_here(dummy_file, row['Source'])
expected = defaultdict(list, zip([from_dummy], [[]]))
self.assertEqual(gff_files, expected)
def test_load_samples_single_cmd(self):
dummy_file = '/dev/null'
inp_file = "test.fastq"
exp_file = from_here(dummy_file, "test_aligned_seqs.sam")
row = {'File': inp_file, 'Group': "test_group", 'Rep': "0"}
csv = self.csv("samples.csv", [row.values()])
with patch('tiny.rna.configuration.open', mock_open(read_data=csv)):
inputs_step = counter.load_samples(dummy_file, is_pipeline=False)
expected_lib_name = f"{row['Group']}_rep_{row['Rep']}"
expected_result = [{'File': exp_file, 'Name': expected_lib_name}]
self.assertEqual(inputs_step, expected_result)
def test_load_config_duplicate_rules(self):
# Features CSV with two duplicate rules/rows
row = self.csv_feat_row_dict.values()
csv = self.csv("features.csv", [row, row]) # Duplicate rows
with patch('tiny.rna.configuration.open', mock_open(read_data=csv)):
dummy_filename = '/dev/null'
ruleset, gff_files = counter.load_config(dummy_filename, False)
r = self.csv_feat_row_dict
expected_gff_file = from_here(dummy_filename, r['Source'])
expected_gff_ret = defaultdict(list,
zip([expected_gff_file], [[r['Name']]]))
expected_ruleset = self.feat_rule
self.assertEqual(gff_files, expected_gff_ret)
self.assertEqual(ruleset, expected_ruleset)
def test_load_config_single_cmd(self):
# Features CSV with a single rule/row
row = self.csv_feat_row_dict.values()
csv = self.csv("features.csv", [row])
with patch('tiny.rna.configuration.open', mock_open(read_data=csv)):
dummy_file = '/dev/null'
ruleset, gff_files = counter.load_config(dummy_file,
is_pipeline=False)
r = self.csv_feat_row_dict
expected_gff_file = from_here(dummy_file, r['Source'])
expected_gff_ret = defaultdict(list,
zip([expected_gff_file], [[r['Name']]]))
expected_ruleset = self.feat_rule
self.assertEqual(gff_files, expected_gff_ret)
self.assertEqual(ruleset, expected_ruleset)
def load_config(features_csv: str,
is_pipeline: bool) -> Tuple[List[dict], Dict[str, list]]:
"""Parses the Features Sheet to provide inputs to FeatureSelector and build_reference_tables
Args:
features_csv: a csv file which defines feature sources and selection rules
is_pipeline: helps locate GFF files defined in the Features Sheet. If true,
GFF files are assumed to reside in the working directory.
Returns:
rules: a list of dictionaries, each representing a parsed row from input.
Note that these are just rule definitions which FeatureSelector will
further digest to produce its rules table.
gff_files: a dict of GFF files and associated Name Attribute preferences
"""
rules, gff_files = list(), defaultdict(list)
for row in CSVReader(features_csv, "Features Sheet").rows():
rule = {
col: row[col]
for col in ["Strand", "Hierarchy", "nt5end", "Length", "Strict"]
}
rule['nt5end'] = rule['nt5end'].upper().translate(
{ord('U'): 'T'}) # Convert RNA base to cDNA base
rule['Identity'] = (row['Key'], row['Value']) # Create identity tuple
rule['Hierarchy'] = int(
rule['Hierarchy']) # Convert hierarchy to number
rule['Strict'] = rule['Strict'].lower(
) # Built later in ReferenceTables
gff = os.path.basename(row['Source']) if is_pipeline else from_here(
features_csv, row['Source'])
# Duplicate Name Attributes and rule entries are not allowed
if row['Name'] not in ["ID", *gff_files[gff]]:
gff_files[gff].append(row['Name'])
if rule not in rules: rules.append(rule)
return rules, gff_files
def get_library_filename(csv_row_file: str, samples_csv: str) -> str:
"""The input samples.csv may contain either fastq or sam files"""
file_ext = os.path.splitext(csv_row_file)[1].lower()
# If the sample file has a fastq(.gz) extension, infer the name of its pipeline-produced .sam file
if file_ext in [".fastq", ".gz"]:
# Fix relative paths to be relative to sample_csv's path, rather than relative to cwd
csv_row_file = os.path.basename(
csv_row_file) if is_pipeline else from_here(
samples_csv, csv_row_file)
csv_row_file = os.path.splitext(
csv_row_file)[0] + "_aligned_seqs.sam"
elif file_ext == ".sam":
if not os.path.isabs(csv_row_file):
raise ValueError(
"The following file must be expressed as an absolute path:\n%s"
% (csv_row_file, ))
else:
raise ValueError(
"The filenames defined in your Samples Sheet must have a .fastq(.gz) or .sam extension.\n"
"The following filename contained neither:\n%s" %
(csv_row_file, ))
return csv_row_file