def test_parse_gtf_lines_without_expand_attributes():
parsed_dict = parse_gtf_lines(gtf_lines, expand_attribute_column=False)
# convert to list since Py3's dictionary keys are a distinct collection type
eq_(list(parsed_dict.keys()), REQUIRED_COLUMNS)
eq_(parsed_dict["seqname"], ["1", "1"])
# convert to list for comparison since numerical columns may be NumPy arrays
eq_(list(parsed_dict["start"]), [11869, 11869])
eq_(list(parsed_dict["end"]), [14409, 14409])
# can't compare NaN with equality
scores = list(parsed_dict["score"])
assert np.isnan(scores).all(), "Unexpected scores: %s" % scores
assert len(parsed_dict["attribute"]) == 2
def test_parse_gtf_lines_with_expand_attributes():
parsed_dict = parse_gtf_lines(gtf_lines, expand_attribute_column=True)
# excluding 'attribute' column from required names
expected_columns = REQUIRED_COLUMNS[:8] + [
"gene_id",
"gene_name",
"gene_source",
"gene_biotype",
"transcript_id",
"transcript_name",
"transcript_source",
]
# convert to list since Py3's dictionary keys are a distinct collection type
eq_(list(parsed_dict.keys()), expected_columns)
eq_(parsed_dict["seqname"], ["1", "1"])
# convert to list for comparison since numerical columns may be NumPy arrays
eq_(list(parsed_dict["start"]), [11869, 11869])
eq_(list(parsed_dict["end"]), [14409, 14409])
# can't compare NaN with equality
scores = list(parsed_dict["score"])
assert np.isnan(scores).all(), "Unexpected scores: %s" % scores
eq_(parsed_dict["gene_id"], ["ENSG00000223972", "ENSG00000223972"])
eq_(parsed_dict["transcript_id"], ["", "ENST00000456328"])
from gtfparse import create_missing_features, parse_gtf_lines
import pandas
# two lines from the Ensembl 54 human GTF containing only a stop_codon and
# exon features, but from which gene and transcript information could be
# inferred
GTF_DATA = """
# seqname biotype feature start end score strand frame attribute
18\tprotein_coding\tstop_codon\t32630766\t32630768\t.\t-\t0\tgene_id "ENSG00000134779"; transcript_id "ENST00000334295"; exon_number "7"; gene_name "C18orf10"; transcript_name "C18orf10-201";
18\tprotein_coding\texon\t32663078\t32663157\t.\t+\t.\tgene_id "ENSG00000150477"; transcript_id "ENST00000383055"; exon_number "1"; gene_name "KIAA1328"; transcript_name "KIAA1328-202";
"""
GTF_LINES = GTF_DATA.split("\n")
GTF_DICT = parse_gtf_lines(GTF_LINES)
GTF_DATAFRAME = pandas.DataFrame(GTF_DICT)
def test_create_missing_features_identity():
df_should_be_same = create_missing_features(GTF_DATAFRAME, {})
assert len(GTF_DATAFRAME) == len(df_should_be_same), \
"GTF DataFrames should be same size"
def _check_expanded_dataframe(df):
assert "gene" in set(df["feature"]), \
"Extended GTF should contain gene feature"
assert "transcript" in set(df["feature"]), \
"Extended GTF should contain transcript feature"
C18orf10_201_transcript_mask = ((df["feature"] == "transcript") &
from gtfparse import create_missing_features, parse_gtf_lines
import pandas
# two lines from the Ensembl 54 human GTF containing only a stop_codon and
# exon features, but from which gene and transcript information could be
# inferred
GTF_DATA = """
# seqname biotype feature start end score strand frame attribute
18\tprotein_coding\tstop_codon\t32630766\t32630768\t.\t-\t0\tgene_id "ENSG00000134779"; transcript_id "ENST00000334295"; exon_number "7"; gene_name "C18orf10"; transcript_name "C18orf10-201";
18\tprotein_coding\texon\t32663078\t32663157\t.\t+\t.\tgene_id "ENSG00000150477"; transcript_id "ENST00000383055"; exon_number "1"; gene_name "KIAA1328"; transcript_name "KIAA1328-202";
"""
GTF_LINES = GTF_DATA.split("\n")
GTF_DICT = parse_gtf_lines(GTF_LINES)
GTF_DATAFRAME = pandas.DataFrame(GTF_DICT)
def test_create_missing_features_identity():
df_should_be_same = create_missing_features(GTF_DATAFRAME, {})
assert len(GTF_DATAFRAME) == len(df_should_be_same), \
"GTF DataFrames should be same size"
def _check_expanded_dataframe(df):
assert "gene" in set(df["feature"]), \
"Extended GTF should contain gene feature"
assert "transcript" in set(df["feature"]), \
"Extended GTF should contain transcript feature"
C18orf10_201_transcript_mask = (
(df["feature"] == "transcript") &
(df["transcript_name"] == "C18orf10-201"))
def test_parse_gtf_lines_error_too_few_fields():
bad_gtf_lines = [line.replace("\t", " ") for line in gtf_lines]
# pylint: disable=no-value-for-parameter
with assert_raises(ParsingError):
parse_gtf_lines(bad_gtf_lines)
