def test_query_biomart_38_xml(ensembl_biomart_xml_query):
"""Prepare a test xml document for the biomart service build 38
and query the service using it
"""
# GIVEN client with a xml query for a gene
build = "38"
# percent encode query before concatenating with url
url = "".join(
[ensembl_rest_clients.BIOMART_38,
quote(ensembl_biomart_xml_query)])
response = (b"ACTR3\tENST00000263238\n"
b"ACTR3\tENST00000443297\n"
b"ACTR3\tENST00000415792\n"
b"ACTR3\tENST00000446821\n"
b"ACTR3\tENST00000535589\n"
b"ACTR3\tENST00000489779\n"
b"ACTR3\tENST00000484165\n"
b"ACTR3\tENST00000478928\n"
b"[success]")
responses.add(responses.GET, url, body=response, status=200)
# WHEN querying ensembl
client = ensembl_rest_clients.EnsemblBiomartClient(
build=build, xml=ensembl_biomart_xml_query, header=False)
# THEN assert that the result is correct
for line in client:
# THEN assert that either the correct gene is fetched or that an
assert "ACTR3" in line
def test_test_query_biomart_38_xml():
"""Prepare a test xml document for the biomart service build 38 and query the service using it"""
## GIVEN a xml file in biomart format
build = "38"
xml = """<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE Query>
<Query virtualSchemaName = "default" formatter = "TSV" header = "0" uniqueRows = "0" count = "" datasetConfigVersion = "0.6" completionStamp = "1">
<Dataset name = "hsapiens_gene_ensembl" interface = "default" >
<Filter name = "ensembl_gene_id" value = "ENSG00000115091"/>
<Attribute name = "hgnc_symbol" />
<Attribute name = "ensembl_transcript_id" />
</Dataset>
</Query>
"""
## WHEN querying ensembl
client = eracs.EnsemblBiomartClient(build="38", xml=xml, header=False)
## THEN assert that the result is correct
i = 0
for i, line in enumerate(client, 1):
## THEN assert that either the correct gene is fetched or that an HTML page is returned (if the service is down)
assert ("ACTR3" in line or line == "<!doctype html>"
or "BioMart::Exception::Database" in line)
break
def test_xml_attributes():
"""test method that creates attribute lines for the biomart xml file"""
## Given a list of biomart attributes
name = ["test_name"]
client = eracs.EnsemblBiomartClient()
## WHEN creating the xml attribute lines
attribute_lines = client._xml_attributes(name)
## THEN make sure that attributes lines are formatted as they should
assert attribute_lines == ['<Attribute name = "test_name" />']
def test_xml_filters():
"""test method that creates filter lines for the biomart xml file"""
## GIVEN a dictionary of biomart filters
filters = {
'string_filter': 'string_value',
'list_filter': ['1', 'X', 'MT'],
}
client = eracs.EnsemblBiomartClient()
xml_lines = client._xml_filters(filters)
# make sure lines are formatted as they should
assert '<Filter name = "{0}" value = "{1}"/>'.format(
'string_filter', 'string_value') in xml_lines
assert '<Filter name = "{0}" value = "{1}"/>'.format(
'list_filter', '1,X,MT') in xml_lines
def test_xml_filters():
"""test method that creates filter lines for the biomart xml file"""
## GIVEN a dictionary of biomart filters
filters = {
"string_filter": "string_value",
"list_filter": ["1", "X", "MT"]
}
client = eracs.EnsemblBiomartClient()
xml_lines = client._xml_filters(filters)
# make sure lines are formatted as they should
assert ('<Filter name = "{0}" value = "{1}"/>'.format(
"string_filter", "string_value") in xml_lines)
assert ('<Filter name = "{0}" value = "{1}"/>'.format(
"list_filter", "1,X,MT") in xml_lines)
def test_test_query_biomart_37_no_xml():
"""Prepare a test xml document for the biomart service build 37 and query the service using it"""
## GIVEN defined biomart filters and attributes
build = '37'
filters = {'ensembl_gene_id': 'ENSG00000115091'}
attributes = ['hgnc_symbol', 'ensembl_transcript_id']
## WHEN querying ensembl
client = eracs.EnsemblBiomartClient(build='38',
filters=filters,
attributes=attributes,
header=False)
i = 0
for i, line in enumerate(client):
## THEN assert the correct gene is fetched
assert 'ACTR3' in line
## THEN assert there was a result
assert i > 0
def test_test_query_biomart_37_no_xml():
"""Prepare a test xml document for the biomart service build 37 and query the service using it"""
## GIVEN defined biomart filters and attributes
build = "37"
filters = {"ensembl_gene_id": "ENSG00000115091"}
attributes = ["hgnc_symbol", "ensembl_transcript_id"]
## WHEN querying ensembl
client = eracs.EnsemblBiomartClient(build="38",
filters=filters,
attributes=attributes,
header=False)
i = 0
for i, line in enumerate(client):
## THEN assert that either the correct gene is fetched or that an HTML page is returned (if the service is down)
assert ("ACTR3" in line or line == "<!doctype html>"
or "BioMart::Exception::Database" in line)
break
def test_query_biomart_38_no_xml():
"""Prepare a test xml document for the biomart service build 38 and
query the service using it
"""
# GIVEN defined biomart filters and attributes
filters = {"ensembl_gene_id": "ENSG00000115091"}
attributes = ["hgnc_symbol", "ensembl_transcript_id"]
query = '<?xml version="1.0" encoding="UTF-8"?>\
<!DOCTYPE Query>\
<Query virtualSchemaName = "default" formatter = "TSV" header = "0" uniqueRows = "0" count = "" \
datasetConfigVersion = "0.6" completionStamp = "1">\
\t<Dataset name = "hsapiens_gene_ensembl" interface = "default" >\
\t\t<Filter name = "ensembl_gene_id" value = "ENSG00000115091"/>\
\t\t<Attribute name = "hgnc_symbol" />\
\t\t<Attribute name = "ensembl_transcript_id" />\
\t</Dataset>\
</Query>'
# percent encode query before concatenating with url
url = "".join([ensembl_rest_clients.BIOMART_38, quote(query)])
response = (b"ACTR3\tENST00000263238\n"
b"ACTR3\tENST00000443297\n"
b"ACTR3\tENST00000415792\n"
b"ACTR3\tENST00000446821\n"
b"ACTR3\tENST00000535589\n"
b"ACTR3\tENST00000489779\n"
b"ACTR3\tENST00000484165\n"
b"ACTR3\tENST00000478928\n"
b"[success]")
responses.add(responses.GET, url, body=response, status=200)
# WHEN querying ensembl
client = ensembl_rest_clients.EnsemblBiomartClient(build="38",
filters=filters,
attributes=attributes,
header=False)
for line in client:
# THEN assert that either the correct gene is fetched or that an
assert "ACTR3" in line
def test_test_query_biomart_38_xml():
"""Prepare a test xml document for the biomart service build 38 and query the service using it"""
## GIVEN a xml file in biomart format
build = '38'
xml = """<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE Query>
<Query virtualSchemaName = "default" formatter = "TSV" header = "0" uniqueRows = "0" count = "" datasetConfigVersion = "0.6" completionStamp = "1">
<Dataset name = "hsapiens_gene_ensembl" interface = "default" >
<Filter name = "ensembl_gene_id" value = "ENSG00000115091"/>
<Attribute name = "hgnc_symbol" />
<Attribute name = "ensembl_transcript_id" />
</Dataset>
</Query>
"""
## WHEN querying ensembl
client = eracs.EnsemblBiomartClient(build='38', xml=xml, header=False)
## THEN assert that the result is correct
i = 0
for i, line in enumerate(client, 1):
assert 'ACTR3' in line
assert i > 0
def ensembl_biomart_client_37():
"""Return a client to the ensembl biomart, build 37"""
return ensembl_rest_clients.EnsemblBiomartClient("37")