def simpleCohort():
delete('Cohort', 'TestCohort', force=True)
# Create the simple cohort
a = Accession('Sample1', files=['file1.txt', 'file2.txt'], type='WGS')
b = Accession('Sample2', files=['file1.txt', 'file2.txt'], type='WGS')
c = Accession('Sample3', files=['file1.txt', 'file2.txt'], type='CHIP')
d = Accession('Sample4', files=['file1.txt', 'file2.txt'], type='CHIP')
x = Cohort('TestCohort')
for acc in [a, b, c, d]:
x.add_accession(acc)
return x
def simpleCohort():
delete("Cohort", "TestCohort", force=True)
# Create the simple cohort
a = Accession("Sample1", files=["file1.txt", "file2.txt"], type="WGS")
b = Accession("Sample2", files=["file1.txt", "file2.txt"], type="WGS")
c = Accession("Sample3", files=["file1.txt", "file2.txt"], type="CHIP")
d = Accession("Sample4", files=["file1.txt", "file2.txt"], type="CHIP")
x = Cohort("TestCohort")
for acc in [a, b, c, d]:
x.add_accession(acc)
return x
def simpleCohort():
FreezableAPI.delete("Cohort", "TestCohort")
# Create the simple cohort
a = Accession("Sample1", files=["file1.txt", "file2.txt"], type="WGS")
b = Accession("Sample2", files=["file1.txt", "file2.txt"], type="WGS")
c = Accession("Sample3", files=["file1.txt", "file2.txt"], type="CHIP")
d = Accession("Sample4", files=["file1.txt", "file2.txt"], type="CHIP")
x = Cohort("TestCohort")
for acc in [a, b, c, d]:
x.add_accession(acc)
yield x
FreezableAPI.delete(x.m80.dtype, x.m80.name)
def test_add_accession(simpleCohort):
a = Accession('Sample4', files=['file1.txt', 'file2.txt'], type='CHIP')
if a in simpleCohort:
del simpleCohort[a]
start_len = len(simpleCohort)
simpleCohort.add_accession(a)
assert len(simpleCohort) == start_len + 1
def test_add_accession(simpleCohort):
a = Accession("Sample4", files=["file1.txt", "file2.txt"], type="CHIP")
if a in simpleCohort:
del simpleCohort[a]
start_len = len(simpleCohort)
simpleCohort.add_accession(a)
assert len(simpleCohort) == start_len + 1
def test_delitem(simpleCohort):
a = Accession("TESTSAMPLE_IGNORE", files=["file1.txt", "file2.txt"], type="CHIP")
if a not in simpleCohort:
simpleCohort.add_accession(a)
start_len = len(simpleCohort)
del simpleCohort["TESTSAMPLE_IGNORE"]
assert len(simpleCohort) == start_len - 1
def __getitem__(self, name):
'''
Get an accession from the database the pythonic way.
Paremeters
----------
name : object
Can be a string, i.e. the name or alias of an Accession,
it can be an Actual Accession OR the AID which
is an internal ID for accession
'''
AID = self._get_AID(name)
cur = self._db.cursor()
# Get the name based on AID
name, = cur.execute('SELECT name FROM accessions WHERE AID = ?',
(AID, )).fetchone()
metadata = {
k: v
for k, v in cur.execute(
'''
SELECT key, val FROM metadata WHERE AID = ?;
''', (AID, )).fetchall()
}
metadata['AID'] = AID
files = [
x[0] for x in cur.execute(
'''
SELECT url FROM files WHERE AID = ?;
''', (AID, )).fetchall()
]
return Accession(name, files=files, **metadata)
def test_delitem(simpleCohort):
a = Accession('TESTSAMPLE_IGNORE',
files=['file1.txt', 'file2.txt'],
type='CHIP')
if a not in simpleCohort:
simpleCohort.add_accession(a)
start_len = len(simpleCohort)
del simpleCohort['TESTSAMPLE_IGNORE']
assert len(simpleCohort) == start_len - 1
def RNAAccession2():
a = Accession('RNAAccession2',
files=[
'./data/Sample2_ATGTCA_L005_R1_001.fastq',
'./data/Sample2_ATGTCA_L005_R2_001.fastq',
'./data/Sample2_ATGTCA_L006_R1_001.fastq',
'./data/Sample2_ATGTCA_L006_R2_001.fastq',
],
type='RNASeq')
return a
def RNAAccession1():
a = Accession('RNAAccession1',
files=[
'./data/Sample1_ATGTCA_L007_R1_001.fastq',
'./data/Sample1_ATGTCA_L007_R2_001.fastq',
'./data/Sample1_ATGTCA_L008_R1_001.fastq',
'./data/Sample1_ATGTCA_L008_R2_001.fastq',
],
type='RNASeq')
return a
def RNAAccession2():
a = Accession(
"RNAAccession2",
files=[
"./data/Sample2_ATGTCA_L005_R1_001.fastq",
"./data/Sample2_ATGTCA_L005_R2_001.fastq",
"./data/Sample2_ATGTCA_L006_R1_001.fastq",
"./data/Sample2_ATGTCA_L006_R2_001.fastq",
],
type="RNASeq",
)
return a
def RNAAccession1():
a = Accession(
"RNAAccession1",
files=[
"./data/Sample1_ATGTCA_L007_R1_001.fastq",
"./data/Sample1_ATGTCA_L007_R2_001.fastq",
"./data/Sample1_ATGTCA_L008_R1_001.fastq",
"./data/Sample1_ATGTCA_L008_R2_001.fastq",
],
type="RNASeq",
)
return a
def add_accessions_from_data_frame(self, df, name_col):
'''
Add accessions from data frame. This assumes
each row is an Accession and that the properties
of the accession are stored in the columns.
Parameters
----------
df : pandas.DataFrame
The pandas data frame containing one accession
per row
name_col : string
The column containing the accession names
Example
-------
>>> df = pd.DataFrame(
[['S1' 23 'O'],
['S2' 30 'O+']],
columns = ['Name','Age','Type']
)
>>> x = m80.add_accessions_from_data_frame(df,'Name')
Would yield two Accessions: S1 and S2 with Age and Type
properties.
'''
if name_col not in df.columns:
raise ValueError(f'{name_col}S not a valid column name')
# filter out rows with NaN name_col values
# The tilda operator is a boolean inversion
df = df.loc[~df[name_col].isnull(), :]
accessions = []
# Iterate over the rows and create and accessions from each one
for i, row in df.iterrows():
d = dict(row)
name = d[name_col]
del d[name_col]
# Get rid of missing data
for k, v in list(d.items()):
if isinstance(v, numbers.Number) and math.isnan(v):
del d[k]
else:
d[k] = str(v)
accessions.append(Accession(name, files=None, **d))
self.add_accessions(accessions)
def test_add_files():
x = Accession("empty")
x.add_files(["./test.txt", "test2.txt", "test3.txt"])
def test_add_relative_path():
x = Accession("empty")
x.add_file("./test.txt")
def simpleAccession():
# Create a simple Accession
return Accession("Sample1",
files=["file1.txt", "file2.txt"],
type="sample")
def test_from_accessions():
a = Accession("Sample1", files=["file1.txt", "file2.txt"], type="WGS")
b = Accession("Sample2", files=["file1.txt", "file2.txt"], type="WGS")
c = Accession("Sample3", files=["file1.txt", "file2.txt"], type="CHIP")
d = Accession("Sample4", files=["file1.txt", "file2.txt"], type="CHIP")
Cohort.from_accessions("TestCohort", [a, b, c, d])
def test_add_files():
x = Accession('empty')
x.add_files(['./test.txt', 'test2.txt', 'test3.txt'])
def test_bare_accession():
x = Accession("empty")
assert isinstance(x, Accession)
str(x)
repr(x)
def test_from_accessions():
a = Accession('Sample1', files=['file1.txt', 'file2.txt'], type='WGS')
b = Accession('Sample2', files=['file1.txt', 'file2.txt'], type='WGS')
c = Accession('Sample3', files=['file1.txt', 'file2.txt'], type='CHIP')
d = Accession('Sample4', files=['file1.txt', 'file2.txt'], type='CHIP')
x = Cohort.from_accessions('TestCohort', [a, b, c, d])
def simpleAccession():
# Create a simple Accession
return Accession('Sample1',
files=['file1.txt', 'file2.txt'],
type='sample')
def test_load_from_yaml():
Accession.from_yaml("data/test_accession.yaml")
def test_add_relative_path():
x = Accession('empty')
x.add_file('./test.txt')
