class TestHash(unittest.TestCase):
def make_rand_ascii_hash(self, n):
"""
This helper function creates a random string of length 10 and
calculates its hash value using the h_ascii strategy.
"""
self.asciiDict = {chr(i): i for i in range(129)}
self.rand_string = ''.join(random.choices(string.ascii_letters, k=10))
self.sum = 0
for char in self.rand_string:
self.sum += self.asciiDict[char]
return self.rand_string, self.sum % n
def make_rand_mult_hash(self, n):
"""
This helper function creates a random string of length 1- and
calculates its hash value using the multiplicative hash strategy.
"""
self.A = (sqrt(5) - 1) / 2
self.asciiDict = {chr(i): i for i in range(129)}
self.rand_string = ''.join(random.choices(string.ascii_letters, k=10))
self.sum = 0
for char in self.rand_string:
self.sum += self.asciiDict[char]
return self.rand_string, floor(self.n * ((self.sum * self.A) % 1))
def make_rand_rolling_hash(self, n):
"""
This helper function creates a random string of length 10 and
calculates its hash value using the poly. rolling hash strategy.
"""
self.m = 2**64
self.p = 53
self.asciiDict = {chr(i): i for i in range(129)}
self.rand_string = ''.join(random.choices(string.ascii_letters, k=10))
self.sum = 0
for c, char in enumerate(self.rand_string):
self.sum += self.asciiDict[char] * self.p**c
return self.rand_string, (self.sum % self.m) % self.n
def test_ascii_hash_empty(self):
"""
This test checks h_ascii function on an empty string
"""
self.n = 10
self.assertEqual(0, h_ascii("", self.n))
def test_ascii_hash(self):
"""
This test checks h_ascii function on a random string of length 10
"""
self.n = 10
self.rand_string, self.hash_value = self.make_rand_ascii_hash(self.n)
self.assertEqual(self.hash_value, h_ascii(self.rand_string, self.n))
def test_asci_hash_no_table(self):
"""
This test checks h_ascii function on a random string of length 10
when n = 0
"""
self.n = 0
self.rand_string = ''.join(random.choices(string.ascii_letters, k=10))
self.assertRaises(ZeroDivisionError and SystemExit, h_ascii,
self.rand_string, self.n)
def test_mult_hash_empty(self):
"""
This test checks h_ascii function on an empty string
"""
self.n = 10
self.assertEqual(0, h_mult("", self.n))
def test_mult_hash(self):
"""
This test checks h_ascii function on a random string of length 10
"""
self.n = 10
self.rand_string, self.hash_value = self.make_rand_mult_hash(self.n)
self.assertEqual(self.hash_value, h_mult(self.rand_string, self.n))
def test_rolling_hash_empty(self):
"""
This test checks h_rolling function on an empty string
"""
self.n = 10
self.assertEqual(0, h_rolling("", self.n))
def test_rolling_hash(self):
"""
This test checks h_rolling function on a random string of length 10
"""
self.n = 10
self.rand_string, self.hash_value = self.make_rand_rolling_hash(self.n)
self.assertEqual(self.hash_value, h_rolling(self.rand_string, self.n))
def test_rolling_hash_no_table(self):
"""
This test checks h_rolling function on a random string of length 10
when n = 0
"""
self.n = 0
self.rand_string = ''.join(random.choices(string.ascii_letters, k=10))
self.assertRaises(ZeroDivisionError and SystemExit, h_rolling,
self.rand_string, self.n)
def test_hash_lp_empty(self):
"""
This test checks LinearProbe add function on a empty table
"""
self.n = 10
self.value = "Hello"
self.inst_table = LinearProbe(self.n, h_ascii)
self.key, self.hash_value = self.make_rand_ascii_hash(self.n)
self.assertTrue(self.inst_table.add(self.key, self.value))
def test_hash_lp_full(self):
"""
This test checks LinearProbe add function on a full table
"""
self.n = 10
self.value = "Hello"
self.key, self.hash_value = self.make_rand_ascii_hash(self.n)
self.inst_table = LinearProbe(self.n, h_ascii)
self.inst_table.table = ["Full" for i in range(self.n)]
self.assertFalse(self.inst_table.add(self.key, self.value))
def test_hash_lp_search_empty(self):
"""
This test checks LinearProbe search function on empty table
where the key does not exist
"""
self.n = 10
self.key, self.hash_value = self.make_rand_ascii_hash(self.n)
self.inst_table = LinearProbe(self.n, h_ascii)
self.assertEqual(None, self.inst_table.search(self.key))
def test_hash_lp_search_full(self):
"""
This test checks LinearProbe search function on a full table
where the key exists
"""
self.n = 10
self.inst_table = LinearProbe(self.n, h_ascii)
self.inst_table.table = [(chr(x), x) for x in range(self.n)]
self.assertTrue(1, self.inst_table.search(chr(1)))
def test_hash_lp_search_partial(self):
"""
This test checks LinearProbe search function on a partially full table
where the key does not exist
"""
self.n = 10
self.inst_table = LinearProbe(self.n, h_ascii)
self.inst_table.table = [(chr(x), x) for x in range(self.n)]
self.inst_table.table[5] = (chr(15), 15)
self.inst_table.table[6] = None
self.assertEqual(None, self.inst_table.search(chr(5)))
def test_hash_chain_empty(self):
"""
This test checks ChainedHash add function on a empty table
"""
self.n = 10
self.value = "Hello"
self.inst_table = ChainedHash(self.n, h_ascii)
self.key, self.hash_value = self.make_rand_ascii_hash(self.n)
self.assertTrue(self.inst_table.add(self.key, self.value))
def test_hash_chain_full(self):
"""
This test checks ChainedHash add function on a table with
one entry in each linked list
"""
self.n = 10
self.value = "Hello"
self.inst_table = ChainedHash(self.n, h_ascii)
self.inst_table.table = [[(chr(x), x)] for x in range(self.n)]
self.key, self.hash_value = self.make_rand_ascii_hash(self.n)
self.assertTrue(self.inst_table.add(self.key, self.value))
def test_hash_chain_search_full(self):
"""
This test checks LinearProbe search function on a full table
with multiple elementes in each linked list
"""
self.n = 10
self.inst_table = ChainedHash(self.n, h_ascii)
self.inst_table.table = [[(chr(x), x), (chr(x + self.n), x + self.n)]
for x in range(self.n)]
self.assertEqual(5, self.inst_table.search(chr(5))[0])
def test_hash_chain_search_empty(self):
"""
This test checks LinearProbe search function on an empty table
where the key exists
"""
self.n = 10
self.inst_table = ChainedHash(self.n, h_ascii)
self.assertEqual(None, self.inst_table.search(chr(5)))
def test_boxplot_exist(self):
"""
This test checks to see if file exists after scatterplot is created
"""
self.x_data = [x for x in range(10)]
self.y_data = self.x_data
self.x_label = "X"
self.y_label = "Y"
self.title = "Title"
self.outfile = "test.png"
scatter_plot(self.x_data, self.y_data, self.x_label, self.y_label,
self.title, self.outfile)
self.assertEqual(True, os.path.exists(self.outfile))
def test_keys_exist(self):
"""
This test checks to see if keys are tracked in the keys list
"""
self.n = 10
self.key = "TEST"
self.value = 1
self.inst_table = ChainedHash(self.n, h_ascii)
self.inst_table.add(self.key, self.value)
self.assertEqual(self.key, self.inst_table.keys[0])
def test_keys_not_exist(self):
"""
This test checks to see if keys list is empty when initialized
"""
self.n = 10
self.inst_table = ChainedHash(self.n, h_ascii)
self.assertEqual([], self.inst_table.keys)
def testChainedHashSearchChain(self):
test_obj = ChainedHash(10, hf.h_ascii)
test_obj.add("a", 1)
test_obj.add("k", "return this, not 1")
self.assertEqual("return this, not 1", test_obj.search("k"))
def hash_process(gene_reads, sample_attributes, gene, group_types,
output_file):
"""
This function calculates the gene expression distribution across either
tissue groups (SMTS) or tissue type (SMTSD) for a target gene. It uses
hash tables for O(1) lookups. A series of box plots is generated.
Parameters:
- gene_reads: (see next line)
GTEx_Analysis_2017-06-05_v8_RNASeQCv1.1.9_gene_reads.acmg_59.gct.gz
- sample_attributes: GTEx_Analysis_v8_Annotations_SampleAttributesDS.txt
- gene: The target gene
- group_types: Tissue group or type
- output_file: File for saving the box plot (.png)
"""
sample_id_col_name = 'SAMPID'
sample_info_header = None
# Initialize tissue-sample hash table
h_table_samples = ChainedHash(1000, h_rolling)
# Read and preprocess the data lines
for line in open(sample_attributes):
if sample_info_header is None:
sample_info_header = line.rstrip().split('\t')
# Find the proper columns containing the group types and sample id's
# using linear search
group_col_idx = linear_search(group_types, sample_info_header)
sample_id_col_idx = linear_search(sample_id_col_name,
sample_info_header)
# Add samples to hash table
else:
line = line.rstrip().split('\t')
group = line[group_col_idx]
sample = line[sample_id_col_idx]
h_table_samples.add(key=group, value=sample)
version = None
dim = None
data_header = None
gene_name_col = 1
# Initialize sample-count hash table
h_table_counts = ChainedHash(100000, h_rolling)
# Read and preprocess the gene reads data lines
for line in gzip.open(gene_reads, 'rt'):
if version is None:
version = line
continue
if dim is None:
dim = [int(x) for x in line.rstrip().split()]
continue
if data_header is None:
data_header = line.rstrip().split('\t')
continue
line = line.rstrip().split('\t')
if line[gene_name_col] == gene:
gene_row = line
for sample, count in zip(data_header, gene_row):
h_table_counts.add(sample, count)
# Get the counts for each sample of each tissue type
group_counts = []
for tissue in h_table_samples.keys:
counts = []
for sample in h_table_samples.search(tissue):
count = h_table_counts.search(sample)
if count is not None:
counts.append(int(count[0]))
group_counts.append(counts)
# Generate box plot
boxplot(group_counts, h_table_samples.keys, gene, group_types,
"Gene read counts", output_file)