def get_distribution(distribution_string, state_name, valid_states):
"""
This high level function returns a state distribution.
Each valid state is expected to have exactly one weight.
The weights in the returned dictionary are normalized to sum to one.
@param distribution_string: a string of lines of text that each look something like 'A : 0.245'
@param state_name: something like 'amino acid'
@param valid_states: a set of valid states
@return: a dictionary mapping each state to a probability
"""
if not distribution_string:
raise HandlingError('no %s distribution was specified' % state_name)
state_to_weight = {}
for line in smallutil.stripped_lines(StringIO(distribution_string)):
state, weight = get_weight_pair(line, state_name, valid_states)
if state in state_to_weight:
raise HandlingError('duplicate %s: %s' % (state_name, state))
state_to_weight[state] = weight
if len(state_to_weight) < len(valid_states):
raise HandlingError('one or more %s was not assigned a weight' %
state_name)
total_weight = float(sum(state_to_weight.values()))
if not total_weight:
raise HandlingError('each %s weight is zero' % state_name)
for state in state_to_weight:
state_to_weight[state] /= total_weight
return state_to_weight
def multiline_state_to_ndarray(multiline_state):
arr = []
for line in stripped_lines(multiline_state.splitlines()):
row = []
for s in line:
v = int(s)
if v not in (0, 1):
raise ValueError('invalid allele')
row.append(v)
arr.append(row)
return np.array(arr)
def multiline_state_to_ndarray(multiline_state):
arr = []
for line in stripped_lines(multiline_state.splitlines()):
row = []
for s in line:
v = int(s)
if v not in (0, 1):
raise ValueError('invalid allele')
row.append(v)
arr.append(row)
return np.array(arr)