def ex3_grand_position(source, destination, recall):
file_params = {
"agents": [10, 20, 30],
"jump_rate": [5, 10, 20],
# "source" : "/home/rob/dust/Projects/ABM_DA/experiments/ukf_experiments/ukf_results/agg_ukf_",
"source": source,
"destination": destination
}
"initialise plot for observed/unobserved agents"
g_plts = depickle.grand_plots(file_params,
True,
restrict=None,
observed=True)
if not recall:
"make dictionary"
L2 = g_plts.data_extractor()
"make pandas dataframe for seaborn"
error_frame = g_plts.data_framer(L2)
error_frame.to_pickle(
os.path.split(source)[:-1][0] + "/rjukf_locations_frame.pkl")
else:
error_frame = pd.read_pickle(
os.path.split(source)[:-1][0] + "/rjukf_locations_frame.pkl")
"make choropleth numpy array"
error_array = g_plts.choropleth_array(error_frame)
"make choropleth"
g_plts.choropleth_plot(error_array, "Numbers of Agents",
"Reversible Jump Window", "")
"make boxplot"
g_plts.boxplot(error_frame, "Reversible Jump Window",
"Grand Median L2s (log(metres))", "")
def ex1_grand_no_split(source, destination, recall):
""" Same as ex1_grand but doesn't split into observed/unobserved agents
Parameters
----------
source, destination : str
`source` of experiment pickles and `destination` of the resulting plots
recall : bool
`recall` previously aggregated pickles from a np/pandas frame for speed?
Returns
-------
None.
"""
prefix = "ukf*"
file_params = {
"agents": [10, 20, 30],
"prop": [0.25, 0.5, 0.75, int(1)],
# "source" : "/home/rob/dust/Projects/ABM_DA/experiments/ukf_experiments/ukf_results/agg_ukf_",
"source": source,
"destination": destination
}
"initialise plot for observed/unobserved agents"
g_plts = grand_plots(file_params, True, restrict=None, observed=True)
"make dictionary"
with HiddenPrints():
L2 = g_plts.data_extractor()
if not recall:
L2 = g_plts.data_extractor()
"make pandas dataframe for seaborn"
error_frame = g_plts.data_framer(L2)
"make choropleth numpy array"
error_array = g_plts.choropleth_array(error_frame)
error_frame.to_pickle(source + "Mixed_distance_error_array.pkl")
np.save(source + "Mixed_distance_error_array", error_array)
else:
error_array = np.load(source + "Mixed_distance_error_array.npy")
error_frame = pd.read_pickle(source + "Mixed_distance_error_array.pkl")
"make pandas dataframe for seaborn"
error_frame = g_plts.data_framer(L2)
"make choropleth numpy array"
error_array = g_plts.choropleth_array(error_frame)
"make choropleth"
g_plts.choropleth_plot(error_array, "Numbers of Agents",
"Proportion Observed", "Mixed")
"make boxplot"
g_plts.boxplot(error_frame, "Proportion Observed",
"Grand Median L2s", "Mixed")
def ex1_grand(source, destination, recall):
""" main function for parsing experiment 1 results into summary plots
Parameters
----------
source, destination : str
`source` of experiment pickles and `destination` of the resulting plots
recall : bool
`recall` previously aggregated pickles from a np/pandas frame for speed?
Returns
-------
None.
"""
prefix = "ukf*"
file_params = {
"agents": [10, 20, 30],
"prop": [0.25, 0.5, 0.75, int(1)],
# "source" : "/home/rob/dust/Projects/ABM_DA/experiments/ukf_experiments/ukf_results/agg_ukf_",
"source": source + prefix,
"destination": destination,
}
"plot observed/unobserved plots"
obs_bools = [True, False]
obs_titles = ["Observed", "Unobserved"]
for i in range(len(obs_bools)):
"initialise plot for observed/unobserved agents"
g_plts = grand_plots(file_params, True, restrict=ex1_restrict,
observed=obs_bools[i])
"make dictionary"
if not recall:
L2 = g_plts.data_extractor()
"make pandas dataframe for seaborn"
error_frame = g_plts.data_framer(L2)
"make choropleth numpy array"
error_array = g_plts.choropleth_array(error_frame)
np.save(source + obs_titles[i] + "_distance_error_array", error_array)
error_frame.to_pickle(source + obs_titles[i] + "_distance_error_array.pkl")
else:
error_frame = pd.read_pickle(source + obs_titles[i] + "_distance_error_array.pkl")
error_array = np.load(source + obs_titles[i] + "_distance_error_array.npy")
"make choropleth"
g_plts.choropleth_plot(error_array, "Numbers of Agents", "Proportion Observed",
obs_titles[i])
"make boxplot"
g_plts.boxplot(error_frame, "Proportion Observed", "Grand Median L2s",
obs_titles[i])
def ex2_grand(source, destination, recall):
""" main function for parsing experiment 2 results into summary plots
Parameters
----------
source, destination : str
`source` of experiment pickles and `destination` of the resulting plots
recall : bool
`recall` previously aggregated pickles from a np/pandas frame for speed?
Returns
-------
None.
"""
prefix = "agg*"
file_params = {
"agents": [10, 20],
"bin": [10, 25],
# "source" : "/home/rob/dust/Projects/ABM_DA/experiments/ukf_experiments/ukf_results/agg_ukf_",
"source": source,
"destination": destination,
}
"init plot class"
g_plts = grand_plots(file_params, True)
"make dictionary"
if not recall:
L2 = g_plts.data_extractor()
"make pandas dataframe for seaborn"
error_frame = g_plts.data_framer(L2)
"make choropleth numpy array"
error_array = g_plts.choropleth_array(error_frame)
error_frame.to_pickle(source + "Aggregate_distance_error_array.pkl")
np.save(source + "Aggregate_distance_error_array", error_array)
else:
error_array = np.load(source + "Aggregate_distance_error_array.npy")
error_frame = pd.read_pickle(source + "Aggregate_distance_error_array.pkl")
with HiddenPrints():
L2 = g_plts.data_extractor()
"make choropleth"
g_plts.choropleth_plot(error_array, "Numbers of Agents",
"Proportion Observed", "Aggregate")
"make boxplot"
g_plts.boxplot(error_frame, "Grid Square Size",
"Grand Median L2s", "Aggregate")
def ex3_grand_gates(source, destination, recall):
""" main function for parsing experiment 3 results into summary plots
Parameters
----------
source, destination : str
`source` of experiment pickles and `destination` of the resulting plots
recall : bool
`recall` previously aggregated pickles from a np/pandas frame for speed?
Returns
-------
None.
"""
file_params = {
"agents": [10, 20, 30],
"jump_rate": [5, 10, 20],
"source": source,
"destination": destination
}
"initialise plot for observed/unobserved agents"
g_plts = depickle.grand_plots(file_params, True)
"make dictionary"
if not recall:
gate_distances = g_plts.gates_extractor()
data_frame = g_plts.gates_data_frame(gate_distances)
data_frame.to_pickle(
os.path.split(source)[:-1][0] + "/rjukf_gates_frame.pkl")
else:
data_frame = pd.read_pickle(
os.path.split(source)[:-1][0] + "/rjukf_gates_frame.pkl")
"make pandas dataframe for seaborn"
"make choropleth numpy array"
g_plts.gates_data_lineplot(data_frame, "Time", "Error (Number of Gates)",
"rjmcmc_UKF_")