summaryrefslogtreecommitdiff
path: root/lib
diff options
context:
space:
mode:
authorDaniel Friesel <derf@finalrewind.org>2018-03-28 16:50:09 +0200
committerDaniel Friesel <derf@finalrewind.org>2018-03-28 16:50:09 +0200
commit4b3ff326b8f97b5f849b8619d984d2e2d50ab9b9 (patch)
tree84f9f944f328b5f32f8398563f85b6497d3b74e2 /lib
parentf5690d1ca3042fcbbe92ebc7f4fcc2785e6cfee3 (diff)
export plot data for pgfplots (experimental and ugly)
Diffstat (limited to 'lib')
-rwxr-xr-xlib/dfatool.py5
-rwxr-xr-xlib/plotter.py65
2 files changed, 57 insertions, 13 deletions
diff --git a/lib/dfatool.py b/lib/dfatool.py
index a591be3..17e7014 100755
--- a/lib/dfatool.py
+++ b/lib/dfatool.py
@@ -984,6 +984,11 @@ class EnergyModel:
return self._parameter_names.index(param_name)
return len(self._parameter_names) + int(param_name)
+ def param_name(self, param_index):
+ if param_index < len(self._parameter_names):
+ return self._parameter_names[param_index]
+ return str(param_index)
+
def get_fitted(self):
if 'fitted_model_getter' in self.cache and 'fitted_info_getter' in self.cache:
diff --git a/lib/plotter.py b/lib/plotter.py
index 6ee2691..2ec635f 100755
--- a/lib/plotter.py
+++ b/lib/plotter.py
@@ -3,6 +3,7 @@
import itertools
import numpy as np
import matplotlib.pyplot as plt
+import re
from matplotlib.patches import Polygon
def float_or_nan(n):
@@ -97,7 +98,7 @@ def plot_xy(X, Y, xlabel = None, ylabel = None, title = None):
def _param_slice_eq(a, b, index):
return (*a[1][:index], *a[1][index+1:]) == (*b[1][:index], *b[1][index+1:]) and a[0] == b[0]
-def plot_param(model, state_or_trans, attribute, param_idx, xlabel = None, ylabel = None, title = None, extra_functions = []):
+def plot_param(model, state_or_trans, attribute, param_idx, xlabel = None, ylabel = None, title = None, extra_function = None):
fig, ax1 = plt.subplots(figsize=(10,6))
if title != None:
fig.canvas.set_window_title(title)
@@ -107,10 +108,19 @@ def plot_param(model, state_or_trans, attribute, param_idx, xlabel = None, ylabe
ax1.set_ylabel(ylabel)
plt.subplots_adjust(left = 0.05, bottom = 0.05, right = 0.99, top = 0.99)
+ param_name = model.param_name(param_idx)
+
+ function_filename = 'plot_param_{}_{}_{}.txt'.format(state_or_trans, attribute, param_name)
+ data_filename_base = 'measurements_{}_{}_{}'.format(state_or_trans, attribute, param_name)
+
param_model, param_info = model.get_fitted()
by_other_param = {}
+ XX = []
+
+ legend_sanitizer = re.compile(r'[^0-9a-zA-Z]+')
+
for k, v in model.by_param.items():
if k[0] == state_or_trans:
other_param_key = (*k[1][:param_idx], *k[1][param_idx+1:])
@@ -118,30 +128,59 @@ def plot_param(model, state_or_trans, attribute, param_idx, xlabel = None, ylabe
by_other_param[other_param_key] = {'X': [], 'Y': []}
by_other_param[other_param_key]['X'].extend([float(k[1][param_idx])] * len(v[attribute]))
by_other_param[other_param_key]['Y'].extend(v[attribute])
+ XX.extend(by_other_param[other_param_key]['X'])
+
+ XX = np.array(XX)
+ x_range = int((XX.max() - XX.min()) * 10)
+ xsp = np.linspace(XX.min(), XX.max(), x_range)
+ YY = [xsp]
+ YY_legend = [param_name]
+ YY2 = []
+ YY2_legend = []
cm = plt.get_cmap('brg', len(by_other_param))
- for i, k in enumerate(by_other_param):
+ for i, k in sorted(enumerate(by_other_param), key = lambda x: x[1]):
v = by_other_param[k]
v['X'] = np.array(v['X'])
v['Y'] = np.array(v['Y'])
plt.plot(v['X'], v['Y'], "rx", color=cm(i))
- x_range = int((v['X'].max() - v['X'].min()) * 2)
- xsp = np.linspace(v['X'].min(), v['X'].max(), x_range)
+ YY2_legend.append(legend_sanitizer.sub('_', 'X_{}'.format(k)))
+ YY2.append(v['X'])
+ YY2_legend.append(legend_sanitizer.sub('_', 'Y_{}'.format(k)))
+ YY2.append(v['Y'])
+
+ sanitized_k = legend_sanitizer.sub('_', str(k))
+ with open('{}_{}.txt'.format(data_filename_base, sanitized_k), 'w') as f:
+ print('X Y', file=f)
+ for i in range(len(v['X'])):
+ print('{} {}'.format(v['X'][i], v['Y'][i]), file=f)
+
+ #x_range = int((v['X'].max() - v['X'].min()) * 10)
+ #xsp = np.linspace(v['X'].min(), v['X'].max(), x_range)
if param_model:
ysp = []
for x in xsp:
xarg = [*k[:param_idx], x, *k[param_idx:]]
ysp.append(param_model(state_or_trans, attribute, param = xarg))
plt.plot(xsp, ysp, "r-", color=cm(i), linewidth=0.5)
- if len(extra_functions) != 0:
- for f in extra_functions:
- ysp = []
- with np.errstate(divide='ignore', invalid='ignore'):
- for x in xsp:
- xarg = [*k[:param_idx], x, *k[param_idx:]]
- ysp.append(f(*xarg))
- plt.plot(xsp, ysp, "r--", color=cm(i), linewidth=1, dashes=(3, 3))
-
+ YY.append(ysp)
+ YY_legend.append(legend_sanitizer.sub('_', 'regr_{}'.format(k)))
+ if extra_function != None:
+ ysp = []
+ with np.errstate(divide='ignore', invalid='ignore'):
+ for x in xsp:
+ xarg = [*k[:param_idx], x, *k[param_idx:]]
+ ysp.append(extra_function(*xarg))
+ plt.plot(xsp, ysp, "r--", color=cm(i), linewidth=1, dashes=(3, 3))
+ YY.append(ysp)
+ YY_legend.append(legend_sanitizer.sub('_', 'symb_{}'.format(k)))
+
+ with open(function_filename, 'w') as f:
+ print(' '.join(YY_legend), file=f)
+ for elem in np.array(YY).T:
+ print(' '.join(map(str, elem)), file=f)
+
+ print(data_filename_base, function_filename)
plt.show()