Created
February 20, 2021 06:41
-
-
Save keimina/78caef973a38356ff69b94d943937b9c to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
import pandas as pd | |
import matplotlib | |
matplotlib.use('Qt5Agg') | |
import matplotlib.pyplot as plt | |
data_points = np.array([[0,0],[0,1],[1,1],[1,0],[2,2],[2,3],[3,3],[3,2]]) | |
df = pd.DataFrame(data_points, columns=["x", "y"]) | |
fig, ax = plt.subplots(3, 8, figsize=(24,9), sharex=True, sharey=True) | |
df2 = df.copy() | |
df2 = df2.set_index(["x", "y"]) | |
df2.loc[:, "xy"] = df2.index | |
df2.loc[:, "key"] = 1 | |
df3 = pd.merge(df2, df2, on="key", suffixes=["_a", "_b"]) | |
del df3["key"] | |
df3.loc[:,"pairs"] = df3.apply(lambda s: s.tolist(), axis=1) | |
df3 = df3.pivot(index="xy_a", columns="xy_b", values="pairs") | |
df4 = df3.copy() | |
df4 = df4.applymap(lambda x: np.array(x[0]) - np.array(x[1])) | |
df5 = df4.applymap(lambda x: np.linalg.norm(x)) | |
df6 = df5.copy() | |
df6 = df6.applymap(lambda i: np.exp(-i**2)) | |
df6.values[np.diag_indices_from(df6.values)] = 0.0 | |
df6 = df6.apply(lambda s: s/s.sum(), axis=1) | |
# create low dimensional data point | |
np.random.seed(1) | |
data_points_2 = np.random.normal(size=(len(data_points), 1, 1)).tolist() | |
df7 = pd.DataFrame(data_points_2, columns=["xy"]) | |
df7 = df7.applymap(lambda i: tuple(i) + tuple([0])) | |
df7.set_index(["xy"]) | |
df7.loc[:, "key"] = 1 | |
df7 = pd.merge(df7, df7, on="key", suffixes=["_a", "_b"]) | |
del df7["key"] | |
df7.loc[:,"pairs"] = df7.apply(lambda s: s.tolist(), axis=1) | |
df7 = df7.pivot(index="xy_a", columns="xy_b", values="pairs") | |
def vec_diff(x): | |
return np.array(x[1]) - np.array(x[0]) | |
def annotate1(i, ax): | |
ax.annotate("", xy=i[1], xytext=i[0], arrowprops=dict(arrowstyle="->")) | |
ax.text(*i[1], "{:.02f}".format(np.linalg.norm(vec_diff(i)))) | |
def annotate2(s, ax, zero_index): | |
s2 = s.apply(lambda i: np.exp(-np.linalg.norm(vec_diff(i))**2)) | |
s2.iloc[zero_index] = 0.0 | |
s2 = s2/s2.sum() | |
s2 = s2.apply(lambda i: [i]) | |
s += s2 | |
s.apply(lambda i: ax.annotate("", xy=i[1], xytext=i[0], arrowprops=dict(arrowstyle="->"))) | |
s.apply(lambda i: ax.text(*i[1], "{:.02f}".format(i[2]))) | |
def annotate3(s, ax, zero_index): | |
s2 = s.apply(lambda i: np.exp(-np.linalg.norm(vec_diff(i))**2)) | |
s2.iloc[zero_index] = 0.0 | |
s2 = s2/s2.sum() | |
s2 = s2.apply(lambda i: [i]) | |
s += s2 | |
t = s.iloc[0] | |
s.apply(lambda i: ax.annotate("", xy=i[1], xytext=i[0], arrowprops=dict(arrowstyle="->", connectionstyle="arc3,rad=-0.8"))) | |
s.apply(lambda i: ax.text(*i[1], "{:.02f}".format(i[2]))) | |
for i in range(8): | |
df_tmp = df3.iloc[[i],:] | |
df_tmp.applymap(lambda j: annotate1(j, ax[0][i])) | |
for i in range(8): | |
df_tmp = df3.iloc[[i],:] | |
df_tmp.apply(lambda s: annotate2(s, ax[1][i], i), axis=1) | |
df_tmp = df7.iloc[[0],:] | |
df_tmp.apply(lambda s: annotate3(s, ax[2][0], 0), axis=1) | |
# df2 = df2.index.to_frame() | |
#df2.loc[:,"xy"] = df2.apply(lambda r: r.tolist(), axis=1) | |
# visualize table and dataframe | |
set_aspect = np.vectorize(lambda ax: ax.set_aspect("equal")) | |
set_xlim = np.vectorize(lambda ax: ax.set_xlim([0,4])) | |
set_ylim = np.vectorize(lambda ax: ax.set_ylim([0,4])) | |
set_xlim(ax) | |
set_ylim(ax) | |
set_aspect(ax) | |
fig.subplots_adjust(hspace=0.0, wspace=0.0) | |
fig.savefig("./fig.png") | |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment