Spaces:
Sleeping
Sleeping
refrac
Browse files- __pycache__/data_processor.cpython-311.pyc +0 -0
- __pycache__/hypergraph_drawer.cpython-311.pyc +0 -0
- __pycache__/visualizer.cpython-311.pyc +0 -0
- app.py +65 -132
- data_processor.py +25 -0
- hypergraph_drawer.py +20 -0
- visualizer.py +43 -0
__pycache__/data_processor.cpython-311.pyc
ADDED
Binary file (1.8 kB). View file
|
|
__pycache__/hypergraph_drawer.cpython-311.pyc
ADDED
Binary file (1.17 kB). View file
|
|
__pycache__/visualizer.cpython-311.pyc
ADDED
Binary file (1.92 kB). View file
|
|
app.py
CHANGED
@@ -1,136 +1,69 @@
|
|
1 |
import streamlit as st
|
2 |
-
import
|
3 |
-
import
|
4 |
-
|
5 |
-
import hypernetx as hnx
|
6 |
-
import matplotlib.pyplot as plt
|
7 |
-
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
|
8 |
-
from io import BytesIO
|
9 |
-
import time
|
10 |
|
11 |
-
#
|
12 |
-
df = pd.read_csv("gmm_point_tracking_with_centroids.csv")
|
13 |
st.set_page_config(layout="wide")
|
14 |
|
15 |
-
#
|
16 |
-
|
17 |
-
|
18 |
-
|
19 |
-
|
20 |
-
|
21 |
-
|
22 |
-
|
23 |
-
|
24 |
-
|
25 |
-
|
26 |
-
|
27 |
-
|
28 |
-
|
29 |
-
|
30 |
-
|
31 |
-
|
32 |
-
|
33 |
-
|
34 |
-
|
35 |
-
st.session_state
|
36 |
-
|
37 |
-
|
38 |
-
st.session_state.
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
|
57 |
-
|
58 |
-
|
59 |
-
|
60 |
-
|
61 |
-
|
62 |
-
|
63 |
-
|
64 |
-
|
65 |
-
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
71 |
-
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
canvas = FigureCanvas(fig_hnx)
|
77 |
-
buffer = BytesIO()
|
78 |
-
canvas.print_png(buffer)
|
79 |
-
buffer.seek(0)
|
80 |
-
|
81 |
-
# 用 Plotly 可视化高斯混合分布
|
82 |
-
fig_gmm = px.scatter(
|
83 |
-
sampled_df,
|
84 |
-
x="x",
|
85 |
-
y="y",
|
86 |
-
color="cluster",
|
87 |
-
hover_data=["title", "keywords", "rating_avg", "confidence_avg", "author", "site"],
|
88 |
-
title=f"高斯混合分布聚类(迭代 {iteration})",
|
89 |
-
)
|
90 |
-
|
91 |
-
# 添加聚类中心点
|
92 |
-
for cluster in sampled_df["cluster"].unique():
|
93 |
-
centroid_x = sampled_df[sampled_df["cluster"] == cluster]["centroid_x"].iloc[0]
|
94 |
-
centroid_y = sampled_df[sampled_df["cluster"] == cluster]["centroid_y"].iloc[0]
|
95 |
-
fig_gmm.add_scatter(
|
96 |
-
x=[centroid_x],
|
97 |
-
y=[centroid_y],
|
98 |
-
mode="markers",
|
99 |
-
marker=dict(size=15, color="black", symbol="x"),
|
100 |
-
name=f"Cluster {cluster} Center",
|
101 |
-
)
|
102 |
-
|
103 |
-
# 并排展示超图和高斯混合分布
|
104 |
-
col1, col2 = st.columns(2)
|
105 |
-
col1.header("超图可视化")
|
106 |
-
col1.image(buffer, caption="超图可视化", use_column_width=True)
|
107 |
-
|
108 |
-
col2.header("高斯混合分布聚类结果")
|
109 |
-
col2.plotly_chart(fig_gmm, use_container_width=True)
|
110 |
-
|
111 |
-
# 显示采样论文的详细信息
|
112 |
-
st.header("采样论文详细信息")
|
113 |
-
st.dataframe(sampled_df[["title", "keywords", "rating_avg", "confidence_avg", "site"]])
|
114 |
-
|
115 |
-
# 增加第二种可视化方式
|
116 |
-
st.header("论文评分分布")
|
117 |
-
|
118 |
-
# 创建柱状图
|
119 |
-
fig_bar = px.bar(
|
120 |
-
sampled_df,
|
121 |
-
x="title",
|
122 |
-
y="rating_avg",
|
123 |
-
color="cluster",
|
124 |
-
title="论文评分分布",
|
125 |
-
hover_data=["keywords", "confidence_avg", "author"],
|
126 |
-
)
|
127 |
-
|
128 |
-
# 调整布局
|
129 |
-
fig_bar.update_layout(
|
130 |
-
xaxis_title="论文标题",
|
131 |
-
yaxis_title="平均评分",
|
132 |
-
xaxis_tickangle=-45,
|
133 |
-
)
|
134 |
-
|
135 |
-
# 显示柱状图
|
136 |
-
st.plotly_chart(fig_bar, use_container_width=True)
|
|
|
1 |
import streamlit as st
|
2 |
+
from data_processor import load_data, process_data
|
3 |
+
from visualizer import visualize_gmm, visualize_ratings
|
4 |
+
from hypergraph_drawer import draw_hypergraph
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
+
# 设置页面配置
|
|
|
7 |
st.set_page_config(layout="wide")
|
8 |
|
9 |
+
# 主应用
|
10 |
+
def main():
|
11 |
+
st.title("高斯混合分布聚类可视化")
|
12 |
+
|
13 |
+
# 使用 sidebar 控制参数
|
14 |
+
with st.sidebar:
|
15 |
+
st.header("控制面板")
|
16 |
+
autoplay = st.button("自动播放")
|
17 |
+
if autoplay:
|
18 |
+
for i in range(1, 11):
|
19 |
+
with st.spinner(f"迭代 {i}"):
|
20 |
+
time.sleep(1)
|
21 |
+
st.session_state.iteration = i
|
22 |
+
st.session_state.autoplay = False
|
23 |
+
st.experimental_rerun()
|
24 |
+
|
25 |
+
# 主页面布局
|
26 |
+
if 'autoplay' not in st.session_state:
|
27 |
+
st.session_state.autoplay = True
|
28 |
+
|
29 |
+
if 'iteration' not in st.session_state:
|
30 |
+
st.session_state.iteration = 1
|
31 |
+
|
32 |
+
if st.session_state.autoplay:
|
33 |
+
# 隐藏迭代次数滑条
|
34 |
+
iteration = st.session_state.iteration
|
35 |
+
else:
|
36 |
+
# 显示迭代次数滑条
|
37 |
+
iteration = st.slider("选择迭代次数", min_value=1, max_value=10, value=st.session_state.iteration, step=1)
|
38 |
+
|
39 |
+
# 动态限制采样数量的最大值
|
40 |
+
df = load_data()
|
41 |
+
max_samples = len(df)
|
42 |
+
num_samples = st.slider("选择采样论文数量", min_value=1, max_value=min(100, max_samples), value=min(10, max_samples), step=1)
|
43 |
+
|
44 |
+
# 处理数据
|
45 |
+
sampled_df, probabilities, hyperedges = process_data(df, iteration, num_samples)
|
46 |
+
|
47 |
+
# 并排展示超图和高斯混合分布
|
48 |
+
col1, col2 = st.columns(2)
|
49 |
+
with col1:
|
50 |
+
st.header("超图可视化")
|
51 |
+
hypergraph_image = draw_hypergraph(hyperedges)
|
52 |
+
st.image(hypergraph_image, caption="超图可视化", use_container_width=True)
|
53 |
+
|
54 |
+
with col2:
|
55 |
+
st.header("高斯混合分布聚类结果")
|
56 |
+
fig_gmm = visualize_gmm(sampled_df, iteration)
|
57 |
+
st.plotly_chart(fig_gmm, use_container_width=True)
|
58 |
+
|
59 |
+
# 显示采样论文的详细信息
|
60 |
+
st.header("采样论文详细信息")
|
61 |
+
st.dataframe(sampled_df[["title", "keywords", "rating_avg", "confidence_avg", "site"]])
|
62 |
+
|
63 |
+
# 增加第二种可视化方式
|
64 |
+
st.header("论文评分分布")
|
65 |
+
fig_bar = visualize_ratings(sampled_df)
|
66 |
+
st.plotly_chart(fig_bar, use_container_width=True)
|
67 |
+
|
68 |
+
if __name__ == "__main__":
|
69 |
+
main()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
data_processor.py
ADDED
@@ -0,0 +1,25 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import pandas as pd
|
2 |
+
import numpy as np
|
3 |
+
|
4 |
+
def load_data():
|
5 |
+
return pd.read_csv("gmm_point_tracking_with_centroids.csv")
|
6 |
+
|
7 |
+
def process_data(df, iteration, num_samples):
|
8 |
+
# 随机采样论文
|
9 |
+
sampled_df = df.sample(n=num_samples, random_state=iteration)
|
10 |
+
|
11 |
+
# 计算每个论文属于各个 cluster 的概率
|
12 |
+
probabilities = []
|
13 |
+
for idx, row in sampled_df.iterrows():
|
14 |
+
prob_str = row["probabilities"].strip("[]")
|
15 |
+
prob_list = list(map(float, prob_str.split(", ")))
|
16 |
+
probabilities.append(prob_list)
|
17 |
+
|
18 |
+
# 找到每个论文概率最高的 3 个 cluster
|
19 |
+
k = 3
|
20 |
+
hyperedges = {}
|
21 |
+
for idx, prob in enumerate(probabilities):
|
22 |
+
top_k = np.argsort(prob)[-k:][::-1]
|
23 |
+
hyperedges[idx] = [f"Cluster {c}" for c in top_k]
|
24 |
+
|
25 |
+
return sampled_df, probabilities, hyperedges
|
hypergraph_drawer.py
ADDED
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import hypernetx as hnx
|
2 |
+
import matplotlib.pyplot as plt
|
3 |
+
from matplotlib.backends.backend_agg import FigureCanvasAgg as FigureCanvas
|
4 |
+
from io import BytesIO
|
5 |
+
|
6 |
+
def draw_hypergraph(hyperedges):
|
7 |
+
# 构建超图
|
8 |
+
H = hnx.Hypergraph(hyperedges)
|
9 |
+
|
10 |
+
# 绘制超图
|
11 |
+
fig, ax = plt.subplots(figsize=(12, 8))
|
12 |
+
hnx.draw(H, ax=ax)
|
13 |
+
|
14 |
+
# 将超图保存为图像
|
15 |
+
canvas = FigureCanvas(fig)
|
16 |
+
buffer = BytesIO()
|
17 |
+
canvas.print_png(buffer)
|
18 |
+
buffer.seek(0)
|
19 |
+
|
20 |
+
return buffer
|
visualizer.py
ADDED
@@ -0,0 +1,43 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import plotly.express as px
|
2 |
+
|
3 |
+
def visualize_gmm(sampled_df, iteration):
|
4 |
+
fig = px.scatter(
|
5 |
+
sampled_df,
|
6 |
+
x="x",
|
7 |
+
y="y",
|
8 |
+
color="cluster",
|
9 |
+
hover_data=["title", "keywords", "rating_avg", "confidence_avg", "author", "site"],
|
10 |
+
title=f"高斯混合分布聚类(迭代 {iteration})",
|
11 |
+
)
|
12 |
+
|
13 |
+
# 添加聚类中心点
|
14 |
+
for cluster in sampled_df["cluster"].unique():
|
15 |
+
centroid_x = sampled_df[sampled_df["cluster"] == cluster]["centroid_x"].iloc[0]
|
16 |
+
centroid_y = sampled_df[sampled_df["cluster"] == cluster]["centroid_y"].iloc[0]
|
17 |
+
fig.add_scatter(
|
18 |
+
x=[centroid_x],
|
19 |
+
y=[centroid_y],
|
20 |
+
mode="markers",
|
21 |
+
marker=dict(size=15, color="black", symbol="x"),
|
22 |
+
name=f"Cluster {cluster} Center",
|
23 |
+
)
|
24 |
+
|
25 |
+
return fig
|
26 |
+
|
27 |
+
def visualize_ratings(sampled_df):
|
28 |
+
fig = px.bar(
|
29 |
+
sampled_df,
|
30 |
+
x="title",
|
31 |
+
y="rating_avg",
|
32 |
+
color="cluster",
|
33 |
+
title="论文评分分布",
|
34 |
+
hover_data=["keywords", "confidence_avg", "author"],
|
35 |
+
)
|
36 |
+
|
37 |
+
fig.update_layout(
|
38 |
+
xaxis_title="论文标题",
|
39 |
+
yaxis_title="平均评分",
|
40 |
+
xaxis_tickangle=-45,
|
41 |
+
)
|
42 |
+
|
43 |
+
return fig
|