mirror of
https://github.com/tencentmusic/cube-studio.git
synced 2024-12-15 06:09:57 +08:00
549 lines
14 KiB
Python
549 lines
14 KiB
Python
import numpy
|
||
import pandas,csv,json,os,sys,datetime,time
|
||
import pysnooper
|
||
|
||
# @pysnooper.snoop()
|
||
def draw_line(df):
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
data = df.iloc[:, 1:] # 后续的列为多条折线数据
|
||
|
||
series=[]
|
||
for index,legend in enumerate(legends):
|
||
serie = data.iloc[:, index].to_list()
|
||
serie_str = {
|
||
"name": legend,
|
||
"type": 'line',
|
||
"smooth": True,
|
||
"stack": 'Total',
|
||
"data": serie
|
||
}
|
||
series.append(serie_str)
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"legend": {
|
||
"data": legends
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"xAxis": {
|
||
"type": 'category',
|
||
"boundaryGap": False,
|
||
"data": xs
|
||
},
|
||
"yAxis": {
|
||
"type": 'value'
|
||
},
|
||
"series": series
|
||
}
|
||
options = json.dumps(options,indent=4,ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
|
||
def draw_bar(df):
|
||
columns = df.columns.tolist()
|
||
# 将列名和数据合并为一个二维数组
|
||
data = numpy.vstack([columns, df.values]).tolist()
|
||
print(data)
|
||
options = {
|
||
"legend": {},
|
||
"tooltip": {},
|
||
"dataset": {
|
||
"source": data
|
||
},
|
||
"xAxis": { "type": 'category' },
|
||
"yAxis": {},
|
||
"series": [{ "type": 'bar' }, { "type": 'bar' }, { "type": 'bar' }]
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
def draw_pie(df):
|
||
data = df.iloc[:, 1:]
|
||
# print(data)
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
series=[]
|
||
titles=[]
|
||
for index,legend in enumerate(legends):
|
||
serie = data.iloc[:, index].to_list()
|
||
serie_data=[]
|
||
for x_index,x in enumerate(xs):
|
||
serie_data.append({
|
||
"name":x,
|
||
"value":serie[x_index]
|
||
})
|
||
left = int((index+1)*100/(len(legends)+1))
|
||
radius = 100//(len(legends)+1)
|
||
series.append({
|
||
"type": 'pie',
|
||
"radius": f"{radius}%",
|
||
"center": [f'{left}%','50%'],
|
||
"datasetIndex": index+1,
|
||
"label": {
|
||
"position": 'outer'
|
||
},
|
||
"data": serie_data,
|
||
})
|
||
titles.append({
|
||
"subtext": legend,
|
||
"left": f'{left}%',
|
||
"top": '80%',
|
||
"textAlign": 'center'
|
||
})
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'item'
|
||
},
|
||
"legend": {
|
||
"orient": 'horizontal'
|
||
},
|
||
"title":titles,
|
||
"series": series
|
||
}
|
||
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
|
||
def draw_scatter(df):
|
||
x = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
data = df.iloc[:, 1:] # 后续的列为多条折线数据
|
||
|
||
series = []
|
||
for index, legend in enumerate(legends):
|
||
serie = data.iloc[:, index].to_list()
|
||
serie_str = {
|
||
"name": legend,
|
||
"type": 'scatter',
|
||
"smooth": True,
|
||
"stack": 'Total',
|
||
"data": serie
|
||
}
|
||
series.append(serie_str)
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"legend": {
|
||
"data": legends
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"xAxis": {
|
||
"type": 'category',
|
||
"boundaryGap": False,
|
||
"data": x
|
||
},
|
||
"yAxis": {
|
||
"type": 'value'
|
||
},
|
||
"series": series
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
def draw_radar(df):
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
data = df.iloc[:, 1:] # 后续的列为多条折线数据
|
||
|
||
series = []
|
||
|
||
for index, legend in enumerate(legends):
|
||
serie = data.iloc[:, index].to_list()
|
||
serie_str = {
|
||
"name": legend,
|
||
"value": serie
|
||
}
|
||
series.append(serie_str)
|
||
|
||
# 获取每个x的最大值
|
||
indicator = []
|
||
for index, x in enumerate(xs):
|
||
max_value = max(data.iloc[index].to_list())
|
||
indicator.append({
|
||
"name":x,
|
||
"max":max_value
|
||
})
|
||
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"legend": {
|
||
"data": legends
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"radar": {
|
||
"shape": 'circle',
|
||
"indicator": indicator
|
||
|
||
},
|
||
"series": [
|
||
{
|
||
"type": 'radar',
|
||
"data":series
|
||
}
|
||
]
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
# 绘制k线图，第一列是x，后面4列是k线，后面的列是折线
|
||
# @pysnooper.snoop()
|
||
def draw_candlestick(df):
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
ys = df.iloc[:, 1:]
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
|
||
if len(legends)<4: # 至少要4列
|
||
return ''
|
||
|
||
k_series=df.iloc[:, 1:5].values.tolist()
|
||
line_series =[]
|
||
if len(legends)>4:
|
||
for index,legend in enumerate(legends):
|
||
if index<4:
|
||
continue
|
||
line_series.append(
|
||
{
|
||
"name": legend,
|
||
"type": 'line',
|
||
"data": ys.iloc[:, index].to_list(),
|
||
"smooth": True,
|
||
"lineStyle": {
|
||
"opacity": 0.5
|
||
}
|
||
}
|
||
)
|
||
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"legend": {
|
||
"data": legends
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"xAxis": {
|
||
"data": xs
|
||
},
|
||
"yAxis": {},
|
||
"series": [
|
||
{
|
||
"type": 'candlestick',
|
||
"data": k_series
|
||
}
|
||
]+line_series
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
|
||
def draw_heatmap(df):
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
ys = df.iloc[:, 1:]
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
|
||
data = pandas.melt(df.reset_index(), id_vars=['index'], value_vars=df.columns)
|
||
data = data.values.tolist()
|
||
# 获取所有元素的最大值和最小值
|
||
max_value = float(ys.max().max())
|
||
min_value = float(ys.min().min())
|
||
# print(max_value,min_value)
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"legend": {
|
||
"data": legends
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"xAxis": {
|
||
"type": 'category',
|
||
"data": xs,
|
||
"splitArea": {
|
||
"show": True
|
||
}
|
||
},
|
||
"yAxis": {
|
||
"type": 'category',
|
||
"data": legends,
|
||
"splitArea": {
|
||
"show": True
|
||
}
|
||
},
|
||
"visualMap": {
|
||
"calculable": True,
|
||
"orient": 'horizontal',
|
||
"left": 'center',
|
||
"min":min_value,
|
||
"max":max_value
|
||
},
|
||
"series": [
|
||
{
|
||
"type": 'heatmap',
|
||
"data": data,
|
||
"label": {
|
||
"show": True
|
||
}
|
||
}
|
||
]
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
def draw_tree(data):
|
||
|
||
# 如果不是一个根，就聚合成一个根
|
||
if type(data)==list:
|
||
data = {
|
||
"name": 'root',
|
||
"children": data
|
||
}
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'axis'
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"series": [
|
||
{
|
||
"type": 'tree',
|
||
"top": '10%',
|
||
"left": '10%',
|
||
"bottom": '10%',
|
||
"right": '10%',
|
||
"edgeShape": 'polyline',
|
||
"initialTreeDepth": 10,
|
||
"label": {
|
||
"backgroundColor": '#fff',
|
||
"position": 'left',
|
||
"verticalAlign": 'middle',
|
||
"align": 'right'
|
||
},
|
||
"leaves": {
|
||
"label": {
|
||
"position": 'right',
|
||
"verticalAlign": 'middle',
|
||
"align": 'left'
|
||
}
|
||
},
|
||
"expandAndCollapse": True,
|
||
"data": [data]
|
||
}
|
||
]
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
|
||
|
||
def draw_sunburst(data):
|
||
options = {
|
||
"series": {
|
||
"type": 'sunburst',
|
||
"data": data,
|
||
"radius": [0, '90%'],
|
||
"label": {
|
||
"rotate": 'radial'
|
||
}
|
||
}
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
def draw_funnel(df):
|
||
xs = df.iloc[:, 0].to_list() # 读取第一例为x
|
||
legends = df.columns.tolist()[1:] # 首行为header
|
||
data = df.iloc[:, 1:] # 后续的列为多条折线数据
|
||
series=[]
|
||
for index,legend in enumerate(legends):
|
||
serie = data.iloc[:, index].to_list()
|
||
opacity = 1-index/len(legends)
|
||
serie_data = []
|
||
for x_index, x in enumerate(xs):
|
||
serie_data.append({
|
||
"name": x,
|
||
"value": serie[x_index]
|
||
})
|
||
|
||
serie_str = {
|
||
"name": legend,
|
||
"type": 'funnel',
|
||
"labelLine": {
|
||
"show": False
|
||
},
|
||
"itemStyle": {
|
||
"opacity": opacity
|
||
},
|
||
"data": serie_data
|
||
}
|
||
series.append(serie_str)
|
||
|
||
|
||
options = {
|
||
"tooltip": {
|
||
"trigger": 'item',
|
||
"formatter": '{a} <br/>{b} : {c}%'
|
||
},
|
||
"legend": {
|
||
"data": xs
|
||
},
|
||
"toolbox": {
|
||
"feature": {
|
||
"saveAsImage": {}
|
||
}
|
||
},
|
||
"series": series
|
||
}
|
||
options = json.dumps(options, indent=4, ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
# 绘制平行坐标系
|
||
# @pysnooper.snoop()
|
||
def draw_parallel(df):
|
||
columns_type=df.dtypes.tolist()
|
||
columns=df.columns.tolist()
|
||
parallelAxis=[]
|
||
visualMap_min=None
|
||
visualMap_max=None
|
||
for index,column in enumerate(columns):
|
||
schema={
|
||
"dim":index,
|
||
"name":column
|
||
}
|
||
column_type = columns_type[index]
|
||
# 对于字符串枚举类型，设置类目
|
||
from numpy import dtype
|
||
if column_type==dtype('O'):
|
||
schema['type']='category'
|
||
else:
|
||
# 设置最大最小值
|
||
one_column_data = df[column].to_list()
|
||
min_value,max_value = min(one_column_data),max(one_column_data)
|
||
schema['min'] = min_value
|
||
schema['max'] = max_value
|
||
visualMap_min = min_value
|
||
visualMap_max = max_value
|
||
parallelAxis.append(schema)
|
||
|
||
options = {
|
||
"parallelAxis": parallelAxis,
|
||
"tooltip": {
|
||
"padding": 10,
|
||
"backgroundColor": '#222',
|
||
"borderColor": '#777',
|
||
"borderWidth": 1
|
||
},
|
||
"visualMap": {
|
||
"show": True,
|
||
"min":visualMap_min,
|
||
"max":visualMap_max,
|
||
"color": ['#d94e5d', '#eac736', '#50a3ba']
|
||
},
|
||
"series": {
|
||
"type": 'parallel',
|
||
"lineStyle": {
|
||
"width": 2
|
||
},
|
||
"data": df.values.tolist() # 完成的把数据集取出来
|
||
}
|
||
}
|
||
options = json.dumps(options,indent=4,ensure_ascii=False)
|
||
print(options)
|
||
return options
|
||
|
||
# @pysnooper.snoop()
|
||
def draw(chart_type,file_path):
|
||
try:
|
||
if '.csv' in file_path:
|
||
example = pandas.read_csv(file_path, header=0)
|
||
elif '.json' in file_path:
|
||
example = json.load(open(file_path))
|
||
else:
|
||
return ''
|
||
if chart_type=='line':
|
||
return draw_line(df=example)
|
||
if chart_type=='bar':
|
||
return draw_bar(df=example)
|
||
if chart_type=='pie':
|
||
return draw_pie(df=example)
|
||
if chart_type=='scatter':
|
||
return draw_scatter(df=example)
|
||
if chart_type=='radar':
|
||
return draw_radar(df=example)
|
||
if chart_type=='candlestick':
|
||
return draw_candlestick(df=example)
|
||
if chart_type=='heatmap':
|
||
return draw_heatmap(df=example)
|
||
if chart_type=='tree':
|
||
return draw_tree(data=example)
|
||
if chart_type=='sunburst':
|
||
return draw_sunburst(data=example)
|
||
if chart_type=='parallel':
|
||
return draw_parallel(df=example)
|
||
except Exception as e:
|
||
print(e)
|
||
if __name__=="__main__":
|
||
example = pandas.read_csv('data.csv', header=0)
|
||
# draw_line(df=example)
|
||
# draw_bar(df=example)
|
||
# draw_pie(df=example)
|
||
# draw_scatter(df=example)
|
||
# draw_radar(df=example)
|
||
# draw_candlestick(df=example)
|
||
# draw_heatmap(df=example)
|
||
# draw_funnel(df=example)
|
||
draw_parallel(df=example)
|
||
|
||
# example = json.load(open('data.json'))
|
||
# draw_tree(data=example)
|
||
# draw_sunburst(data=example)
|
||
|
||
# import json
|
||
# metric = {
|
||
# "metric_type":"echart-parallel",
|
||
# "file_path":"/mnt/admin/pipeline/ray_hyperparams/hyperparams_r_lightgbm.csv"
|
||
# }
|
||
# json.dump(metric,open('/metric.json',mode='w')) |