我将此作为input，我们称之为tree
if ( device_type_id <= 1 )
39 Clicks - 0.61%
2135 Conversions - 33.32%
else ( device_type_id > 1 )
if ( country_id <= 216 )
1097 Clicks - 17.12%
else ( country_id > 216 )
if ( browser_id <= 2 )
296 Clicks - 4.62%
else ( browser_id > 2 )
if ( browser_id <= 4 )
if ( browser_id <= 3 )
if ( operating_system_id <= 2 )
262 Clicks - 4.09%
1094 Impressions - 17.08%
else ( operating_system_id > 2 )
if ( operating_system_id <= 4 )
281 Clicks - 4.39%
220 Impressions - 3.43%
else ( operating_system_id > 4 )
if ( operating_system_id <= 6 )
4 Clicks - 0.06%
20 Impressions - 0.31%
else ( operating_system_id > 6 )
70 Impressions - 1.09%
else ( browser_id > 3 )
if ( operating_system_id <= 2 )
19 Clicks - 0.3%
21 Impressions - 0.33%
else ( operating_system_id > 2 )
19 Clicks - 0.3%
707 Impressions - 11.03%
else ( browser_id > 4 )
113 Clicks - 1.76%
然后我使用tree作为input：
def function_one(tree):
network = []
for line in tree.splitlines() :
if line.strip():
line = line.strip()
network.append(line)
else : break
if not line : break
res = []
res.append({'name':'Prediction Result', 'children':parser(network[1:])})
with open('static/json/structure_sklearn.json', 'w') as outfile:
json.dump(res, outfile)
return tree
正如您所看到的，我使用了parser函数：
def parser(lines):
block = []
while lines :
if lines[0].startswith('if'):
bl = ' '.join(lines.pop(0).split()[1:]).replace('(', '').replace(')', '')
block.append({'name':bl, 'children':parser(lines)})
if lines[0].startswith('else'):
be = ' '.join(lines.pop(0).split()[1:]).replace('(', '').replace(')', '')
block.append({'name':be, 'children':parser(lines)})
elif not lines[0].startswith(('if','else')):
block2 = lines.pop(0)
block.append({'name':block2})
else:
break
return block
我的问题是，我不知道在哪个阶段我错过了什么，因为从json创建的function_one文件只是：
[
{
"children":[
{
"name":"39 Clicks - 0.61%"
},
{
"name":"2135 Conversions - 33.32%"
}
],
"name":"Prediction Result"
}
]
答案 0 :(得分：1)
以下是我建议的修改：
def parser(lines):
block = []
while lines :
if lines[0].startswith('if'):
bl = ' '.join(lines.pop(0).split()[1:]).replace('(', '').replace(')', '')
block.append({'name':bl, 'children':parser(lines)})
elif lines[0].startswith('else'):
be = ' '.join(lines.pop(0).split()[1:]).replace('(', '').replace(')', '')
block.append({'name':be, 'children':parser(lines)})
elif not lines[0].startswith(('if','else')):
block2 = lines.pop(0)
block.append({'name':block2})
else:
break
return block
通过此更改，我得到以下JSON缩进输出：
[
{
"name": "Prediction Result",
"children": [
{
"name": "39 Clicks - 0.61%"
},
{
"name": "2135 Conversions - 33.32%"
},
{
"name": " device_type_id > 1 ",
"children": [
{
"name": " country_id <= 216 ",
"children": [
{
"name": "1097 Clicks - 17.12%"
},
{
"name": " country_id > 216 ",
"children": [
{
"name": " browser_id <= 2 ",
"children": [
{
"name": "296 Clicks - 4.62%"
},
{
"name": " browser_id > 2 ",
"children": [
{
"name": " browser_id <= 4 ",
"children": [
{
"name": " browser_id <= 3 ",
"children": [
{
"name": " operating_system_id <= 2 ",
"children": [
{
"name": "262 Clicks - 4.09%"
},
{
"name": "1094 Impressions - 17.08%"
},
{
"name": " operating_system_id > 2 ",
"children": [
{
"name": " operating_system_id <= 4 ",
"children": [
{
"name": "281 Clicks - 4.39%"
},
{
"name": "220 Impressions - 3.43%"
},
{
"name": " operating_system_id > 4 ",
"children": [
{
"name": " operating_system_id <= 6 ",
"children": [
{
"name": "4 Clicks - 0.06%"
},
{
"name": "20 Impressions - 0.31%"
},
{
"name": " operating_system_id > 6 ",
"children": [
{
"name": "70 Impressions - 1.09%"
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
}
]
我无法判断这是否是您正在寻找的实际表单或输出，但它似乎包含原始树源中的所有节点。
编辑：
这是使用迭代器遍历行列表的更简洁（也可能更快）的方法。
def parser(lines):
lines_iter = iter(lines)
block = []
for line in lines_iter:
if line.startswith('if'):
bl = ' '.join(line.split()[1:]).replace('(', '').replace(')', '')
block.append({'name':bl, 'children':parser(lines_iter)})
elif line.startswith('else'):
be = ' '.join(line.split()[1:]).replace('(', '').replace(')', '')
block.append({'name':be, 'children':parser(lines_iter)})
elif not line.startswith(('if','else')):
block2 = line
block.append({'name':block2})
else:
break
return block
我更改了function_one中的调用以调用parser(network)而不是parser(network[1:])，但其他代码保持不变。 list.pop(0)很慢，特别是对于大型列表，因为底层数组必须将每个元素向下移动一个。