0%

在Flask中的一种无需生成zip文件的下载方式

背景

在Flask中对文件进行归档并下载时,一般会使用zipfile库。由于zipfile库必须先生成zip文件再进行下载,所以当生成的zip文件较多且较大时,会逐渐消耗磁盘的容量。为了避免这种情况,我们可以使用zipstream库。

使用zipstream库

zipstream可以不生成zip文件,直接下载:

1
2
3
4
5
6
7
8
@app.route('/package.zip', methods=['GET'], endpoint='zipball')
def zipball():
z = zipstream.ZipFile(mode='w', compression=ZIP_DEFLATED)
z.write('/path/to/file')

response = Response(z, mimetype='application/zip')
response.headers['Content-Disposition'] = 'attachment; filename={}'.format('files.zip')
return response

使用生成器的方式进行流式下载:

1
2
3
4
5
6
7
8
9
10
11
@app.route('/package.zip', methods=['GET'], endpoint='zipball')
def zipball():
def zip_generator():
z = zipstream.ZipFile(mode='w', compression=ZIP_DEFLATED)
z.write('/path/to/file')
for chunk in z:
yield chunk

response = Response(generator(), mimetype='application/zip')
response.headers['Content-Disposition'] = 'attachment; filename={}'.format('files.zip')
return response

以迭代器的形式生成zip文件,避免将大文件存储在内存中:

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
@app.route('/package.zip', methods=['GET'], endpoint='zipball')
def zipball():
def read_file(path):
with open(path, 'rb') as f:
while True:
data = f.read(20 * 1024 * 1024)
if not data:
break
yield data

def zip_generator():
z = zipstream.ZipFile(mode='w', compression=ZIP_DEFLATED)
z.write_iter('arcname', read_file('/path/to/file'))
for chunk in z:
yield chunk

response = Response(generator(), mimetype='application/zip')
response.headers['Content-Disposition'] = 'attachment; filename={}'.format('files.zip')
return response

参考文档

python-zipstream github