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:
@app.route('/package.zip', methods=['GET'])
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:
@app.route('/package.zip', methods=['GET'])
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(zip_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:
@app.route('/package.zip', methods=['GET'])
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(zip_generator(), mimetype='application/zip')
response.headers['Content-Disposition'] = 'attachment; filename={}'.format('files.zip')
return response