如何使用 Seafile 生成的上传链接，无需从命令行进行身份验证令牌

问题描述 投票：0回答：3

使用 Seafile，人们可以创建一个公共上传链接（例如 

https://cloud.seafile.com/u/d/98233edf89/
），通过浏览器上传文件，无需身份验证。

Seafile webapi 不支持任何不带身份验证令牌的上传。

我如何从带有curl的命令行或从python脚本使用这种链接？

python curl urllib2 http-upload seafile-server
3个回答
9
投票

用curl花了2个小时才找到解决方案，需要两个步骤：

  1. 使用 
    repo-id
    作为查询参数向公共上行链路 url 发出 get 请求，如下所示：

curl 'https://cloud.seafile.com/ajax/u/d/98233edf89/upload/?r=f3e30b25-aad7-4e92-b6fd-4665760dd6f5' -H 'Accept: application/json' -H 'X-Requested-With: XMLHttpRequest'

答案是 (json) 在下一个上传帖子中使用的 id 链接，例如：

{"url": "https://cloud.seafile.com/seafhttp/upload-aj/c2b6d367-22e4-4819-a5fb-6a8f9d783680"}

  1. 使用此链接启动上传帖子：

curl 'https://cloud.seafile.com/seafhttp/upload-aj/c2b6d367-22e4-4819-a5fb-6a8f9d783680' -F file=@./tmp/index.html -F filename=index.html -F parent_dir="/my-repo-dir/"

答案又是json，例如

[{"name": "index.html", "id": "0a0742facf24226a2901d258a1c95e369210bcf3", "size": 10521}]

完成;)

1
投票
#!/usr/bin/env bash

# this script depend on jq,check it first
RED='\033[0;31m'
NC='\033[0m' # No Color
if ! command -v jq &> /dev/null
then
    echo -e "${RED}jq could not be found${NC}, installed and restart plz!\n"
    exit
fi


usage () { echo "Usage : $0 -u <username> -p <password> -h <seafile server host> -f <upload file path> -d <parent dir default value is /> -r <repo id> -t <print debug info switch off/on,default off>"; }

# parse args
while getopts "u:p:h:f:d:r:t:" opts; do
   case ${opts} in
      u) USER=${OPTARG} ;;
      p) PASSWORD=${OPTARG} ;;
      h) HOST=${OPTARG} ;;
      f) FILE=${OPTARG} ;;
      d) PARENT_DIR=${OPTARG} ;;
      r) REPO=${OPTARG} ;;
      t) DEBUG=${OPTARG} ;;
      *) usage; exit;;
   esac
done

# those args must be not null
if [ ! "$USER" ] || [ ! "$PASSWORD" ] || [ ! "$HOST" ] || [ ! "$FILE" ] || [ ! "$REPO" ]
then
    usage
    exit 1
fi

# optional args,set default value

[ -z "$DEBUG" ] && DEBUG=off

[ -z "$PARENT_DIR" ] && PARENT_DIR=/

# print vars key and value when DEBUG eq on
[[ "on" == "$DEBUG" ]] && echo -e "USER:${USER} PASSWORD:${PASSWORD} HOST:${HOST} FILE:${FILE} PARENT_DIR:${PARENT_DIR} REPO:${REPO} DEBUG:${DEBUG}"

# login and get token
TOKEN=$(curl -s --location --request POST "${HOST}/api2/auth-token/" --header 'Content-Type: application/x-www-form-urlencoded' --data-urlencode "username=${USER}" --data-urlencode "password=${PASSWORD}" | jq -r ".token")

[ -z "$TOKEN" ] && echo -e "${RED}login seafile faild${NC}, call your administrator plz!\n" && exit 1

# gen upload link 
UPLOAD_LINK=$(curl -s --header "Authorization: Token ${TOKEN}" "${HOST}/api2/repos/${REPO}/upload-link/?p=${PARENT_DIR}" | jq -r ".")

[ -z "$UPLOAD_LINK" ] && echo -e "${RED}get upload link faild${NC}, call your administrator plz!\n" && exit 1

# upload file
UPLOAD_RESULT=$(curl -s --header "Authorization: Token ${TOKEN}" -F file="@${FILE}" -F filename=$(basename ${FILE}) -F parent_dir="${PARENT_DIR}" -F replace=1 "${UPLOAD_LINK}?ret-json=1")

[ -z "$UPLOAD_RESULT" ] && echo -e "${RED}faild to upload ${FILE}${NC}, call your administrator plz!\n" && exit 1

# print upload result
[[ "on" == "$DEBUG" ]] && echo -e "TOKEN:${TOKEN} UPLOAD_LINK:${UPLOAD_LINK} UPLOAD_RESULT:${UPLOAD_RESULT}"

另存为 

seafile-upload.sh

# ubuntu 
apt install -y jq
# centos 
# yum install -y jq
chmod +x ./seafile-upload.sh

./seafile-upload.sh -u <username> -p <password> -h <seafile server host> -f <upload file path> -d <parent dir default value is /,must be start with /> -r <repo id> -t <print debug info switch off/on,default off>

我的博客060-轻量级基于curl的seafile脚本

官方文档web-api/v2.1/file-upload.md

0
投票

根据我对浏览器中文件上传行为的观察，我编写了一个用于命令行使用的 Python 脚本。只需要上传页面链接（由 Seafile 生成）和本地文件路径。

我还将脚本作为 GitHub gist 上传。

使用方法

设置：

  • 安装 pip 包 
    requests
    和 
    beautifulsoup4
    。
  • （可选）安装
    requests_toolbelt
    和
    tqdm
    以查看实时上传进度。
usage: upload_seafile.py [-h] -l LINK -f FILE [FILE ...] [--verbose]

options:
  -h, --help            show this help message and exit
  -l LINK, --link LINK  upload page link (generated by seafile)
  -f FILE [FILE ...], --file FILE [FILE ...]
                        file(s) to upload
  --verbose             show detailed output

脚本

关键逻辑：

  1. 打开上传页面并提取
    parent_dir
    和
    token
    （无需登录）。
  2. 请求文件上传网址。该部分包括两种方法：一种是指浏览器中使用的seafile API，另一种结合了lmen6e的答案中的方法。
  3. 上传文件，其中需要
    parent_dir
    参数。

代码：

import re
import argparse
import sys
import copy
from pathlib import Path
from urllib.parse import urlparse

import requests
from bs4 import BeautifulSoup

optional_packages = True
try:
    # optional, for upload progess updates
    from requests_toolbelt import MultipartEncoder, MultipartEncoderMonitor
    from tqdm import tqdm
except ImportError:
    optional_packages = False


def extract_var(script_text, variable_name, default=None):
    if variable_name in script_text:
        # match: var_name: "value" or var_name: 'value' or var_name = "value" or var_name = 'value'
        pattern = re.compile(r'{}\s*[:=]\s*(["\'])(.*?)\1'.format(re.escape(variable_name)))
        match = pattern.search(script_text)
        if match:
            return match.group(2)
    return default

def extract_info_from_html(html_content):
    soup = BeautifulSoup(html_content, 'html.parser')
    scripts = soup.find_all('script')
    token = parent_dir = repo_id = dir_name = None
    for script in scripts:
        token = extract_var(script.text, 'token', token)
        parent_dir = extract_var(script.text, 'path', parent_dir)
        repo_id = extract_var(script.text, 'repoID', repo_id)
        dir_name = extract_var(script.text, 'dirName', dir_name)
    return token, parent_dir, repo_id, dir_name

def get_html_content(url):
    response = requests.get(url)
    return response.text

def get_upload_url(api_url):
    response = requests.get(api_url)
    if response.status_code == 200:
        return response.json().get('upload_link')
    return None

def get_upload_url2(api_url):
    headers = {
        'Accept': 'application/json',
        'X-Requested-With': 'XMLHttpRequest'
    }
    response = requests.get(api_url, headers=headers)
    if response.status_code == 200:
        return response.json().get('url')
    return None

def upload_file(upload_url, file_path, fields):
    fields = copy.deepcopy(fields)
    path = Path(file_path)
    filename = path.name
    total_size = path.stat().st_size
    
    if not optional_packages:
        with open(file_path, 'rb') as f:
            fields["file"] = (filename, f)
            response = requests.post(upload_url, files=fields, params={'ret-json': 'true'})
        return response

    # ref: https://stackoverflow.com/a/67726532/11854304
    with tqdm(
        desc=filename,
        total=total_size,
        unit="B",
        unit_scale=True,
        unit_divisor=1024,
    ) as bar:
        with open(file_path, "rb") as f:
            fields["file"] = (filename, f)
            encoder = MultipartEncoder(fields=fields)
            monitor = MultipartEncoderMonitor(
                encoder, lambda monitor: bar.update(monitor.bytes_read - bar.n)
            )
            headers = {"Content-Type": monitor.content_type}
            response = requests.post(upload_url, headers=headers, data=monitor, params={'ret-json': 'true'})
    return response

def upload_seafile(upload_page_link, file_path_list, verbose):
    parsed_results = urlparse(upload_page_link)
    base_url = f"{parsed_results.scheme}://{parsed_results.netloc}"
    if verbose:
        print(f"Input:")
        print(f" * Upload page url: {upload_page_link}")
        print(f" * Files to be uploaded: {file_path_list}")
        print(f"Preparation:")
        print(f" * Base url: {base_url}")

    # get html content
    html_content = get_html_content(upload_page_link)

    # extract variables from html content
    token, parent_dir, repo_id, dir_name = extract_info_from_html(html_content)
    if not parent_dir:
        print(f"Cannot extract parent_dir from HTML content.", file=sys.stderr)
        return 1
    if verbose:
        print(f" * dir_name: {dir_name}")
        print(f" * parent_dir: {parent_dir}")

    # get upload url
    upload_url = None
    if token:
        # ref: https://github.com/haiwen/seafile-js/blob/master/src/seafile-api.js#L1164
        api_url = f'{base_url}/api/v2.1/upload-links/{token}/upload/'
        upload_url = get_upload_url(api_url)
    elif repo_id:
        # ref: https://stackoverflow.com/a/38743242/11854304
        api_url = upload_page_link.replace("/u/d/", "/ajax/u/d/").rstrip('/') + f"/upload/?r={repo_id}"
        upload_url = get_upload_url2(api_url)
    if not upload_url:
        print(f"Cannot get upload_url.", file=sys.stderr)
        return 1
    if verbose:
        print(f" * upload_url: {upload_url}")

    # upload each file
    print(f"Upload:")
    for idx, file_path in enumerate(file_path_list):
        print(f"({idx+1}) {file_path}")
        try:
            response = upload_file(upload_url, file_path, {'parent_dir': parent_dir})
            if response.status_code == 200:
                print(f"({idx+1}) upload completed: {response.json()}")
            else:
                print(f"({idx+1}) {file_path} ERROR: {response.status_code} {response.text}", file=sys.stderr)
        except Exception as e:
            print(f"({idx+1}) {file_path} EXCEPTION: {e}", file=sys.stderr)

    return 0


if __name__ == '__main__':
    parser = argparse.ArgumentParser()
    parser.add_argument('-l', '--link', required=True, help='upload page link (generated by seafile)')
    parser.add_argument('-f', '--file', required=True, nargs='+', help='file(s) to upload')
    parser.add_argument('--verbose', action='store_true', help='show detailed output')
    args = parser.parse_args()
    sys.exit(upload_seafile(args.link, args.file, args.verbose))
最新问题
© www.soinside.com 2019 - 2024. All rights reserved.