53 Commits
0.1 ... 0.2.1

Author SHA1 Message Date
wenead99
9c495c4a54 Update README.md 2019-05-27 23:57:39 +08:00
wenead99
51e0467be0 Update README.md 2019-05-27 23:31:15 +08:00
wenead99
59720d8c09 Update README.md 2019-05-27 23:25:01 +08:00
wenead99
150943694b Update README.md 2019-05-27 23:13:55 +08:00
wenead99
f40683e5ec Update README.md 2019-05-27 22:57:58 +08:00
wenead99
912e5ed0fb Update README.md 2019-05-27 22:52:25 +08:00
wenead99
f17994ecf8 Update README.md 2019-05-27 22:49:22 +08:00
wenead99
a414c8eae9 Update README.md 2019-05-27 22:46:14 +08:00
wenead99
8f0eed64c7 Update README.md 2019-05-27 21:57:41 +08:00
wenead99
5f8131f984 Update README.md 2019-05-27 21:55:25 +08:00
wenead99
a574d3a275 Update README.md 2019-05-27 18:37:33 +08:00
wenead99
1813102b0b Update README.md 2019-05-27 18:17:43 +08:00
wenead99
0ef0fed958 Add files via upload 2019-05-27 17:52:07 +08:00
wenead99
64be029c78 Delete readme5.png 2019-05-27 17:51:36 +08:00
wenead99
0f77b78133 Delete single.gif 2019-05-27 17:49:10 +08:00
wenead99
b87f41dc8e Delete readme5.png 2019-05-27 17:49:01 +08:00
wenead99
ecdacdbcb8 Delete readme4.PNG 2019-05-27 17:48:51 +08:00
wenead99
8308f66ae7 Delete readme3.PNG 2019-05-27 17:48:42 +08:00
wenead99
44c89590ee Delete readme2.PNG 2019-05-27 17:48:33 +08:00
wenead99
e7b0980524 Delete readme1.PNG 2019-05-27 17:48:23 +08:00
wenead99
1cc328bbd7 Delete flow_chart2.png 2019-05-27 17:48:14 +08:00
wenead99
66fbb5efbb Update README.md 2019-05-27 17:47:51 +08:00
wenead99
88fded90ee 这是教程图片文件夹
This is readms.md's images folder
2019-05-27 17:47:01 +08:00
wenead99
b055f6ca7f Create This is readms.md's images folder 2019-05-27 17:45:12 +08:00
wenead99
254e37d9cf Update README.md 2019-05-27 17:25:41 +08:00
wenead99
1c87c26f32 Update README.md 2019-05-27 17:24:18 +08:00
wenead99
10f13f882f 福利 2019-05-27 17:20:28 +08:00
wenead99
df17ee59f0 Update README.md 2019-05-27 17:19:53 +08:00
wenead99
07fa18080b Update README.md 2019-05-27 17:11:12 +08:00
wenead99
2bdf2ff283 Update README.md 2019-05-27 17:10:31 +08:00
wenead99
5445f1773c 教程gif 2019-05-27 17:08:34 +08:00
wenead99
b629fb4615 Add files via upload 2019-05-27 17:00:16 +08:00
wenead99
d50f5d2f34 Update README.md 2019-05-27 16:59:55 +08:00
wenead99
13f36ddf8b Delete flow_chart.png 2019-05-27 16:59:39 +08:00
wenead99
295ea2d174 新流程图 0.2 Beta 2019-05-27 16:58:19 +08:00
wenead99
ad4fc237b1 删除旧流程图 2019-05-27 16:57:55 +08:00
wenead99
55d8f02eee 删除旧流程图 2019-05-27 16:57:25 +08:00
wenead99
d6cbd3bdb2 新流程图 for 0.2 Beta 2019-05-27 16:56:57 +08:00
wenead99
99942745c4 Update README.md 2019-05-27 16:55:40 +08:00
wenead99
4daee989e6 新流程图 for 0.2 Beta 2019-05-27 16:55:20 +08:00
wenead99
99b04ef8b5 0.2 beta 更新
2019.5.27更新:
1.支持多媒体格式
2.更智能的程序终止机制
3.改善程序源码本体
2019-05-27 16:47:26 +08:00
wenead99
9707f1b38a 0.2 beta 更新
2019.5.27更新:
1.支持多媒体格式
2.更智能的程序终止机制
3.改善程序源码本体
2019-05-27 16:46:27 +08:00
wenead99
0a9bdd8729 Update README.md 2019-05-27 16:09:16 +08:00
wenead99
8a091a710e Update README.md 2019-05-27 15:39:08 +08:00
wenead99
c5a0ef1f77 Update README.md 2019-05-27 15:38:00 +08:00
wenead99
00b03a4ec4 优化readme 2 2019-05-27 15:37:36 +08:00
wenead99
6b4d9ca829 Add files via upload 2019-05-27 15:36:20 +08:00
wenead99
ef93ba5eb5 Update README.md
优化readme.md
2019-05-27 15:35:32 +08:00
wenead99
46b4557471 Update README.md 2019-05-27 00:01:52 +08:00
wenead99
6b2205887f Update AV_Data_Capture.py 2019-05-26 22:38:22 +08:00
wenead99
3e0dc60bad Update AV_Data_Capture.py 2019-05-26 22:37:11 +08:00
wenead99
424ede24eb Update README.md 2019-05-26 21:29:43 +08:00
wenead99
e307ec21e3 Update README.md 2019-05-26 20:38:00 +08:00
11 changed files with 118 additions and 37 deletions

View File

@@ -2,12 +2,38 @@ import glob
import os
import time
#a=glob.glob(os.getcwd()+r"\*.py")
a=glob.glob(os.getcwd()+r"\*\**\*.mp4")
b=glob.glob(os.getcwd()+r"\*.mp4")
for i in b:
a.append(i)
def movie_lists():
#MP4
a1 = glob.glob(os.getcwd() + r"\*\**\*.mp4")
a2 = glob.glob(os.getcwd() + r"\*.mp4")
# AVI
b1 = glob.glob(os.getcwd() + r"\*\**\*.avi")
b2 = glob.glob(os.getcwd() + r"\*.avi")
# RMVB
c1 = glob.glob(os.getcwd() + r"\*\**\*.rmvb")
c2 = glob.glob(os.getcwd() + r"\*.rmvb")
# WMV
d1 = glob.glob(os.getcwd() + r"\*\**\*.wmv")
d2 = glob.glob(os.getcwd() + r"\*.wmv")
# MOV
e1 = glob.glob(os.getcwd() + r"\*\**\*.mov")
e2 = glob.glob(os.getcwd() + r"\*.mov")
# MKV
f1 = glob.glob(os.getcwd() + r"\*\**\*.mkv")
f2 = glob.glob(os.getcwd() + r"\*.mkv")
# FLV
g1 = glob.glob(os.getcwd() + r"\*\**\*.flv")
g2 = glob.glob(os.getcwd() + r"\*.flv")
total = a1+a2+b1+b2+c1+c2+d1+d2+e1+e2+f1+f2+g1+g2
return total
def lists_from_test(custom_nuber):
a=[]
a.append(custom_nuber)
return a
os.chdir(os.getcwd())
for i in a:
for i in movie_lists():
os.system('python core.py'+' "'+i+'"')
print("[+]All finished!!!")

View File

@@ -1,28 +1,66 @@
日本AV元数据抓取工具
========================
**关于本软件**
# 日本AV元数据抓取工具 (刮削器)
## 关于本软件 ~路star谢谢
目前我下的AV越来越多也意味着AV要集中地管理形成媒体库。现在有两款主流的AV元数据获取器"EverAver"和"Javhelper"。前者的优点是元数据获取比较全,缺点是不能批量处理;后者优点是可以批量处理,但是元数据不够全。
为此我写出了本软件为了方便的管理本地AV和更好的手冲体验。
为此,综合上述软件特点,我写出了本软件为了方便的管理本地AV和更好的手冲体验。没女朋友怎么办ʅ(‾◡◝)ʃ
**如何使用**
1. 请安装requests,pyquery,lxml,Beautifulsoup,pillow模块,可在CMD逐条输入以下命令安装
`pip install requests`
`pip install pyquery`
`pip install lxml`
`pip install beautifulsoup4`
`pip install pillow`
2. 你的AV在被软件管理前最好命名为番号:**COSQ-004.mp4**
文件名中间要有减号"-",没有多余元数据只有番号为最佳,可以让软件更好获取元数据
**tg官方电报群:https://t.me/AV_Data_Capture_Official**
![](readme1.PNG)
### **请认真阅读下面使用说明再使用**
![](readme/readme5.png)
## 软件流程图 (下下一个为使用教程)
![](readme/flow_chart2.png)
## 如何使用 (使用前请认真阅读下文)
**release的程序可脱离python环境运行可跳过第一步仅限windows平台)**
**下载地址(Windows):https://github.com/wenead99/AV_Data_Capture/releases**
1. 请安装requests,pyquery,lxml,Beautifulsoup4,pillow模块,可在CMD逐条输入以下命令安装
```python
pip install requests
```
###
```python
pip install pyquery
```
###
```python
pip install lxml
```
###
```python
pip install Beautifulsoup4
```
###
```python
pip install pillow
```
2. 你的AV在被软件管理前最好命名为番号:
```
COSQ-004.mp4
```
文件名中间要有减号"-",没有多余的内容只有番号为最佳,可以让软件更好获取元数据
对于多影片重命名可以用ReNamer来批量重命名
软件官网:http://www.den4b.com/products/renamer
![](readme/readme1.PNG)
3. 把软件拷贝到AV的所在目录下运行程序中国大陆用户必须挂VPNShsadowsocks开全局代理
4. 运行AV_Data_capture.py
5. 软件会自动把元数据获取成功的电影移动到JAV_output文件夹中根据女优分类失败的电影移动到failed文件夹中。
5. **你也可以把单个影片拖动到core程序**
![](readme/single.gif)
6. 软件会自动把元数据获取成功的电影移动到JAV_output文件夹中根据女优分类失败的电影移动到failed文件夹中。
![](readme/readme2.PNG)
![](readme/readme3.PNG)
![](readme/readme4.PNG)
![](readme2.PNG)
![](readme3.PNG)
![](readme4.PNG)

40
core.py
View File

@@ -129,33 +129,49 @@ if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument("file", help="Write the file path on here")
args = parser.parse_args()
#===============================
#获取文件名称
filename=str(os.path.basename(args.file)) #\[\d{4}(\-|\/|.)\d{1,2}\1\d{1,2}\]
#去除文件名中日期
#print(filename)
deldate=str(re.sub("\[\d{4}-\d{1,2}-\d{1,2}\] - ","",filename))
#print(deldate)
number=str(re.search('\w+-\w+',deldate).group())
#检测是否可以获取番号,如果不行终止本程序
def number_getter():
print("[!]Making Data for ["+filename+"]")
try:
a = str(re.search('\w+-\w+', deldate).group())
return a
except:
print('[-]File '+filename+'`s number can not be caught')
print('[-]Move ' + filename + 'to failed folder')
filepath = str(args).replace("Namespace(file='", '').replace("')", '').replace('\\\\', '\\')
if not os.path.exists('failed/'): # 新建failed文件夹
os.makedirs('failed/')
if not os.path.exists('failed/'):
print("[-]failed!Dirs can not be make (Please run as Administrator)")
time.sleep(3)
os._exit(0)
shutil.move(filepath, str(os.getcwd())+'/failed/')
os._exit(0)
number=number_getter()
#print(number)
#获取网页信息
#获取网页HTML
html = get_html("https://www.javbus.com/"+str(number))
html_outline=get_html("https://www.dmm.co.jp/mono/dvd/-/detail/=/cid="+number.replace("-",''))
#处理超长文件夹名称
if len(getActor(html)) > 240:
path = 'JAV_output' + '/' + '超多人' + '/' + getNum(html)
path = 'JAV_output' + '/' + '超多人' + '/' + getNum(html) #path为影片+元数据所在目录
else:
path = 'JAV_output' + '/' + getActor(html) + '/' + getNum(html)
if not os.path.exists(path):
os.makedirs(path)
#文件路径处理
#print(str(args))
filepath = str(args).replace("Namespace(file='",'').replace("')",'').replace('\\\\', '\\')
#print(filepath)
houzhui = str(re.search('[.](AVI|RMVB|WMV|MOV|MP4|MKV|FLV|avi|rmvb|wmv|mov|mp4|mkv|flv)$',filepath).group())
print("[!]Making Data for ["+number+houzhui+"]")
#下载元数据
if not os.path.exists('failed/'):
#如果DownloadFileWithFilename返回为failed就退出本程序
if not os.path.exists('failed/'): #新建failed文件夹
os.makedirs('failed/')
if not os.path.exists('failed/'):
print("[-]failed!Dirs can not be make (Please run as Administrator)")
@@ -180,7 +196,7 @@ if __name__ == '__main__':
# 电源文件位置处理
os.rename(filepath, number + houzhui)
shutil.move(number + houzhui, path)
#处理元数据
#下载元数据
PrintFiles(path)
print('[!]Finished!')
time.sleep(3)
time.sleep(3)

View File

@@ -0,0 +1 @@
1

BIN
readme/flow_chart2.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 101 KiB

View File

Before

Width:  |  Height:  |  Size: 1.1 KiB

After

Width:  |  Height:  |  Size: 1.1 KiB

View File

Before

Width:  |  Height:  |  Size: 3.4 KiB

After

Width:  |  Height:  |  Size: 3.4 KiB

View File

Before

Width:  |  Height:  |  Size: 1.3 KiB

After

Width:  |  Height:  |  Size: 1.3 KiB

View File

Before

Width:  |  Height:  |  Size: 16 KiB

After

Width:  |  Height:  |  Size: 16 KiB

BIN
readme/readme5.png Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 457 KiB

BIN
readme/single.gif Normal file

Binary file not shown.

After

Width:  |  Height:  |  Size: 68 KiB