|
来自danboor的根据角色名爬取的图片和对应TXT文件。 |
|
图片如果尺寸超过2048,resize至2048内。 |
|
但TXT文件有命名的不妥,还未处理。 |
|
如图片:000e31bd5414370380390abb08b16a40.png |
|
但TXT:000e31bd5414370380390abb08b16a40.png.txt |
|
其中Video或者Gif不进行处理。 |
|
且考虑到有Video或者Gif文件,其有对应的TXT,但截取其中一帧,但依旧有帧不同,因为对应Viedo和gif包含多帧,其只对应着一个TXT文件,极大容易TXT的TAG内容与某一帧对不上。 |
|
1.尺寸超过2048,reszie至2048内。 |
|
2.对视频文件不进行处理。 |
|
3.以角色名进行爬取,仅爬取danbooru,角色图片数量>500的。 |
|
4.TXT文件TAG已去除下划线,换成空格。 |
|
----------------------------------------------------------------- |
|
Images and corresponding TXT files crawled by character names from danbooru. |
|
If the image size exceeds 2048, resize it to within 2048. |
|
However, the TXT file is not properly named and has not been processed. |
|
For example, the image: 000e31bd5414370380390abb08b16a40.png |
|
But TXT: 000e31bd5414370380390abb08b16a40.png.txt |
|
Among them, Video or Gif is not processed. |
|
And considering that there are Video or Gif files, they have corresponding TXT, but one frame is intercepted, but there are still different frames, because the corresponding Viedo and gif contain multiple frames, which only correspond to one TXT file, and it is very easy for the TAG content of TXT to not match a certain frame. |
|
1. If the size exceeds 2048, resize it to within 2048. |
|
2. The video file is not processed. |
|
3. Crawl by character name, only crawl danbooru, and character images with > 500. |
|
4. The underline in TXT file TAG has been removed and replaced with a space. |