Pixiv Fanbox Downloader?

[NHOrus]

Possible option to get images from people whom I patron?

[Nandaka]

I never use that feature, how does it works?

[NHOrus]

In simplest words, Patreon, but directly linked to pixiv account of all participants.

Do you need test pages?

[Nandaka]

Checked one of the pixiv fanbox, I can get the ajax call for retriving the posts
https://www.pixiv.net/ajax/fanbox/creator?userId=648285

and the post details:
https://www.pixiv.net/ajax/fanbox/post?postId=123882

both with empty body element, because I'm not the patron.

Do you have samples for those messages?

[Nandaka]

Got one with open post: https://www.pixiv.net/ajax/fanbox/post?postId=481
Looks like the body element is not structured, are you want to download the imageForShare or coverImageUrl?

[NHOrus]

originalUrl from body - type:image - body - images array
also, url from body - type:file - body - files array

Two examples are included.
post102135.txt
post132421.txt

[Nandaka]

any referer need to be passed, or you can just download the file based on the current cookies value?

[NHOrus]

If tests are correct - 403 when not logged in, image from empty page with just URL when logged in, cookies are enough.

[NHOrus]

... That was poorly worded.
Using any link results in 403 when not logged in
Link provides image when logged in, even not from pixiv page.

So, cookies.
Any other way I can help?

[NHOrus]

Maybe release beta3 before this?

[NHOrus]

https://www.pixiv.net/ajax/fanbox/index is the start url for the process

Following URLs look like https://www.pixiv.net/ajax/fanbox/supporting_latest_post?maxPublishedDatetime=2018-07-06%2018%3A35%3A12&maxId=96862&limit=10 and https://www.pixiv.net/ajax/fanbox/supporting_latest_post?maxPublishedDatetime=2018-07-23%2022%3A17%3A59&maxId=109759&limit=10

Attaching JSON files that were returned to me.
index.txt
supporting_latest_post_1.txt
supporting_latest_post_2.txt

[NHOrus]

Alternatively, it could be better to get list of people supported from https://www.pixiv.net/ajax/fanbox/support to extract userID from, then iterate over all https://www.pixiv.net/ajax/fanbox/creator?userId=15521131 and get from nextUrl link in the form of https://www.pixiv.net/ajax/fanbox/creator/post?userId=15521131&maxPublishedDatetime=2018-05-01%2022%3A11%3A59&maxId=26783&limit=10 with more posts. Triple indirection, but works generically.

nextURL.txt
creator.txt
page.txt

[NHOrus]

Let me be overly direct. What I need to donate for you to consider working on this?

[Nandaka]

Time? a bit busy with real works, so might not be possible for me to create new features.

[Nandaka]

@NHOrus is it possible for you to send you username/password to my mail, so I can test the feature, as I don't follow any fanbox (or do they have 0 yen plan?)

[NHOrus]

Well, hope it works successfully.

[Nandaka]

You can try this one.
pixivutil20181005-alpha1.zip

Please check the changelog.

[NHOrus]

It downloads one image from post when there's multiple. Because it tries to save them into same filename and overwrite is disabled.

[Nandaka]

Check your filenamemangaformat in config.ini. It is used to generate the images filename.
https://github.com/Nandaka/PixivUtil2/blob/master/PixivUtil2.py#L1865

As for the cover image, it uses the filenameFormat.
https://github.com/Nandaka/PixivUtil2/blob/master/PixivUtil2.py#L1812

I'm using %urlFilename% to make it unique, else you can try the combination of %image_id% and %page_index% or %page_number%?

The given filename for the images are randomized from the API, fyi.

[Nandaka]

also, I forgot to add the post id to the cover image, so for now it will shows only the original filename...

[NHOrus]

filenameformat = %member_id%\%image_id% - %title%
filenamemangaformat = %member_id%\%image_id% - %title%_%page_index%

I will try with %urlFilename%

[NHOrus]

Or with %page_number%, file names end with _

[NHOrus]

%page_number% and %page_index% don't work, only %urlFilename% does

[NHOrus]

with %title%_%urlFilename% I see stuff like 109759_p8_bvEkWxw2K8MXWTf9NWUtDGj7.png
Something got concatenated somewhere?

[Nandaka]

I'm using %urlFilename% - %title% and it does append the title in the back.

109759_p8_bvEkWxw2K8MXWTf9NWUtDGj7.png is the %urlFilename% portion.

[NHOrus]

I completely misunderstood what %urlFilename% does, I expected only hash-looking thing for that.
Thank you.

[Nandaka]

updated for cover filename, now it will include the post id, but without the _p.
pixivutil20181006-alpha2.zip

[NHOrus]

Will test it a bit later.

[NHOrus]

Well, it appears to be working correctly.

Thank you very much. If I encounter any bugs, I will open new issue.

[1pang2yik]

hey i have some trouble

[1pang2yik]

Error at doLogin(): (<class 'mechanize._response.httperror_seek_wrapper'>, <httperror_seek_wrapper (urllib2.HTTPError instance) at 0x33732d0 whose wrapped object = <closeable_response at 0x33b2bc0 whose fp = <response_seek_wrapper at 0x33b22d8 whose wrapped object = <closeable_response at 0x33b2670 whose fp = <socket._fileobject object at 0x033A7DF0>>>>>, <traceback object at 0x03361A58>)
Cannot Login!

[1pang2yik]

and if i remove cookies from config.ini it will show

Server reply: {u'captcha': u'Complete the reCAPTCHA verification'}
press enter to exit.

[Nandaka]

see #505