- Fix title, tags, and caption parsing.
- Add new filename format for download by tags:
- %searchTags% ==> for download by tags, add searched tag keyword.
Download link for pixiv downloader 20121215, source code in GitHub.
72 thoughts on “pixiv downloader 20121215”
I think Pixiv may have changed their layout again. All was running fine when all of a sudden I started to get this error message.
‘NoneType’ object has no attribute ‘find’
1 2 3 4
I installed the latest version of Pixiv, and just to be safe, deleted the config.ini file. Didn’t make any difference…I still kept getting the above error.
Sorry; I’ve been away for a bit. At the time, when I downloaded the latest version, that one wasn’t available yet. I’ll give it a try, and if there are any problems, I’ll reply in that thread. If not…then all is well.
I seem to be having trouble downloading from the “IbXXXusers入り” tags. It ends up download Tiger & Bunny stuff instead.
Thanks for the awesome program by the way.
It looks like the problem is that when I use the download by tags list feature, it doesn’t search the first letter of what I want. So when I look for “IbXXXusers入り”, it just searches for “bXXXusers入り”.
do you have the log file? what is the generated url from the log file?
Hello, thanks for all your work!
I didn’t updated the program for a while, today I tried to use this version and download by tags as usual using URL Encoder (http://meyerweb.com/eric/tools/dencoder/), pasting the endoded tag and using Wildcards = y with rest default. But now the program gives me this error. Trying to the older one gave me no image saved, though it always worked fine using Encoded tags.
Try to set UseRobot = False in config.ini
Thanks, it remove the error, but it doesn’t download nothing still. “No more images” (Trying http://www.pixiv.net/search.php?s_mode=s_tag&word=%E3%83%A9%E3%83%BC%E3%83%A1%E3%83%B3%E5%AD%90)
I also tried to download by user id, but it loops with “NoneType has no attribute “find”
Try the latest version here: http://nandaka.wordpress.com/2013/01/28/pixiv-downloader-20130128/
I love the program. I’m just having a problem when I download by tags. If I search by a particular tag and leave everything else at defaults (bookmark count, starting page, date, etc.) it works fine. If I try to add a “bookmark count” number (e.g. 100) than I’ll only download an image or two even when there are thousands of images with at least 100 bookmarks.
sample query and log file please.
http://pastebin.com/XhPEsLCi — Log file
query: “trap” (no quotes)
Looks correct to me, the (-1) value is because there is no bookmark counts on the search result page (the blue star with number)
Thanks for looking into it. I hate to sound silly but how do I download a tag with at least 100 bookmarks then?
Try to download with start page = 1, page limit = 0, but set the bookmark count = 100. It will try to list the images (I think up to 100 page) and only download the images with 100++ bookmark count. I don’t think there is a way from Pixiv to list images with specific bookmark count.
I figured out what I was doing wrong. 100% my fault. Thanks for the help.
Ok, now there’s another strange bug I just witnessed. The last page of an artist’s profile (specifically http://www.pixiv.net/member_illust.php?id=2637975&p=5) feature ssome mangas; however, the program detected them as being made of a single page and downloaded only “p0” of each of them, Oddly, previous mangas in the same profile have been downloaded correctly. What may have happened? It’s the first time I witness such an error and I hope to not find it has happened already.
(For precision, it’s on the laptop with Win8)
Some of the manga on the page you give is only have 1p, you can verify it by checking it on the pixiv page itself (see near the title, e.g.: 03/02/2011 03:54 Manga 1P ComicStudio )
whoops, my bad then. Never stumbled in a single-page manga before, I thought that mode only from 2 images and up °__°
I hadn’t updated the program for a while (older versions were still fine) and now I’m starting to use this version in Win8. I opened two instances and set one to download from page 1 to 2 of the member_id, and the other from 3 to 4. However, they both stopped after a single page. I then tried only one instance to do from 1 to 3, so after checking page 1 were all duplicates, it downloaded 2 and stopped at the end of it.
Is the end limit now changed to “before page x”? It’s confusing.
Confirmed that, most likely I forget to add = on the page checking 😀
something is apparently off even at page starting, I’m trying with two instances again and even with limits well set for each one, after the first instance reached page 2 it told there were duplicate files…@__@ Since the only functions I use are member_id download and bookmarks download (a differently configured instance, that works without problems though), maybe it’s better if I downgrade to a previous version for now? Or will you fix this soon?
You are using 2 instance from different folder, right? If not, then Can you upload the log and config file for both instances? See whether the generated member’s url is correct or not.
Nandaka, I appreciate your work, the command is clean and easy to use,
I’ve sent you an email in the contact you left for donations,
please give it a look.
Hi, the link you gave me (http://blog.sina.com.cn) is a blog hosting site like wordpress. It is not a imageboard site like pixiv or danbooru.
Each user can have different css styling, so it will be hard to parse. Your best bet is using website crawler to get all the images (HTTrack: http://www.httrack.com/), which basically copying the site to your drive.
Thank you for your quick Answer, I’ll have a look at it.
by the way, did you like the links?
Ok, I guess 😀 I prefer the less ‘oil painted’ style, e.g. Ishikei 😛
Is there an escape key when downloading, so it would finish downloading the image it’s on, then abort? It looks like if you ctrl+c, the image will not finish downloading, but it’s already written in the SQL file, and will not be redownloaded.
Could you add a feature where we can download from pixiv daily ranking? XD
You saved my wrist from incessant right clicking. Thank you SO MUCH!
Could you please make a tutorial on how to set up this because I keep getting the “Nonetype” object has no attribute ‘ul’ error
The default config is enough for running. That error is usually because pixiv is changing their layout, can you tell me what are the step to replicate the error?
I’m using proxy to connect to the internet, and that proxy needs username and password, where in the config.ini I can set that parameter?
See on readme: http://[username]:[password]@[proxy_server]:[port]
If your password or username contain : or @, replace it with %3A or %40
Okay, thanks 😀
seems to be an error…
2013-01-03 12:12:36,757 – PixivUtil20121215 – INFO – Processing Member Id: 142558
2013-01-03 12:12:36,792 – PixivUtil20121215 – INFO – Member Url: http://www.pixiv.net/member_illust.php?id=142558&p=1
2013-01-03 12:12:39,193 – PixivUtil20121215 – INFO – Member ID (142558): ‘Member Error: xe6x8axb1xe6xadx89xefxbcx8cxe8xbfx99xe4xb8xaaxe7x94xa8xe6x88xb7xe7x9ax84xe5xb8x90xe5x8fxb7xe5xb7xb2xe7xbbx8fxe8xbfx90xe8x90xa5xe5x81x9cxe6xadxa2xe4xbdxbfxe7x94xa8xe4xbax86xe3x80x82’
Checked the dump file, are you using Chinese language? This is because I didn’t take count for error message other than Japanese and English.
Based on google translate: 抱歉，这个用户的帐号已经运营停止使用了 ==> Sorry, this user’s account has been in operation to stop using the… , basically the user is not in pixiv already.
Checked the english message: This user account has been suspended.
Yes,I’m using Chinese language in PIXIV.
I saw “Error Code 1003”,I thought it was a serious problem,so…
Actually, this also happens with .bats. I mentioned AutoIt just because that’s what I use. And what has working directory got to do with copying config.ini (I know it’s copied because all the settings are kept the same)?
usually I place the .bat file in the same folder with the application, or you can try to “cd” to the application folder before calling the application in the batch file.
Yep, that indeed helped. But why does this happen?
The application will try to read config.ini from the current active folder, if not found, it will recreate the config.ini by using the default value.
By the way, why does this downloader find duplicates much slower than Danbooru Downloader? (one image per 2-4 seconds versus 10+ images per second)
try set processfromdb = True and overwrite = False in config.ini. Most likely it is because of the duplicate check is done not from db (get the page info from internet and check the filename, compared by checking the downloaded image_id from db)
Both are already set this way.
can you upload the config.ini and the log file to pastebin?
Config file is unchanged from the above, log is here: http://pastebin.com/SsQMNnyG
Note how the last entry (which was pure duplicate checking) took five minutes to check 117 images for duplicates. Third entry from the end (the same artist) was even slower. Fourth entry was the actual download and it took 14 minutes, which is still kinda slow, but still much better.
Set alwayscheckfilesize = False in config.ini
Yaaay. And one last thing. Every time the downloader finishes running, config.ini gets copied to the directory where my AutoIt script that runs it is located. How can I prevent it from doing so?
Dunno about that. Never use Autolt script before. Maybe you need to change the working directory before running the script?
The default: http://pastebin.com/BQ30xxhG . I tried changing pretty much every parameter to True or False, but nothing helped. 🙁
Try to set the rootdirectory to your save folder, e.g. S:ScriptsPixiv Downloader and keep the filenameformat/ filenamemangaformat as it is. The problem is you have double drive path when the application triednto generate the filename (see the error message, it resolve to ‘S:ScriptsPixiv Downloader”S:’
Yep, that helped. Apparently, putting quotes around the path was causing this glitch. However, now I am getting the following error: prntscr.com/n8tdz . The artist folder gets created where it’s supposed to be, but all pictures are downloaded in the script directory (I assume this is some kind of fallback algorithm).
Why the filenameformat got inverted triangle there?
Huh. I deleted ‘e_’ and typed it again and everything started working perfectly. Many thanks for the help and this awesome program, Nandaka. 😉
While I haven’t encountered this yet that I know of, I was wondering something:
If I download by member ID, and choose to download by that same member ID later after the member has changed their icon, what happens to the folder.jpg? Is it overwritten, or is a new folder with folder.jpg generated?
Assuming the save path doesn’t change, then it will overwrite the folder.jpg
For some reason I’m getting double slashes instead of single ones: http://prntscr.com/n5vke
This only happens if I set rootdirectory to be something other than ‘.’.
Also, even when I am downloading in the same directory, an empty folder gets created and everything else gets downloaded outside of it. What can I do? 🙁
What is your config.ini setting?
I have downloaded pictures, how do I update them with their tags?
no can do :(.
Downloaded images is cannot be synced with the tags, either you download the images again or you copy the tags and update the filename by yourself.
Do not you can set the default bookmark count config.ini?
I want to work bookmark count when Download by taglist.
No feature for that, sorry.
Found the problem. alwayscheckfilesize = True. Silly me…
Anyways thanx Nandaka, you and your work are awesome!
Noticed that the program writes in the db.sqlite but can’t seem read it (instead of skipping the already downloaded file, always says “file already exist(same size)”
have you check
Yup, it is set on true and it doesn’t work. It creates the db with no prob, The db.sqlite grows larger as long as I download but when I go back and download things I already downloaded (by memberid for example) instead of skipping the already downloaded ones it checks every picture “file already exist(same size)”.
Bless you based nandaka!
Hmmm, just noticed something… when using the command line, the program will not automatically decode the tags?
Update the source code. I’ll upload the binary later on 😀
Comments are closed.