Should be fixed now if you update to v1.2.1! I just made it skip the messed up documents, from looking into it those should contain no text anyway.
EDIT: I made a thread to properly announce the program, any further discussion about it should definitely be moved to that thread: Vocabulary Extractor: Make your own Decks from Manga/Anime/EBooks
Yes ! it works great ! Thank you once more !!!
When you import decks, all missing vocab is thrown to the bottom of the deck, which is great for fixing them all in one place. but after you fix them and replace them with what youre looking for they stay at the bottom of the deck. which messes with the chronological order as they appear in the book. trying to reorder them where they should go takes too much time, to both find the location it should have been in and draging it up through 500 items
is there a way to maintain its original location in the import after the missing vocab error is fixed?
Would it be an idea to have an optional field for a custom study sentence in the import process? It would allow the user to preserve the context of where the word came from.
It might be awkward to fit it in conceptually between shareable custom decks and personal custom sentences, but I thought it’d be an interesting feature.
Are we able to add grammar items to custom decks or just vocab?
You can add Grammar, just not through the importer.
Ah that makes sense, thanks (. ❛ ᴗ ❛.)
Any idea when maintence will be done on importing decks? I cooked up some Ace Attorney csvs I cant wait to throw in the deck pool
I did get half a chapter of harry potter done before the maintence, which had to be broken into parts because of the 500 unit limit. I can see this number being a lil low for book chapters or anime episodes. Dont suppose increasing the number of vocab allowed in units is part of the maintenence 
Probably back on the live site without any changes tomorrow.
We thought it was the thing causing the server issues we were having, but it wasn’t.
Was working on adding the “Add to Deck” button for the individual Grammar/Vocab pages + on the Search results page.
So that should be out tomorrow at least.
Made a not of the other stuff , but just wanted to mention:
The parser only searches for Vocab, not Grammar!
Hence the ぞ stuff not working
If I accidentally delete something right now it’s gone forever
Probs not gonna implement saving of the not-found items, but I will add a Confirmation model before allowing the deletion of items.
Do you have the text that generated this error?
If so I will try to reproduce and fix it before re-publishing the importer.
I assume it’s the process of what I was doing rather than the list specifically so I’ll describe that in detail as well, but here you go: A Silent Voice Vocab · GitHub
What I did was I manually copy-pasted around 250 words each time from the file, and imported each batch separately. I didn’t rename any of the units at first so I ended up with 5 units with the same name in the import page. As I tried to rename the first one, all the ones below it duplicated and turned into the frozen state. The duplicate ones were inserted between the edited unit and the other 4 that still work at the bottom, if I remember correctly. If I now edited the next editable unit (the 2nd one, now at position 6 i believe), the same thing happened with the units below that one, and so on until I finished all of them. Only editing the last one didn’t cause any problems.
Also, another thing I’d like to add: If I remember correctly, any time I tried to import those 250 items, it lost a couple of them even though they were not shown as a fixable mistake. I ended up with a couple words missing from every Unit even though the JMDict IDs should all be entirely unique. Maybe it’s very similar words being consolidated? Or counters being combined into just the standalone counter vocab?
Wasn’t sure to mention it here or suggestions. But when importing csv (non id numbers), the wrong word will sometimes be chosen for kana words or even single character kanji. Which is understandable of course. But, sometimes I’ll miss that it did this until after the import. Which to fix, you have to delete>add correct vocab>move into chronological spot
Can there be a way to simply edit/exchange vocab and replace it with the correct vocab exactly where it is in the unit lineup? Being able to pull up the vocab chooser you see for missing items on import anytime during deck edit would be really useful
Also big thanks to the adjustment of how it handles missing vocab on import😭 my missed vocab doesn’t sit at the bottom of the unit, but actually where it belongs thank youuuuu
Just pushed a bunch of fixes for issues that have recently been mentioned.
Also switched to “batching” system for both searching for items and saving content.
This should fix those mass-import timeouts.
Tomorrow I will remove the Unit item (currently is 500) and total item (currently 2000) limits.
I just tried this out with a bunch more anime and it works great! That ghost unit bug when renaming units still exists though, sadly. And of course it’ll be amazing once the limits are gone. Thank you so much for all the hard work!
@eefara @Magyarapointe @haldo @entropyofchaos
I just published the Import tool again, this time without total count/Unit limits!
EDIT: Sadly had to re-introduce the 500 item Unit limit.
@Flutter
I can’t replicate the "ghost unit"bug (anymore?).
Are you able to do me a huge favor and try reproduce it for me now plz? 
Hey! Sure, here you have a video of the bug: https://youtu.be/kBNaETnPaOI
I’m using Librewolf so maybe this is a Firefox thing.
Also, is the Unit limit reinstated temporarily or is it going to stay for now? I found multiple anime in which the first episode exceeds 500 items so I think it’s rather important personally. Thanks!
For novels, using @Flutter さん 's wonderful tools after segmenting epub into chapters, and using the vocab_combined file, the 500-item unit limit at least takes care of itself, I simply renamed overflow units into part 1/part 2, etc… No 500 item limit for units would be even better though…
@veritas_nz I found a bug : when importing with JMDICT ids, sometimes a word won’t be found. Previously, the ID number showed up in the red rectangle marking the missed word. Now, the importer tells me a word is missing, I get the red rectangle, but it’s completely empty : there’s no way of knowing which word failed.
Also, if I create units first, use the import tool to populate them, and there’s an overflow unit because of the 500-item limit, I can edit the title of the overflow unit, but not the title of the original one. If I use the import tool first and use #whatever to mark units, I can edit the titles of both the main unit, and the overflow; however, to save the edited title of the main unit, I first HAVE to add something in the Unit description sub-level (even if I erase it right away). If I don’t add anything there, it won’t save edits to the unit name. This doesn’t happen with overflow units, presumably because Bunpro automatically adds a unit description.
I don’t have anything useful to contribute, but I just wanted to say that I’m super excited for this feature. I’ve been wanting to see more decks on the site and now with the prospect of importing jpdb decks into bunpro I can’t wait to use this feature. I have been using games like Ace Attorney to help with my reading skills and this integration will really improve my studying experience since those vocab lists already exist elsewhere. Thank you @veritas_nz for your responsiveness and work with the community, I’ve read this thread and I’ve seen a lot of hard work done. Thanks additionally to all of the community members who have helped guide this feature as it’s being tested. I’ve already added a few community decks to my queue and I can’t wait to add more.
when importing with JMDICT ids, sometimes a word won’t be found. Previously, the ID number showed up in the red rectangle marking the missed word. Now, the importer tells me a word is missing, I get the red rectangle, but it’s completely empty : there’s no way of knowing which word failed.
I have fixed this now.
Thanks for the report!
Also, if I create units first, use the import tool to populate them, and there’s an overflow unit because of the 500-item limit, I can edit the title of the overflow unit, but not the title of the original one. If I use the import tool first and use #whatever to mark units, I can edit the titles of both the main unit, and the overflow; however, to save the edited title of the main unit, I first HAVE to add something in the Unit description sub-level (even if I erase it right away). If I don’t add anything there, it won’t save edits to the unit name. This doesn’t happen with overflow units, presumably because Bunpro automatically adds a unit description.
Sorry to be a PITA, but are you able to make break this into steps please? With the naming of the Units too maybe?
I wasn’t able to replicate this starting from an empty Deck. 
I know there could potentially be some funkiness with duplicate names etc, but I thought that got fixed in my last update 