[GUI Plugin] Power (Full-text) Search

[motherat]

Hi mapozyan -- thank you for replying.

It's completely possible that my use case is way too niche to be worth agonising over. I'll give a quick rundown anyway, just in case:

The multiline search(es) I perform include HTML tags. For instance, there's a common formatting error on AO3 which places each <em> tag in its own separate line, resulting in faulty-looking punctuation around italicised text. If I could search my library for such occurences, I'd be able look into fixing the affected files without checking all of my epubs by hand, one by one.

Unless Power Search does not look for HTML tags at all? I thought it did, but if that's not the case, please disregard my request.

[mapozyan]

Hi motherat,

Indeed this can be an interesting feature.

But you guessed right, Power Search currently does not respect any HTML tags and works with plaintext content only. While supporting HTML is possible in principle, it is not an easy task. Can't promise that you'll see it soon.

[Rellwood]

Ok, it looks like I might be over my head here, but is there some way of getting the elastic search without having to purchase a monthly subscription? I love the idea of being able to search within my books - especially to find the ones that have editing errors that need fixing, but it seems like this plugin requires more than I thought.

Am I mistaken?

I just tried the "free trial" and it's creating a deployment - whatever that means...

[mapozyan]

Hi Rellwood and thanks for your comment.

Elasticsearch is free for self-managed usage. It just turned out that the webpage that I was referring to has changed.
Here you can download the latest version for your operating system.
https://www.elastic.co/downloads/elasticsearch
If you are using Windows I would recommend trying "MSI (BETA)" installation, its easier than getting it up and running from ZIP file.

Hope that helps!

[Rellwood]

Originally Posted by mapozyan

Hi Rellwood and thanks for your comment.

Elasticsearch is free for self-managed usage. It just turned out that the webpage that I was referring to has changed.
Here you can download the latest version for your operating system.
https://www.elastic.co/downloads/elasticsearch
If you are using Windows I would recommend trying "MSI (BETA)" installation, its easier than getting it up and running from ZIP file.

Hope that helps!

Thanks! I'll let you know

[ownedbycats]

Small request: For column icon reasons, could you modify the marks that Power Search creates so it's marked:search_results or something similar? I think you can use marked_text for this; Quality Check and FanFicFare both use them if you want to see actual examples.

Thank you

[mapozyan]

Version 2.1.1 released!

What's changed:
- Mark search results with "marked:search_results"
- Fix occasional error message on first ElasticSearch start

Thanks to ownedbycats for great suggestion as usual.

[Rellwood]

Hi!

Just to let you know, once I downloaded the free version and set it to run automatically in the background instead of setting it to manual start when I installed elastic search, the plugin works perfectly.

Now, I wanted to ask if there is some kind of way to indicate if a book has been indexed, also, if there is a way to select specific books to index?

I see the index all new books, but when I hit that I hear my computer start working but I don't know which books were done. I have taken to doing a power search on the word "The" since I know it's in all my books. I found about 900 books that didn't match the search. So, instead of being able to just select the books and index them, I had to reindex all books. But when they are done, how do I know they have been indexed?

Is there a way of either generating a report that includes the titles indexed or a way of marking the books or creating a tag?

Other than that it's great. I just wish I knew how to apply this elastic search program to the rest of my computer. I am using Agent Ransack right now for that.

[DNSB]

Originally Posted by Rellwood

Now, I wanted to ask if there is some kind of way to indicate if a book has been indexed, also, if there is a way to select specific books to index?

I see the index all new books, but when I hit that I hear my computer start working but I don't know which books were done. I have taken to doing a power search on the word "The" since I know it's in all my books. I found about 900 books that didn't match the search. So, instead of being able to just select the books and index them, I had to reindex all books. But when they are done, how do I know they have been indexed?

Basically, when you start Power Search, it will index all new books in that library which translates as the books that are not listed as indexed in that library. If you click on the arrow beside Details, you can watch list of books being indexed. Unless you have done some mucking around in the OS which is "NOT A GOOD IDEA", once you've run Power Search on your library (or each of your libraries if you have multiple libraries), all the books will be indexed. Have you run calibre's Library maintenance => Check library function recently?

Edit: One other thing to remember is that Power Search will not return more than 9999 search results.

[Rellwood]

Originally Posted by DNSB

Basically, when you start Power Search, it will index all new books in that library which translates as the books that are not listed as indexed in that library. If you click on the arrow beside Details, you can watch list of books being indexed. Unless you have done some mucking around in the OS which is "NOT A GOOD IDEA", once you've run Power Search on your library (or each of your libraries if you have multiple libraries), all the books will be indexed. Have you run calibre's Library maintenance => Check library function recently?

Edit: One other thing to remember is that Power Search will not return more than 9999 search results.

Thanks for the info. I haven't been messing with the OS, and I am not searching more than 5000 books. I have indexed the library in whole and have seen the details while it was running. I don't quite know what happened, but I redid another search just using "a" because I guess there are books that don't have "The" in them (actually I noticed a couple of cookbooks listed) but after doing the last search I was left with 300 or so books that hadn't been indexed. So, I copied them to a blank library and reindexed them. I realized that all of these books had an "original_epub" or "original_mobi" file in them, and the one book that didn't index actually was corrupted and I needed to replace it. So, it seems that book files that have "original" in them don't index. When I deleted them (those specific files, but left the edited versions) the search ran much faster and all of them ended up indexing. I copied them back over to my current "All" library and reindexed them as new. Now it's fine.

Soooo, the lesson here is to make sure that you aren't trying to index any files that aren't listed in the powersearch.

As for marking the books, I just create a tag when it's done and that's that!

[mapozyan]

That's interesting! I will try to figure out what this "original" is.

[ownedbycats]

Use polish ebook to embed a cover or metadata or something. Unless you've configured it not to, it'll create an ORIGINAL_EPUB file.

[DNSB]

Polish will do that for Kindle format ebooks as well. The Modify Epub plugin will also create .original_epub backups of the unmodified original epub.

[jbacelar]

I have created 150 user categories (nationalities), corresponding to the 8000 authors that I have in the library.
Before this, when indexing new books, Power Search immediately started indexing them.
Now before starting work, it remains static (with high CPU consumption) for about three minutes, then indexes them with the same joy as always.
In other libraries that do not have user categories defined, Power Search works perfectly.
Will it be possible to solve this problem?

[ownedbycats]

It's kinda painful to do, but what happens if you delete the index and then re-create it?