Calibre Server: 2 Feature Requests/Ideas... "tag cloud"+"word cloud"

[wjBLqf-_9]

It would be something awesome if Calibre's web server had the ability to let people "drill down" by selecting tags or words in book

Feature Request/Idea #1 of 2 --> "tag cloud"
It would be something awesome if Calibre's web server had the ability to let people "drill down" by selecting tags or words in book

https://www.smashingmagazine.com/200...d-practices-2/

Like:
tier 1> Click on Tag "Earth"
tier 2> Click on Tag "North America"
tier 3> Click on Tag "California"
tier 4> Click on Tag "SanFrancisco"
... to see all library books with the tags: Earth, North America, California, SanFrancisco


Feature Request/Idea #2 of 2 --> "word cloud"
The *MUCH* bigger version of Feature Request/Idea #1

Totally identical

Except it would be a cloud representation of the "X" most used words in the library of books

same 'drill down" ability by selecting words you would piece together... to discover book content you never realized you had



FWIW: Tags are good, but they are not enough.

The closest I can come to approximating this is using the plugin "ENF" and generating a word list of each book and display in the comments.


I would even offer to "buy several pints" of beer for the developer who implements either idea... hopefully idea #2

Is the DB engine robust enough? Would something beefier be neeed?

[kovidgoyal]

You can already do (1) simply click the search button and then click on the tags/authors/publishers/custom columns/ etc you want to browse by. It works just like the Tag browser in the main calibre program.

[BetterRed]

You could use a a content search utility, like the one that comes with Windows File Explorer, to search your books for words and terms of interest. Drop the search results into the Drop Search Results plugin, which will Mark the books. Then add the word/term to a tag-like column (My Tag Cloud) for Marked books

If you use File Explorer Search you can save the searches (blah-blah.search-ms) as 'books', then you can easily run them again to find the words/terms in newly acquired books.

You'll probably have to convert books into a format that the content indexer can handle, e.g. DOCX or TXT.

I do content searches of 100,000+ books in < 5secs. It can take longer to add the 'tag' to the Marked books than it takes to find them

Note I'm referring to File Explorer Search not Start Menu/Taskbar/Cortana search. You may have to enable the indexer and configure it via Control Panel.

BR

[wjBLqf-_9]

Friends don't let friends do "File Explorer Search"
...or better put... "once you go 'FileLocator Pro' you never go back!"

BEST $49 ever spent!!!
---> https://www.mythicsoft.com/filelocatorpro/ <---

IMO: Explorer Search is a nasty tool (Win10 has improved it, almost adequately)

Proper index search :-)

[wjBLqf-_9]

But even FileLocator Pro is a rough fit for my idea.

I will gladly put up $50 bounty if someone implements my "word cloud" globally.

I know your coding skills are worth more than my pocket book.

(I am not rich and I could never afford to "hire" you as my private full time software developer all-for-myself... if someone wanted to raise the bounty, let them PM me and I'll negotiate)

[BetterRed]

@wjBLqf-_9 - and how was I supposed to intuit you had a payware search tool.

If you had bothered to read the Drop Search Results thread then you would have seen I tested it with File Explorer Search, FileLocator Pro, X1, and xplorer², and the author tested it with a couple of others.

So, my imperfect solution will work with any content search tool that conforms to the relevant MS APIs, including the Mythicsoft products.

“Don't let the perfect be the enemy of the good.” - Anonymous (not Voltaire)

BR

[wjBLqf-_9]

"BetterRed" email/forum is a poor medium to read between lines. I hope you 'read' my post as humorous :-)

I never considered myself any definitive authority in anything, especially not calibre matters.

I was attempting to express a limited bit of levity in my reply. In my day job I'm a super-kinda-analytical-often-too-literal telecom engineer... and following self assessment of my whole person, I have selected "ability to make people chuckle with my jokes" as one area of my life to improve :-)

My humor also expressed my personal statement of opinion hence my posting stated "IMO" (In My Opinion).

Regarding topic at hand: no argument from. I believe you. I an certain you have a lot more experience in this than me.

*shrug* I like what I found works for me :-)

[wjBLqf-_9]

edit: *shrug* I like what I found to fill-the-void [well enough for now] to mostly work for me

maybe I'm a goofball for that "word-tag-cloud" thingy idea

hopefully I'm a harmless friendly goofball :->

[DaltonST]

Originally Posted by wjBLqf-_9

maybe I'm a goofball for that "word-tag-cloud" thingy idea

No, you are not. CalibreSpy can easily do what you described because it is designed primarily for "filtering", not "searching".

For Tags within an intrinsic (informal) "hierarchy", the CalibreSpy plug-in can filter down from the "top" of the hierarchy to its bottom.

Filter-down example: https://www.mobileread.com/forums/sh...0&postcount=25


CalibreSpy Homepage: https://www.mobileread.com/forums/sh...d.php?t=310103


The latest version as of this post is: 1.0.28.





DaltonST

[wjBLqf-_9]

DaltonST: thank you! Your posting about "CalibreSpy" proves how filter-down-tags-hierarchically is currently possible. You are absolutely 100% correct! I sincerely appreciate your post! Please accept my sincerely gratitude for sharing your information with me!

I would describe CalibreSpy search like this as a 'technical capability'

I would have no doubt technically minded people would certainly become agile/skilled at the necessary syntax to perform this search for themselves.

For the BIG PICTURE comparison............... let us compare the current version of whatever your personally favorite chosen web browser....... to the first browser I ever used to "surf" the internet: LYNX via VT-100 terminal session in MS-DOS v3.3

LYNX was a legit true web browser!
6 months later I was using Trumpet Winsock's SLIP/PPP to browse the web in Netscape Navigator v2

My point: technically capable (LYNX) is worlds from a user friendly system (Netscape)

I remain optimistic that someday I will joyfully payout my aforementioned $50 bounty for a solution that retained enough capability in trade for less technical skills required for use :-)

[Terisa de morgan]

Originally Posted by wjBLqf-_9

I remain optimistic that someday I will joyfully payout my aforementioned $50 bounty for a solution that retained enough capability in trade for less technical skills required for use :-)

Those "mixes" don't usually end well. Big capabilities usually require user's skills (I've seen at work a lot of those tries, and I have seen none working fine).

[DaltonST]

Originally Posted by wjBLqf-_9

I would have no doubt technically minded people would certainly become agile/skilled at the necessary syntax to perform this search for themselves.

Syntax? What syntax?

I have no clue what you mean by "necessary syntax". CalibreSpy has no "syntax".

CalibreSpy is "mouse point-and-click" driven.


Selecting your Calibre Tags from a dropdown list with a mouse in Calibrespy is not "syntax".

The example shown at https://www.mobileread.com/forums/at...1&d=1537301810 shows that CalibreSpy just allows you to filter any Calibre Tags (standard or custom column) that you wish via the mouse.




DaltonST

[wjBLqf-_9]

Update to this thread :-)

Having funneled new text from widely diverse sources into my Calibre Library... approaching 125k now.

I find the fault in "manually-administrated/moderated" tags... or perhaps upper limit to group-standard-universal-compliance (wrt 'scale-ability')

Authors/Publishers who obey common-compliant tags -- are like overlapping clusters. Ven Diagrams of overlapping ovals of familial but different tag nomenclature that eventually circumnavigate the globe.

Since I am just one human, I cannot re-write the tags for an entire world to force compliance to my view..... seems to me it would be a neato application for dynamically-generated "word-tag-cloud" filter system.

At risk of abusing cliche hip-with-the-times term: perhaps this is an interesting application of some form of ML or AI to learn how to quickly synthesize the appropriate translated tag between clusters -- or create a new tag to correlate similarities between compatibile but different cluster obeyed tags.

hehehehe :->

[theducks]

Originally Posted by wjBLqf-_9

Update to this thread :-)

Having funneled new text from widely diverse sources into my Calibre Library... approaching 125k now.

I find the fault in "manually-administrated/moderated" tags... or perhaps upper limit to group-standard-universal-compliance (wrt 'scale-ability')

Authors/Publishers who obey common-compliant tags -- are like overlapping clusters. Ven Diagrams of overlapping ovals of familial but different tag nomenclature that eventually circumnavigate the globe.

Since I am just one human, I cannot re-write the tags for an entire world to force compliance to my view..... seems to me it would be a neato application for dynamically-generated "word-tag-cloud" filter system.

At risk of abusing cliche hip-with-the-times term: perhaps this is an interesting application of some form of ML or AI to learn how to quickly synthesize the appropriate translated tag between clusters -- or create a new tag to correlate similarities between compatibile but different cluster obeyed tags.

hehehehe :->

Calibre has a Tag Mapper. You define each map.
I reduce the 42 ways folk tag Science Fiction to that tag