@Keywords Auto-Group Sort

Started by StanRohrer, December 11, 2014, 01:26:13 AM

Previous topic - Next topic

StanRohrer

IM 5.2.16

My Request:
Left panel / Categories / @Keywords / Property Panel I can turn on Auto-Group and use 1,1 as the grouping. I get 52 groups including lower case "a" and upper case "A". I need a sort that combines upper "A" and lower case "a" into one group. This would result in 26 groups - 1 for each letter, not case sensitive.

Currently we have @Keywords Auto-Grouped:
a
A
b
B
c
C

I want @Keywords:
a (containing a, A)
b
c

or:
A (containing a, A)
B
C

I think what I want is a sort order like that of a dictionary.

My problem:
Over 10+ years of using iMatch I know I have not been consistent with capitalization. So I would like "White House" and "white house" to be next to each other in the keywords sort under Auto-Group "W". Similarly "Hotrod", "HotRod", and "hotrod" under "H".

What Does Not Work Well:
It is possible to drag all of one category (say "a") to another (such as "A"). However this reverts back to the original tree on the next refresh. With over 10,000 keywords, the drag and drop for each letter is very time consuming. It seemed to work on a small test database but hung in my full database.

It is possible for some people to find "White House" and combine with "white house" into the same category. I understand this would modify the files to match. However, I have 110,000 offline files on CD/DVD's so updates can't happen.

Previous discussion here:
https://www.photools.com/community/index.php?topic=3892.0

Carlo Didier

+1 (I'm strictly against all forms of case-sensitivity, except in passwords)

ColinIM

+1

Not so much from a philosophical stance in my case ... it's just that my work is heavily based around keywords, so when I want to correlate and verify a set of images with their intended keywords (by clicking on successive Keywords on the Categories panel) this separation of 'a' and 'A' keywords means that I must do a "double take" each time. 

I note that on the Filter Panel, where I can choose to enable the Categories sub-panel and then (possibly) include Keywords in a filter definition, the list of Keywords there has (logically) the same upper-case / lower-case separation, so that 'Filter' route isn't necessarily better for me.

Colin P.

Ger

-1

Maybe I'm too much of a purist in this. The current setup (keywords, synonyms, all organized in a very flexible thesaurus) allows me to easily maintain a unambiguous structure, where 'White House' is not the same as 'white house', although the 'White House' actually is a 'white house'.

Ger

StanRohrer

Ger,
With this request, I DO NOT want to combine "White House" and "white house". I want them next to each other in the sorted list so I can decide if it is OK or if they need further fixed. If one is under "W" auto-group and the other is under "w", I may never consider that both exist when on a fast search for something. Like ColinIM, I don't like to have to do a double take to check the possibilities. When working in the filter panel it just takes double the hunting to be sure the base is covered. I expect the filter panel to work the same as @Keywords. I think a "dictionary sort" ignoring (or more likely combining) case would be very useful to me.

Now a question that Mario would likely ask next - should this be for all categories listed (as today), or should this be a sort option for only the @Keywords category? I certainly want the @Keywords as a dictionary sort. As for the rest - I've not decided - I think I could live with them also being a dictionary sort. Most of my non-@Keywords categories have been prefixed with a number (star date, 20141211, or sequence, 04) so most sorts keep the same order for me.

Mario

I showed you how to sort W next to w in the category view. See the screen shot in the other thread.

We had keyword unification (A = a) in the early beta but dropped it for good reasons after introducing the advanced features for @Keywords, like making changing in @Keywords is updating the real keywords in the file, moving files within @Keywords also changes the corresponding keywords in the file and the etc. If @Keywords would not reflect the true spelling of keywords, utilizing it to update the actual keywords in the file would not be possible, or would change the spelling of the keywords in the file in probably unwanted ways. Thats why we dropped all the unify functionality (which is still available in standard (read-only) data-driven categories).
-- Mario
IMatch Developer
Forum Administrator
http://www.photools.com  -  Contact & Support - Follow me on 𝕏 - Like photools.com on Facebook

StanRohrer

#6
Mario, this request is NOT to get w and W next to each other. Yes, that works. I am asking to provide a sort that (in your words) folds w and W entries together into one single (wW) group. I am NOT asking to convert one keyword to another with a case change (thus changing the files).

Here is a real example from my database. I have a group c with 720 entries and group C with 470 entries. Flipping back and forth to find dictionary adjacent entries is not useful when I think there has to be a software sort that could take the two lists, make one, and put them in order. Please give us a sort for cC that would include all 1190 entries (c 720 + C 470). This would put "cheerios" and "Cheerios" with one directly after the other in the one "cC" result list.

cheer
cheerful
Cheerful           <- should be lower case based on photos of people
cheerios          <- wrong lower case needs fixed
Cheerios          <- the breakfast cereal food name, correct name
cheese steak
cheesecake      <- generic name for a food, links to a restaurant should be fixed
Cheesecake      <- maybe should be Cheesecake Factory, a restaurant chain name in the USA
cheesehouse    <- name of store. likely should be fixed to Cheese House
Cheetah           <- should be upper or lower case? I'm not an English major so my database is confusing.
chef

With the current sort choices separating w from W I will not go through my categories as it is way to much flipping back and forth for each term to compare. Similarly, I may look for Cheerios and not think about the possibility that my database also contains cheerios. A dictionary sort of @Keywords can cut every @Keyword category access in half for us. This has to work with @Keywords Auto-Group enabled.

With your programming skills I'm sure you can make this function work. Albeit, the Auto-Group parameters will certainly add a fun challenge. Now that I'm here - I think the fix I'm looking for is actually a function of how Auto-Group works and sorts. Perhaps this is not a function of the Categories Gear sort selections, but a sort function of @Keywords Auto-Group.  What do you think?