bulk keyword tool problem?

pilotdave

Trying to use the bulk keyword tool and I'm running into some unusual trouble. I try to add, for example, "joe b" to selected keywords. I end up with: joe b,. I don't want the comma, and I do want the quotes.

Also, the preview button no longer seems to work in IE6. I realize this is not a supported browser, but it has always worked before (and I'm stuck with IE6 for 8 hours a day).

Guessing this has to do with the recent change to allow foreign characters in keywords. Is this just the way it works now, or is this a bug?

Dave

Andy

go add it to two photos, just to see, the , [comma] will show in the tool but not be part of you keywords.. we don't need the quotes anymore!

More to come, including new documentation.

pilotdave

Andy wrote:

go add it to two photos, just to see, the , [comma] will show in the tool but not be part of you keywords.. we don't need the quotes anymore!

More to come, including new documentation.

I THINK this is a good thing but I'm not good with change. I add most of my keywords very quickly manually... I'm SO used to typing in the quotes and putting a space after the last keyword so I'm ready to go back and add another keyword later. This is going to take some getting used to. Are you guys offering any counseling or therapy to help us deal with a change to a tool we use so often?

There are some changes I'd love to see in the caption/keyword tool, but I didn't see this one coming...

Dave

pilotdave

Andy wrote:

go add it to two photos, just to see, the , [comma] will show in the tool but not be part of you keywords.. we don't need the quotes anymore!

More to come, including new documentation.

Out of curiosity (haven't tested it yet), have you stopped creating keywords from filenames? I'd imagine you must have at least stopped creating keywords like 1234 if the filename was "unrelatedtoimage_1234.jpg" because without quotes, that 1234 would now get indexed as a valid keyword.

Dave

Andy

pilotdave wrote:

"unrelatedtoimage_1234.jpg"

unrelatedtoimage is a keyword, 1234 is not, in this case.

Allen

This keyword change, comma vs quotes, is one of the changes that needs to
be brought forward to warn everyone. It messed one of my galleries, haven't
checked others yet. Also found all kinds of garbage numbers and split
"multiple word" words in keyword list. I've fixed what I could find but not
the "logged in" cloud yet.

Smugmug really needs to publish a list of things changing in an update BEFORE
updating. It's a PITA having to go to every corner to see what has changed,
good or bad.

scottkwest

keyword issues

Andy wrote:

unrelatedtoimage is a keyword, 1234 is not, in this case.

I'm seeing the numbers show up as keywords. My file names are things like SKW_1234. I used to get keywords of 'SKW'; '1234' (without quotes) and the 1234 would not be indexed. I would remove the SKW by hand.

Now the 1234 is showing up, and commas are appearing and disappearing.

See my keyword list here:
www.scottwestphotography.com/keywords
There are several 4 digit numbers at the beginning and end. Looks like only at the end when I'm logged out.

Where can I get more information about the changes?

pilotdave

scottkwest wrote:

I'm seeing the numbers show up as keywords. My file names are things like SKW_1234. I used to get keywords of 'SKW'; '1234' (without quotes) and the 1234 would not be indexed. I would remove the SKW by hand.

Now the 1234 is showing up, and commas are appearing and disappearing.

See my keyword list here:
www.scottwestphotography.com/keywords
There are several 4 digit numbers at the beginning and end. Looks like only at the end when I'm logged out.

Where can I get more information about the changes?

I'm guessing that if you never cleaned up the junk that smugmug has been creating all along, you just got a bunch of new keywords. But new uploads shouldn't add more junk... well, other than non-numerical junk.

Creating keywords from filenames is a bad, bad idea. Smugmug needs another workaround for a workaround because of this. The old workaround was to just not index keywords they created that they knew were junk (and only create keywords for the junk they didn't know was junk). Now they apparently just indexed the junk they knew better than to index before.

Smugmug, stop the insanity! Leave keywording to human beings.

Dave

scottkwest

I also used to be able to remove the file numbers by removing one digit at a time. So it would take 10 'removes' per gallery, removing 0, 1, ..., 9.

That was enough of a hassle that I usually didn't bother. But that single digit approach isn't working anymore either.

surffrog

New Keyword Change problem
Hi Smugmug or anyone that can help;


I use a search for your photos function that searches each photo individually by keyword. Each photo is tagged in the keyword by a number that is visible in the picture..For example a person running a road race (5K or 10K) wearing a number pinned to their chest.

I normally would enter "456" in the keyword edit box to identify runner #456. If i just enter 456 now without the double quotes, it works fine. However, if there is multiple runners in the photo (i.e. 456 and 356 and 21) and I try to enter keywords by the new comma separation method only 456 and 356 will show up. The number 21 does not work. I notice that numbers that are 2 digits are not showing up in this new keyword storage function with the comma delimiter. Now if I have multiple galleries of different running events two completely different people may have the same race number. For example: A race called the turkey trot 5k will have a runner with number 124 and a completely different race that was two weeks prior called the halloween bash 5k will also have a participant with number 124. If I am in the gallery for the turkey trot and I enter # 124 in the search field the participant with number 124 from the Halloween bash will show up also.....The way I stop this from happening is to create a keyword that uniquely identifies the name of the event and I place that keyword in all photos as well as the participants race number. With this new comma delimiter change I can no longer do this effectively. I can only do this with 3 digit numbers and the keyword for the event also. ...

Anyone know of a way around this now so I can get my galleries back to being searchable by race number and include my participants that have double digit and single digit numbers not just triple digit and larger?


Thanks,
Josh

dennismullen

Multi-word keywords
If i only have one multi-word keyword, like "all pictures", the comma after the keyword is thrown away so it isn't seen as a multi-word keyword and isn't allowed.

Andy

dennismullen wrote:

If i only have one multi-word keyword, like "all pictures", the comma after the keyword is thrown away so it isn't seen as a multi-word keyword and isn't allowed.

Show us, exactly where, please?

WinsomeWorks

Could someone just show me where to go to understand the ENTIRE new keywording processes as they exist now? Is there a list that exists now with all the latest info. on it? I simply am not understanding all the new ways it is working (not that I understood it 100% before, either). I'm getting a bunch of odd things happening, & cannot make heads nor tails of it. Last week, commas were being added at the end of every list. Now, they're not. I need to know when and where and how to use spaces, commas, quotation marks, semi-colons, numbers, phrases, etc. I had a lot of things correctly keyworded, or so I thought, and some of them are now goofy.

It also seems things are working differently in bulk keywording than they work when I simply click "edit" under the image to fix one photo's keywords. Oh, please- I hope not. How we list something in bulk processes shouldn't work any differently (essentially) from the one-image-at-a-time process. I also hope to goodness that this is all, or soon will all match Photoshop keywording processes. Otherwise it's an almost useless feature. For instance: right now in PS, phrases only need to be separated by a comma or semi-colon.... no quotation marks needed for this. It's a big pain to have to go into every keyword list here then and add quotation marks to each phrase I want kept as a phrase.

devbobo

Hi Anna Lisa,

I believe that keywording should be more straight forward now than previously.

Previously, we supported space, comas or semi colon separated keyword strings. In regard to building the keyword string from iptc keywords or the filename, the string was space separated. Having spaced separated strings leads to obvious issues with regard to multiple keywords, as a result these needed to be quoted.

Moving forward, when we create keyword strings from iptc keywords or the filename, we now use a coma separated format.

So it's pretty straight forward, you can use comas or semi colons as separators. eg...

Melbourne, New York, London
Melbourne; New York; London

Both of these examples result in the keywords 'Melbourne', 'New York' and 'London'.

Previously, when using space separated keyword strings, you would need to enter the following to achieve the same result...

Melbourne "New York" London

This method still works now, but you will notice that if you save that string, and then open the keyword dialog again...The keyword string has been rewritten as....

Melbourne, "New York", London

In regard to bulk keywording, if you can give me some specific examples of it not working how you would expect it, I will take a look.

Cheers,

David

devbobo

pilotdave wrote:

Trying to use the bulk keyword tool and I'm running into some unusual trouble. I try to add, for example, "joe b" to selected keywords. I end up with: joe b,. I don't want the comma, and I do want the quotes.

Also, the preview button no longer seems to work in IE6. I realize this is not a supported browser, but it has always worked before (and I'm stuck with IE6 for 8 hours a day).

Guessing this has to do with the recent change to allow foreign characters in keywords. Is this just the way it works now, or is this a bug?

Dave

The removal of the quotes show no longer be happening. But the commas will be added (under most circumstances) as this is our preferred method of delimiting now.

scottkwest

devbobo wrote:

Hi Lisa,

I believe that keywording should be more straight forward now than previously.

Previously, we supported space, comas or semi colon separated keyword strings. In regard to building the keyword string from iptc keywords or the filename, the string was space separated. Having spaced separated strings leads to obvious issues with regard to multiple keywords, as a result these needed to be quoted.

Moving forward, when we create keyword strings from iptc keywords or the filename, we now use a coma separated format.

So it's pretty straight forward, you can use comas or semi colons as separators. eg...

Melbourne, New York, London
Melbourne; New York; London

Both of these examples result in the keywords 'Melbourne', 'New York' and 'London'.

Previously, when using space separated keyword strings, you would need to enter the following to achieve the same result...

Melbourne "New York" London

This method still works now, but you will notice that if you save that string, and then open the keyword dialog again...The keyword string has been rewritten as....

Melbourne, "New York", London

In regard to bulk keywording, if you can give me some specific examples of it not working how you would expect it, I will take a look.

Cheers,

David

There seem to be a couple problems I can identify.

One, with you example, if the keyword is just New York (no other keywords), this some time becomes

New, York

and sometimes doesn't, for different images in the same gallery, apparently.

The second is with numbers, which for me get pulled out of the file name.

Take a look at this gallery:
http://www.scottwestphotography.com/Children/Zeeshan/

Sometimes the 4 digit filename number is included with the name, sometimes not. Note that this originally had my initials "skw" as a keyword (the other part of the filename), but I removed the skw from all the keyword strings.

devbobo

scottkwest wrote:

One, with you example, if the keyword is just New York (no other keywords), this some time becomes

New, York

and sometimes doesn't, for different images in the same gallery, apparently.

Whether or not the commas were added, it still would have been interpreted as two keywords. The ideal solution in this case would be to add the delimiter to the end of the string as...

New York,
New York;

But it seems that I'm stripping that out, which is a bug (I've fixed internally).

scottkwest wrote:

The second is with numbers, which for me get pulled out of the file name.

Take a look at this gallery:
http://www.scottwestphotography.com/Children/Zeeshan/

Sometimes the 4 digit filename number is included with the name, sometimes not. Note that this originally had my initials "skw" as a keyword (the other part of the filename), but I removed the skw from all the keyword strings.

I posted in another thread, that we are currently implementing a global flag to disable filename keywording at the account level. Is this what you want ? Or are you looking for certain keywords from the filename ?

WinsomeWorks

devbobo wrote:

Whether or not the commas were added, it still would have been interpreted as two keywords. The ideal solution in this case would be to add the delimiter to the end of the string as...

New York,
New York;

But it seems that I'm stripping that out, which is a bug (I've fixed internally).

I posted in another thread, that we are currently implementing a global flag to disable filename keywording at the account level. Is this what you want ? Or are you looking for certain keywords from the filename ?

David, Thanks much for the info. This helps clear up a lot. It would be nice if everyone knew where to go for this info. though-- other than DGrin. Hopefully when the next stage of changes is completed soon, everyone will have a visible note on their page to tell them. I'm very glad I won't have to put quotes in anymore. I suppose this change won't help with old files already uploaded though, right? I'm asking because I was still going through the laborious process of fixing a bunch of them when I realized my phrases from PS didn't "stick", and if it were retrospective I could stop that, but I'm guessing that may not even be possible.

I think with ScottKWest's issues, one thing everyone will want to see in order to even make the decision of whether to allow automated filename keywording or not, is consistency. In his case, if it were me, it would be hard to know whether I wanted that option turned off or on because I'm not seeing it behaving consistently.... sometimes his numbers are added, sometimes not.

A couple more questions I have are:
1. Does the addition of quotations or any other changes made after indexing has happened cause any problems for future keyword changes? (i.e. quotations will basically behave as if they are invisible, which is fine, right?)
2. Could you say what is meant exactly by a "delimiter"?
3. How do numbers behave?
4. How do apostrophes, dots ( . ) dashes & underlines behave?
5. How many keywords are indexed, and will this change soon?

Thanks!!!

devbobo

WinsomeWorks wrote:

1. Does the addition of quotations or any other changes made after indexing has happened cause any problems for future keyword changes? (i.e. quotations will basically behave as if they are invisible, which is fine, right?)

the addition of quotes will have no impact on future indexing...they are stripped from the keyword prior to it being indexed.

WinsomeWorks wrote:

2. Could you say what is meant exactly by a "delimiter"?

from dictionary.com

–noun Computers. a blank space, comma, or other character or symbol that indicates the beginning or end of a character string, word, or data item.

WinsomeWorks wrote:

3. How do numbers behave?

Numbers (3 chars or more) will be indexed regardless of whether they are quoted or not

WinsomeWorks wrote:

4. How do apostrophes, dots ( . ) dashes & underlines behave?

apostrophes, dots and underscores (underlines) are valid.
dashes or hyphens are not.

looks like there could be a bug with keywords that contain apostrophes not displaying correctly.

WinsomeWorks wrote:

5. How many keywords are indexed, and will this change soon?

the first valid 30 keywords are indexed, at this point, I don't think there is any plans to change that.

Andy

WinsomeWorks wrote:

It would be nice if everyone knew where to go for this info. though-- other than DGrin.

Keep in mind that you are extremely advanced - moreo than 99.99 % of all our users who will never go to the detail that you are going to

We're updating the help pages with all this info

scottkwest

devbobo wrote:

I posted in another thread, that we are currently implementing a global flag to disable filename keywording at the account level. Is this what you want ? Or are you looking for certain keywords from the filename ?

Yes, that will solve the problem for new galleries. I don't want the filename keyworded at all.

But there is a potential problem with the existing galleries. Only the ones I've touched recently have the numbers indexed. I'd like a way to remove the numbers from all existing galleries as well, or a way to have them not indexed. This might need some kind of script/tool that with options like "remove all number-only keywords", maybe with a length parameter. Maybe also a "remove all numbers", or a "remove all words that include a number".

For me, just removing all numbers works, but I know others use numbers.

Looking forward to what you come up with. Thanks for listening!

devbobo

scottkwest wrote:

Yes, that will solve the problem for new galleries. I don't want the filename keyworded at all.

But there is a potential problem with the existing galleries. Only the ones I've touched recently have the numbers indexed. I'd like a way to remove the numbers from all existing galleries as well, or a way to have them not indexed. This might need some kind of script/tool that with options like "remove all number-only keywords", maybe with a length parameter. Maybe also a "remove all numbers", or a "remove all words that include a number".

For me, just removing all numbers works, but I know others use numbers.

Looking forward to what you come up with. Thanks for listening!

Scott,

The easiest way to remove numeric indexed keywords (or any indexed keywod for that matter), is to go to the keyword gallery for that keyword then use the bulk tool to remove that offending keyword from all the images.

Cheers,

David

WinsomeWorks

Andy wrote:

Keep in mind that you are extremely advanced - moreo than 99.99 % of all our users who will never go to the detail that you are going to

We're updating the help pages with all this info

Oh, thank you, Andy! I'm so reassured to know I'm such an advanced human, errrrr, Smugger! I feel so smug now . Meh. Well. Part of it is actually the result of doing stock-- and having crossover files here. And happening to love language & words as well, & a humongous family to keep track of & direct them to galleries. It's complicated. Lots of people here are probably lucky they don't need the detail. But then the weirdos like me are lucky to have Smug, which has the detail when we want it. So even when I'm begging for this n' that, know that I'm singing its virtues all over the country! Now if I could just figure out how to customize my unlimited keyword page....!

scottkwest

devbobo wrote:

Scott,

The easiest way to remove numeric indexed keywords (or any indexed keywod for that matter), is to go to the keyword gallery for that keyword then use the bulk tool to remove that offending keyword from all the images.

Cheers,

David

Yeah, but if all of my galleries are indexed again, I will have up to 10,000 keywords from 0000 to 9999. Each of those 10K keyword galleries would have a couple photos. So I need something like a numeric wildcard for that to be practical.

Better might be an option like the 'empty' feature, but only for all-number keywords.

Scott

Erick L

devbobo wrote:

apostrophes, dots and underscores (underlines) are valid.
dashes or hyphens are not.

Would be nice if dashes were valid since they're actually used in composed words.

pilotdave

Out of curiosity (haven't tested it yet), have you stopped creating keywords from filenames? I'd imagine you must have at least stopped creating keywords like 1234 if the filename was "unrelatedtoimage_1234.jpg" because without quotes, that 1234 would now get indexed as a valid keyword.

Andy wrote:

unrelatedtoimage is a keyword, 1234 is not, in this case.

Andy, were you referring to the way it is now, or the way it will be at some point in the future?



This is worse than it used to be. Not only do more numbers in filenames get converted than before (dates), but 3+ digit numbers get indexed so they show up on the keywords list.

What would be the easiest way to leave the keywords I want and remove the keywords I don't want? In reality, I'm going to wait until I've finished uploading to that gallery and then remove all the keywords and start from scratch. But seriously, what do you expect people to do that do their keywording in external software? And how about the people that WANT to be able to use 1- or 2-digit numerical keywords? Wouldn't it have been much simpler to just get rid of filename keywording so you don't have to think all this through?

Dave

Malte

Erick L wrote:

Would be nice if dashes were valid since they're actually used in composed words.

I think they're invalid because they used to make up the URL for multiple keyword galleries, like username.smugmug.com/keyword/keyword1-keyword2. They've since change this to another slash, like username.smugmug.com/keyword/keyword1/keyword2, but the old way is still supported.

Malte

devbobo

Erick L wrote:

Would be nice if dashes were valid since they're actually used in composed words.

Erick,

I'd love to allow dashes/hyphens but we can't while it remains the separator for combined words as there would be no way to determine if a given word with a dash/hyphen was a combine word or a keyword in it's own right.

Cheers,

David

Malte

devbobo wrote:

Erick,

I'd love to allow dashes/hyphens but we can't while it remains the separator for combined words as there would be no way to determine if a given word with a dash/hyphen was a combine word or a keyword in it's own right.

Cheers,

David

So the reason is to not break existing links based on the "old" way of combining keywords? I'm thinking that all the instances where those links exist within SmugMug, people's customization and such, you guys would be able to find and replace. Links that have been handed out outside of SmugMug would be the problem. Maybe you could measure how many visitors come in that way, by a dash-combined keyword galley, from the outside to see what the impact would be.

Might be worth it?

Malte

Malte

devbobo wrote:

Scott,

The easiest way to remove numeric indexed keywords (or any indexed keywod for that matter), is to go to the keyword gallery for that keyword then use the bulk tool to remove that offending keyword from all the images.

Cheers,

David

I'll bet a vast majority of numerical keywords are camera sequence numbers, meaning designed to have no two be identical. In other words, one keyword gallery per keyword.

BTW, this indexing of keywords, what's the difference between an indexed keyword and a non indexed keyword?

Malte

pilotdave

Malte wrote:

I'll bet a vast majority of numerical keywords are camera sequence numbers, meaning designed to have no two be identical. In other words, one keyword gallery per keyword.

BTW, this indexing of keywords, what's the difference between an indexed keyword and a non indexed keyword?

Malte

An indexed keyword is a "real" keyword. It shows up in your keywords list, searches, etc. A non-indexed keyword is an invalid keyword. It'll appear in your keyword editing tools, but it won't appear "publicly" or on any keywords list or in search results.

Dave