keyword change? 3 character minimum?

kny

Has there been a change to keyword behavior? Well, yes, there has definitely been a change, but is it by design or a bug?

It seems that only keywords of at least 3 characters or more are actually indexed.

If I look at the keywords for an image, only those with 3 or more characters are shown. But, if I click to Edit keywords, I see all the keywords that have been set for the image, even those with just 1 or 2 characters. I can make edits and save, but again, only those with 3 or more characters actually show and are actually indexed.

This is critical to my business. My images are all tagged according to a strict set of rules, and some of these tags are only 1 or 2 characters. All images are uploaded and tags set via a program that uses the API.

Here is an example of a recently uploaded photo that does not have all of its keywords actually indexed.
http://tri-to-win-events.smugmug.com/Sports/2009-SavageMan/9725936_f5Yky#727366208_tAcXF

Please assist.

devbobo

Hi kny,

We have had a few teething issues related to a recent keyword rewrite to support UTF-8. This issue has been fixed internally and should be going live some time this week.

Cheers,

David

kny

devbobo wrote:

Hi kny,

We have had a few teething issues related to a recent keyword rewrite to support UTF-8. This issue has been fixed internally and should be going live some time this week.

Cheers,

David

Thanks. I can hold off on uploading new images from events for a few days, but what about the ones that I recently uploaded. Do I need to do anything to get all of the tags indexed for these pictures, or will this happen automatically?

(while you're at it, remove or increase that 30 keyword limit, please.... )

devbobo

You will need to resave the keyword string to re-index the keywords...whether that is individually or in bulk for a whole gallery.

At this point, we aren't removing the 30 keyword limit.

Cheers,

David

kny

devbobo wrote:

You will need to resave the keyword string to re-index the keywords...whether that is individually or in bulk for a whole gallery.

At this point, we aren't removing the 30 keyword limit.

Cheers,

David

THanks again,
I just want to confirm - because this is critical to my business and if this issue isn't going to be fixed I will have to figure something out and pronto - that this specific issue of character keywords is a known issue that is being fixed and rolled out this week. It sounds like there have been a number of keyword-related "teething" issues, and I just want to know for certain that this is one that is known and definitely being addressed.

Also, how can I be notified when the changes are rolled out.
Thanks much.
-Kyle

devbobo

Kyle,

Based on the image you linked above, here's a screenshot of the keywords indexed in my development environment..



It looks fine to me.

Cheers,

David

jcmx

I would like to also echo this issue. My business regularly indexes keywords with 1 or 2 characters (numbers on motorcycles). Without this I'm hosed. When will the fix go live? Didn't seem to be working as of sunday night 10/29.

jcmx

Also, what's the preferred method to get the reindexing? All my keywords are stored in the image information (EXIF I believe). I certainly don't want to have to go back through 250 or so photos using the web interface and add all the bike numbers (I wrote a client program to do that and stuff the keywords in the image files). So what's the best way? Delete all the images and reupload? Ugh....

devbobo

jcmx wrote:

Also, what's the preferred method to get the reindexing? All my keywords are stored in the image information (EXIF I believe). I certainly don't want to have to go back through 250 or so photos using the web interface and add all the bike numbers (I wrote a client program to do that and stuff the keywords in the image files). So what's the best way? Delete all the images and reupload? Ugh....

there is no need to delete and reupload, the bike numbers still exist in the keyword string, just not indexed as keywords.

So all you need to do, is load a gallery up in the bulk caption/keyword tool and click Save without editing anything. The existing keyword strings will be reprocessed and the missing keywords will be indexed.

Cheers,

David

devbobo

in case, you don't know...this is what i mean by the keyword string. You will see the 1 or 2 char quoted string exist there, despite not being indexed.

kny

devbobo wrote:

in case, you don't know...this is what i mean by the keyword string. You will see the 1 or 2 char quoted string exist there, despite not being indexed.

Yup, the keyword string is definitely being stored, just not indexed. How will I know when this fix is in place so I can go and re-save the keywords and have them index? Should I try on Friday after the upgrade Thurs. night?

kny

kny wrote:

Yup, the keyword string is definitely being stored, just not indexed. How will I know when this fix is in place so I can go and re-save the keywords and have them index? Should I try on Friday after the upgrade Thurs. night?

Not fixed as of 12/4 am. Did this fix not get included in the upgrade in the maintenance window last night.

I am holding off sending out email for Christmas sales promotion because people's images are not properly indexed and thus won't return in the search interface provided for them. I was really, really hoping this would be working this morning.

BSing

kny wrote:

Not fixed as of 12/4 am. Did this fix not get included in the upgrade in the maintenance window last night.

I too was waiting for a couple of keyword fixes, and hoping for a "don't junk up my keywords with filenames" master switch. I don't think they made it. :cry

devbobo

hang tight

devbobo

Bug fixes are now live, plus an account level setting to disable filename keywording.

kny

devbobo wrote:

Bug fixes are now live, plus an account level setting to disable filename keywording.

All right! Best news of the day! I was sure I'd be waiting for the next maintenance window for this.

Holiday sales here I come!!!

Thanks!

denisegoldberg

devbobo wrote:

...plus an account level setting to disable filename keywording.

Thank you!

--- Denise

alacranera

While I very much appreciate the disabling of auto-generated keywords, I am wondering if there is some way or removing, in bulk, the meaningless keywords that have already been indexed. My galleries' keywords are completely contaminated with useless sequential numbers that are in no way descriptive, and should never have been considered keywords. I do not look forward to removing them picture by picture, gallery by gallery.

I tried using a couple of wildcards (ie 0*, and 0??) to ID the offending keywords, but no luck. Any other ideas?

devbobo

alacranera wrote:

While I very much appreciate the disabling of auto-generated keywords, I am wondering if there is some way or removing, in bulk, the meaningless keywords that have already been indexed. My galleries' keywords are completely contaminated with useless sequential numbers that are in no way descriptive, and should never have been considered keywords. I do not look forward to removing them picture by picture, gallery by gallery.

I tried using a couple of wildcards (ie 0*, and 0??) to ID the offending keywords, but no luck. Any other ideas?

if you browse to the offending keyword gallery, you can then use the bulk tool to remove that keyword.

alacranera

devbobo wrote:

if you browse to the offending keyword gallery, you can then use the bulk tool to remove that keyword.

Thanks for the quick reply.

I understand that, but what I think maybe you folks at smugmug have never quite understood is that for many people parts of filenames are just sequential numbers. So every picture has a unique number that has been indexed as a keyword. So browsing to a keyword gallery that was based on a sequential filename brings up exactly 1 picture whose keywords need editing. Which means a picture by picture edit.

Just as an example: All of my filenames across a jillion galleries are MyNeatoCamera_1234.jpg, MyNeatoCamera_1235.jpg,
MyNeatoCamera_1236.jpg, etc. Smugmug auto generated and indexed the keywords 1234, 1235, 1236 etc. out of that. But each of those keyword galleries contains only its unique image. And so each keyword removal has to be done image by image.

I guess I'm asking if there is a gallery wide (or bigger) wildcard tool that will allow for selecting, in this example, all four-digit numbers that start with "1" (1???, for instance) for removal. It would be ever so handy to delete all the useless keywords in a fell swoop, rather than one by one.

Alternatively, is there a way to un-Index what smugmug already autogenerated?

Thanks for any help

jfriend

devbobo wrote:

if you browse to the offending keyword gallery, you can then use the bulk tool to remove that keyword.

Thanks devbobo for giving us the ability to prevent further keyword generation from camera generated filenames. Now, in order to get back to actually using keywords on Smugmug, I have a big mess to clean up and I'm looking for ideas on how to do so. These keywords are spread over hundreds of galleries and thousands of images and are intermixed with real keywords from the IPTC that I intend to keep so I don't want to just clear them all. I can't see any doable way to use bulk keywording to fix. Any other ideas how to clean up this mess?

Here's what one piece of my keywords page looks like to give you an idea of the mess it is now:

devbobo

alacranera wrote:

Thanks for the quick reply.

I understand that, but what I think maybe you folks at smugmug have never quite understood is that for many people parts of filenames are just sequential numbers. So every picture has a unique number that has been indexed as a keyword. So browsing to a keyword gallery that was based on a sequential filename brings up exactly 1 picture whose keywords need editing. Which means a picture by picture edit.

Just as an example: All of my filenames across a jillion galleries are MyNeatoCamera_1234.jpg, MyNeatoCamera_1235.jpg,
MyNeatoCamera_1236.jpg, etc. Smugmug auto generated and indexed the keywords 1234, 1235, 1236 etc. out of that. But each of those keyword galleries contains only its unique image. And so each keyword removal has to be done image by image.

I guess I'm asking if there is a gallery wide (or bigger) wildcard tool that will allow for selecting, in this example, all four-digit numbers that start with "1" (1???, for instance) for removal. It would be ever so handy to delete all the useless keywords in a fell swoop, rather than one by one.

Alternatively, is there a way to un-Index what smugmug already autogenerated?

Thanks for any help

just post the url for your gallery and i'll organise to have all the numeric keywords removed

devbobo

jfriend wrote:

Thanks devbobo for giving us the ability to prevent further keyword generation from camera generated filenames. Now, in order to get back to actually using keywords on Smugmug, I have a big mess to clean up and I'm looking for ideas on how to do so. These keywords are spread over hundreds of galleries and thousands of images and are intermixed with real keywords from the IPTC that I intend to keep so I don't want to just clear them all. I can't see any doable way to use bulk keywording to fix. Any other ideas how to clean up this mess?

Here's what one piece of my keywords page looks like to give you an idea of the mess it is now:

John,

I'm happy to remove them for you.

Let me know if there are any you want to keep.

Cheers,

David

Pupator

devbobo wrote:

John,

I'm happy to remove them for you.

Let me know if there are any you want to keep.

Cheers,

David

I also would like every keyword removed from my account - this will enable me to begin using them properly. Thanks for the fix!

kny

jfriend wrote:

Thanks devbobo for giving us the ability to prevent further keyword generation from camera generated filenames. Now, in order to get back to actually using keywords on Smugmug, I have a big mess to clean up and I'm looking for ideas on how to do so. These keywords are spread over hundreds of galleries and thousands of images and are intermixed with real keywords from the IPTC that I intend to keep so I don't want to just clear them all. I can't see any doable way to use bulk keywording to fix. Any other ideas how to clean up this mess?

Here's what one piece of my keywords page looks like to give you an idea of the mess it is now:

If you want to remove all numeric keywords from every keyword string, I've got a python script that can do that for you. Right now it accepts Gallery name as a parameter, which would be a nuisance if you have hundreds of galleries. But, if you're moderately technical and not afraid of the API it could probably easily be modified to iterate through all galleries and do this.

alacranera

devbobo wrote:

just post the url for your gallery and i'll organise to have all the numeric keywords removed

No, Thank you. The last time I made a generic comment about the mysteries of smugmug, one of your lackies took it upon himself to look up my site and "fix" it. It only took a few hours to undo his good deeds. I do not want to repeat that event.

Your offering to "fix" my site does not at all address what seems to be an emerging global problem: lots of people are having problems with their sites being inundated with junk keywords.

Surely, if you have a solution for my particular site, you have a solution for everybody? Right? Rather than offering to fix sites one-by-one, why don't you just publish that solution? Or at least admit you screwed up royally with that decision to generate keywords out of filenames?

devbobo

alacranera wrote:

Or at least admit you screwed up royally with that decision to generate keywords out of filenames?

Filename keywording was in place long before I started working here. I'm just trying to help out, if you want to do it yourself...that's cool.

Andy

alacranera wrote:

No,

Alacranera, I'm looking around, we have no 'lackeys' here, just dedicated SmugMug employees, including Devbobo who wants to help you on his Saturday afternoon (he's in Australia).

Devbobo is a SmugMug sorcerer, and if he says he can fix you, he will

devbobo

jfriend wrote:

Thanks devbobo for giving us the ability to prevent further keyword generation from camera generated filenames. Now, in order to get back to actually using keywords on Smugmug, I have a big mess to clean up and I'm looking for ideas on how to do so. These keywords are spread over hundreds of galleries and thousands of images and are intermixed with real keywords from the IPTC that I intend to keep so I don't want to just clear them all. I can't see any doable way to use bulk keywording to fix. Any other ideas how to clean up this mess?

Here's what one piece of my keywords page looks like to give you an idea of the mess it is now:

John,

I spent a few hours this afternoon cleaning up the majority of the numeric keywords that are indexed on your site.

Cheers,

David

jfriend

devbobo wrote:

John,

I spent a few hours this afternoon cleaning up the majority of the numeric keywords that are indexed on your site.

Cheers,

David

Wow! Thanks David. I was out for a few hours at a holiday party, came back and it was all fixed up. That was above and beyond the call of duty. Many, many thanks!

Ham1

alacranera wrote:

Or at least admit you screwed up royally with that decision to generate keywords out of filenames?

I can put you in touch with an event photographer or two that say filename keywording is critical to their business.

The right thing for us to do a long time ago was put the option to disable/enable the feature, which we now have inside the Control Panel

---

>Settings

---

>Site Privacy & Security

Markham