Support for IPTC
surge_martin - December 30, 2008 - 14:33
| Project: | Imagefield Import |
| Version: | 6.x-1.x-dev |
| Component: | Code |
| Category: | feature request |
| Priority: | normal |
| Assigned: | wulff |
| Status: | fixed |
Description
It would be great this module could access the metadata for pictures, and use it for titles, etc.
#1
subscribe
#2
In the for each loop add the following under the:
foreach ($files as $file) {Add:
$image_name = $dirpath . '/' . basename($file->filename);$size = getimagesize($image_name, $info);
if(isset($info['APP13']))
{
$iptc = iptcparse($info['APP13']);
$caption = $iptc["2#120"][0];
}
else {
$caption = basename($file->name);
}
Replace:
$form['files']['title'][] = array( '#type' => 'textfield', '#size' => 20, '#default_value' => basename($file->name), );With:
$form['files']['title'][] = array( '#type' => 'textfield', '#size' => 20, '#default_value' => $caption, );This will use the iptc caption for the title.
This is on the drupal 6 version.
#3
Sorry for my ignorance on IPTC--
Do all images have such metadata?
How is the caption usually used?
Would that make a more suitable body?
#4
To my knowledge, most image formats have some capacity to store metadata. TIFF, RAW and jpg seem the image formats most widely read from & written to. IPTC is one of a few standards by which the data's written. There isn't a lot of consistency between software support for individual metatags or which tags are written by which cameras, but they all write something that identifies variables of the image origination, and many allow further user modifications, eg user comments, keyword insertion, copyright data, etc.. The caption is used differently in different contexts, but is more flexible than some IPTC tags in that it can hold more characters.
The EXIF module (http://drupal.org/node/226703) provides some strong image metadata-reading functions, but I haven't yet been able to configure it to read the fields I need. I've also been revisiting hedinfoto's caption solution above, trying to adapt it to gather keywords, but the coding is beyond me so far.
For me, the most powerful application of harvesting metadata at the image bulk upload stage would be to allow an interface between metadata content and other drupal entities, eg importing keywords or comments and mapping those to selected taxonomies, as the maqum module (http://drupal.org/project/maqum) currently does for single taxonomies on single images, or importing subject fields and mapping them to user references.
Totally thrilled to see you examining this area. If there are ways an image geek much more comfortable on the camera end can support/assist/encourage development, please let me know.
Thanks!
#5
Subscribing.
Could really need iptc support for d6 as this is what keeps me chained to d5 (with maqum) currently.
#6
The following works for me. The caption of the image goes into the Title and the keywords go into the Body.
$size = getimagesize($dirpath.'/'.basename($file->filename), $info);if(isset($info['APP13']))
{
$iptc = iptcparse($info['APP13']);
// Key 2#120 is the caption
$caption = $iptc["2#120"][0];
// Key 2#025 is the keywords (not sure if this is true only for Photoshop)
$keywords = implode(', ',$iptc["2#025"]);
}
else {
$caption = basename($file->name);
$keywords = " ";
}
// image_get_info is a Drupal Function That Gets Details about an image
The last line is given to anchor the code
The second snippet is straightforward assignment of the Title and Body fields:
$form['files']['title'][] = array('#type' => 'textfield', '#size' => 20, '#default_value' => $caption, );$form['files']['body'][] = array('#type' => 'textfield', '#size' => 20, '#default_value' => $keywords, );
@vordude:
Caption and Keywords are usually used on photo sites that have large number of contributors that upload large batches of files. Typically the node is given the title from the caption field and keywords are used for the database search. Typical sites that need this are photos stock sites such as istockphoto.com
It would be great to extend the given code snippet to automatically create a node type that has several text fields corresponding to EXIF/IPTC fields.
Hope this helps.
#7
This works really well.
We have metadata in a lot of fields -- here's a little script I found that translates various fields into their IPTC codes.
Set up an image in photoshop (or wherever) with your metadata, then feed it to this script. It'll tell you what "2#xyz" keys match the fields you use.
<html>
<head>
<title>Metadata Checker</title>
</head>
<body>
<?php
function output_iptc_data( $image_path ) {
$size = getimagesize ( $image_path, $info);
if(is_array($info)) {
$iptc = iptcparse($info["APP13"]);
foreach (array_keys($iptc) as $s) {
$c = count ($iptc[$s]);
for ($i=0; $i <$c; $i++)
{
echo $s.' = '.$iptc[$s][$i].'<br>';
}
}
}
}
$image_path = "/server/path/to/the/image/with/metadata/in/it/.jpg";
output_iptc_data($image_path);
?>
</body>
</html>
#8
#9
How funny, just the other day I put together a similar patch for myself...
Instead of putting keywords into the body field, however, I wanted to put (IPTC) keywords into (taxonomy) Tags.
As well, it reads the EXIF 'DateTime' into a cck date field.... and, because I don't want duplicates it will ignore filenames that already exist (rather than renaming them).
Basically it adds a 'tags' text field to each row of the import form, and populates it based on the iptc keywords. This list is appended to the taxonomy field already present on the form (so that a tag can be added to the entire batch of files).
I'm attaching the patch for reference only, in case anyone feels like putting together a system like this on their own, but note that some things (like the taxonomy vocabulary, the name of the CCK date field and files location) are hardcoded, and it uses things like $_SERVER['DOCUMENT_ROOT'], to find directories, which is most probably not the drupal way of getting the base path...
But this system is working great for me for images that my lady friend tags using iPhoto. Using taxonomy makes it simple to put together galleries based on views (I use jquery/galleria for this), and I like having the date that the photo was taken stored as well.
Also, I consistently get warnings of invalid values from the exif_read_data function, but this doesn't seem to cause any problems, and might just be a bug in php http://bugs.php.net/bug.php?id=40073 or in the exif data that my camera puts out.
cheers.
edit: I think it would be great if some EXIF and IPTC data was read in by the core image_get_info function()...
#10
Hey rubberglove,
that sounds really useful and I'd like to see your code for reference, as I'm gonna need something similar sooner or later, too.
Did you forget to attach your patch? Can't see it.
Thanks in advance,
Daniel
#11
EXIF and IPTC are related enough that people use them interchangeably and generically to refer to image metadata.
Somewhat more specifically, these names represent families of image metadata tags: The camera itself usually populates EXIF, and you can use editing programs like Graphic Converter or Nikon's View NX to edit IPTC. This latter set of tags is more consistently used than EXIF because journalists and newspapers need a consistent way of labeling and tracking media. In addition, the XMP family of tags is often used by other programs like photoshop or windows applications, but here again, each application extends and adds additional tags. (Not to mention re-use and over-writing of tags named the same in different families).
The IPTC defines a keywords field. I use the contact field as a secondary taxonomy vocabulary to tag people or places. I use various other IPTC fields more-or-less as defined, but sometimes I redefine them for my own purposes.
ALL PRESENT DRUPAL IMAGE MODULES ARE BAD AT READING EXIF DATA
The php exifread function doesn't seem to be very forgiving, but is commonly used because it is available on many (most?) hosting sites. I've abandoned it because I get bad characters and strange entries coming out of my meta tags, which have been edited on an apple. I haven't tried the iptc parse, yet. Instead, I use exiftool to dump my tags to a sidecar file.
The perl utility: exiftool is EXTREMELY robust and very well-maintained. You can examine tags and tag families. It let's you read or write any of your tags and it is scriptable so you can do bulk tag editing across your photographs.
Also, exiftool lets you can dump any or all of your photo tags to a sidecar file in various formats like: tag:value, JSON or XMP. Building a sidecar file may seem like additional hassle, but if you are uploading via ftp, then there is no problem in bringing along additional text files.
Hint1: php has decent read functions for JSON encoded data....
Hint2: fast_gallery automatically builds, updates and deletes galleries out of folders.
Hint3: It is not integrated with nodes or image_field
http://PhotoKinesis.US
#12
ah - not only did I forget to attach my patch (or so it seems), but I got busy and forgot to check back here...
apologies for the delay, hopefully this could still be useful...
#13
I have developed a module which brings API functions for IPTC metadata support (even encoding handling/conversions)
But I won't be able to contribute it since I can't use CVS due to firewall restrictions
If you read my forum post Looking for a CVS committer you can find
apix_iptclistedIf somebody comes up with some idea I would gladly contribute it, maintain it and give support for other module developers which might wan't to use its API
#14
IPTC and EXIF support is on my own wishlist as well. Marking this as a feature request for the Drupal 6 version.
Also, see the idea for adding tags from #501358: Load Keywords automatically for each image (not group of images).
#15
I'm in
I have coded IPTC's API enough, but requires clean up,
actually what it requires is an clean API design,
being alone with this I made it work for my job's use cases, but certainly other developers would bring better ideas
IPTC handling has some pitfalls I have sorted out, so double effort can be avoided
how do we proceed with this?
PS: I have found a CVS committer (sinasalek) which is already co-maintaining a couple of module with me
#16
The first thing I would need is a proper name for the module which will hold the ITPC's API
after that following issues discussing how functions signature (interface) should be better
#17
I wonder if we really need that much extra code.
Wouldn't it be enough to grab the title and body from IPTC (using PHP's iptcparse() and grabbing the fields 2#005 and 2#120)?
#18
The latest version of the module now lets you prepopulate the title and body field with EXIF or IPTC data.
#19
I don't see it as "much extra code" since it isn't "much" but rather "tricky" code
there are other crucial IPTC metadata as Keywords and Category/sub-categories
also there is the encoding problem CP_1252/CP_UTF8
are you sorting this out?
#20
BTW: I was talking about an API module for IPTC, since metadata handing wouldn't be a particular case for this module, but rather for any image-handling module
#21
At the moment, I just grab the title and description and convert them from cp1252 to UTF8 using iconv.
In time it might be a nice addition to let users map arbitrary IPTC fields to the title and body field.
If you want to create a clean API, it might be a good idea to contact the author of the IPTC module: http://drupal.org/project/iptc
#22
no, thanks
1- it depends on image module
2- most of all, it uses external libraries to do so: exiftool binary, exiv2 binary.
3- only has 5.x-dev snapshot
4- seems to be frozen since 2008-Mar-29
the functionality I'm currently handling is all about
iptcparse/iptcembed- providing
iptc_make_tagfunction (by Thies C. Arntzen) with better API to handle an array of IPTC data around (encoding taken into account, etc)- pitfalls safety: handling the tricky cases
- the set of IPTC constants (with their Photoshop equivalents)
- etc
just a centralized place to put the known stuff plus a couple of weirdness
#23
BTW: what I called above "the known stuff" is pretty hard to find well documented altogether in one single place, that's why I think such an API is needed for any PHP framework
despite the fact that XMP is taking IPTC off the table, there are several press agency with old legacy software that will remain handling IPTC and/or require their websites to take care of the existing metadata into archived photos
PS: just replying to your recommendation, I understood it won't be part of this module