I am confused with the order of all functions. Just to summarize. In
my simple example I want to generate TN from PDF like in PDF solution
pack but instead of adding TN as a datastream I want to create
derivative object and add TN there.
<alanjarlathstan
...@gmail.com> wrote:
> You have a potential problem here. You need the PID from F01 to build the
> RELS-EXT stream for f02, but F01 does not yet exist, so no pid is available.
> There would be a few of ways of dealing with this (there always is). I'd
> use a hook_form_alter within my custom module to add an additional submit
> handler to the fedora_repository_ingest_form, and I'd build my additional
> object there. By the time the first submit handler has done its work the
> pid for the new object will be in $form_state['values']['pid'].
> On 2012-05-01, at 12:53 PM, Serhiy Polyakov wrote:
> Alan,
> Your answer is very helpful.
> I decided to model my desired behavior as close as possible on PDF
> content model first and then move to implementing my specifics. I am
> basically creating derivative object while creating main object. I
> think I have everything to implement this. However, if you have any
> comments please let me know.
> So, instead of generating TN for PDF and adding TN as a datastream to
> the same Fedora Object I will do:
> -create main Fedora Object (FO1) with PID1
> -create new derivative FO2 with calculated PID2
> -add PDF datastream to FO1
> -generate TN from PDF
> -add TN to FO2
> -add to RELS-EXT of the FO2:
> <fedora:isMemberOf rdf:resource="info:fedora/PID1"></fedora:isMemberOf>
> -remove TN file from the file system
> Order of steps may be slightly different.
> Thanks,
> Serhiy
> On Tue, May 1, 2012 at 6:35 AM, Alan Stanley
> <alanjarlathstan...@gmail.com> wrote:
> You are correct about the method signature. Then method specified in your
> content model is executed by the execIngestRules in the ContentModel class
> so the method signature is locked down, but you can pass in null for the
> params if they are not used.
> On a standard system files will come from sites/all/files/ This isn't
> always the case, but the files will always be in the directory returned by
> file_directory_path(). You can see exactly what is going on by looking
> in fedora_repository_ingest_form_validate() in the fedora_repository.module
> file.
> The quick answer to question is three is - you can't get what you ant by
> altering that line, but you can do what you want to do. When files are
> added to the fedora_ingest_files array in the $_SESSION scope they go
> through a particular workflow which involves adding that file to the fedora
> object created at ingest time, and then deleting that file. At this point
> in the creation process you can create any derivative you like to be added
> to the object created at ingest time, or none at all. If I'm understanding
> you correctly you wish to create a second object while the first is being
> created. You can create a fedora object here in the usual way with
> $my_item = Fedora_Item::ingest_new_item($my_calculated_pid, 'A',
> $my_object_label);
> You'd then add whatever datastreams you wanted to this object before
> continuing with your myManipulation function.
> I hope this is what you need - let me know if there's anything I can clarify
> for you.
> - Alan
> On 2012-05-01, at 3:38 AM, Serhiy Polyakov wrote:
> I am working with ingest rules defined in a content model. For
> example, PDF CM has this rule:
> <ingest_method class="PDFManipulation" dsid="TN" file="pdf_sp.inc"
> method="createThumbnailFromPDF" modified_files_ext="jpg"
> module="islandora_pdf_sp">
> Then header of the function in pdf_sp.inc is this:
> function createThumbnailFromPDF($parameterArray, $dsid, $file, $file_ext)
> I defined my own function myManipulation for manipulations on my files
> and call it in my content model. I am guessing that I cannot change
> number of parameters in that function even I defined my new function.
> So I cannot have:
> function myManipulation($dsid, $file, $file_ext)
> Question 1. Is it correct?
> Question 2. To which directory files get uploaded (Linux) and from
> where they are taken for processing? For example in
> createThumbnailFromPDF processing line looks like:
> $cmdline = "convert \"$file\"\[0\] -colorspace RGB -thumbnail " .
> $width . "x" . $height . " \"$file$file_suffix\"";
> So does $file value include path beside filename? Is it some temp
> directory? Do files get deleted after uploading and processing?
> I see there are some files I uploaded to the repository are left in
> this directory:
> /var/www/myDrupal/sites/default/files
> Question 3.
> At the end of createThumbnailFromPDF function there is this a line:
> $_SESSION['fedora_ingest_files']["$dsid"] = $file . $file_suffix;
> How can I alter this part to create minimal new Fedora Object instead
> of just datastream for the existing object? I have calculated PID for
> the new object.
> Thanks you,
> Serhiy
> Alan Stanley
> Lead Developer, Islandora Project
> Robertson Library, UPEI
> 550 University Avenue, Charlottetown
> PE, C1A 4P3
> 1 902 566 6007
> Alan Stanley
> Lead Developer, Islandora Project
> Robertson Library, UPEI
> 550 University Avenue, Charlottetown
> PE, C1A 4P3
> 1 902 566 6007
