You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
45 lines
1.8 KiB
45 lines
1.8 KiB
5 years ago
|
# islandora_text_extraction
|
||
|
|
### Connects Islandora 8 to Hypercube microservice and extracts text from PDFs
|
||
|
|
|
||
|
|
Install module in the usual way,
|
||
|
|
then copy `assets/ca.islandora.alpaca.connector.ocr.blueprint.xml`
|
||
|
|
to `/opt/karaf/deploy` on the server.
|
||
|
|
_note:_ This config file assumes a url of `http://localhost:8000/hypercube`.
|
||
|
|
If your service is found elsewhere this must be changed.
|
||
|
|
There is no need to restart.
|
||
|
|
|
||
|
|
In the usual Ansible build this will require no modification.
|
||
|
|
|
||
|
|
If a parent node is tagged as `Digital Document` an `Image` tagged media
|
||
|
|
will extract text from that image at the time of ingestion.
|
||
|
|
The content type of the parent node should be configured to allow multiple tags.
|
||
|
|
|
||
|
|
_note:_ Media are linked to their parent nodes with the `Media Of`
|
||
|
|
entity reference field. If you wish to attach the PDF (or any other ) media type
|
||
|
|
to a parent node which has any content type other than Repository Item
|
||
|
|
(islandora_object) the parent content type will have to be added to the `Media Of`
|
||
|
|
field in the media type description.
|
||
|
|
|
||
|
|
## Prepare module for PDF text extraction
|
||
|
|
Install `texttopdf` on your server if not already present.
|
||
|
|
On an ubuntu/debian machine like the default claw playbook run
|
||
|
|
`sudo apt-get install poppler-utils`
|
||
|
|
|
||
|
|
test to see its been properly installed with `which pdftotext`
|
||
|
|
|
||
|
|
Install php libraries with `composer require spatie/pdf-to-text`
|
||
|
|
|
||
|
|
In the unlikely event that your `pdftotext` binary exists on your server
|
||
|
|
outside of the system path, the path to the binary can be set at
|
||
|
|
`/admin/config/islandora/text_extraction`.
|
||
|
|
|
||
|
|
## Using text extraction ##
|
||
|
|
The containing document must be tagged as `Digital Document`,
|
||
|
|
and the media must be tagged as `Original File`.
|
||
|
|
A new editable `Extracted Text` media will be created and attached when `PDF` or
|
||
|
|
`Image` media types are added to a node.
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|