Skip to content



Islandora Workbench offers three "hooks" that can be used to run scripts at specific points in the Workbench execution lifecycle. The three hooks are:

  1. Bootstrap
  2. CSV preprocessor
  3. Post-action

Hook scripts can be in any language, and need to be executable by the user running Workbench.

Execution (whether successful or not) of hook scripts is logged, including the scripts' exit code.

Bootstrap and shutdown scripts

Bootstrap scripts execute prior to Workbench connecting to Drupal. For an example of using this feature to run a script that generates sample Islandora content, see the "Generating sample Islandora content" section.

To register a bootstrap script in your configuration file, add it to the bootstrap option, like this, indicating the absolute path to the script:

bootstrap: ["/home/mark/Documents/hacking/workbench/scripts/"]

Each bootstrap script gets passed a single argument, the path to the Workbench config file that was specified in Workbench's --config argument. For example, if you are running Workbench with a config file called create.yml, "create.yml" will automatically be passed as the argument to your bootstrap script (you do not specify it in the configuration), like this: create.yml

Shutdown scripts work the same way as bootstrap scripts but they execute after Workbench has finished connecting to Drupal. Like bootstrap scripts, shutdown scripts receive a single argument from Workbench, the path to your configuration file.

A common situation where a shutdown script is useful is to check the Workbench log for failures, and if any are detected, to email someone. The script in the scripts directory shows how this can be used for this.

To register a shutdown script, add it to the shutdown option:

shutdown: ["/home/mark/Documents/hacking/workbench/"]

--check will check for the existence of bootstrap and shutdown scripts, and that they are executable, but does not execute them. The scripts are only executed when Workbench is run without --check.

A shutdown script that you might find useful is scripts/, which generates the IIIF Manifest (book-manifest) for each node in the current create task that is a parent. You might want to do this since pregenerating the manifest usually speeds up rendering of paged content in Mirador and OpenSeadragon. To use it, simply add the following to your create task config file, adjusting the path to your Workbench scripts directory:

shutdown: ["/home/mark/hacking/workbench/scripts/"]


Bootstrap and shutdown scripts get passed the path to your configuration file, but they only have access to the configuration settings explicitly defined in that file. In other words, any configuration setting with a default value, and therefore no necessarily included in your configuration file, is not known to bootstrap/shutdown scripts.

Therefore, it is good practice to include in your configuration file all configuration settings your script will need. The presence of a configuration setting set to its default value has no effect on Workbench.

CSV preprocessor scripts

CSV preprocessor scripts are applied to CSV values in a specific CSV field prior to the values being ingested into Drupal. They apply to the entire value from the CSV field and not split field values, e.g., if a field is multivalued, the preprocessor must split it and then reassemble it back into a string. Note that preprocessor scripts work only on string data and not on binary data like images, etc. and only on custom fields (so not title). Preprocessor scripts are applied in create and update tasks.


If you are interested in seeing preprocessor scripts act on binary data such as images, see this issue.

For example, if you want to convert all the values in the field_description CSV field to sentence case, you can do this by writing a small Python script that uses the capitalize() method and registering it as a preprocessor.

To register a preprocessor script in your configuration file, add it to the preprocessors setting, mapping a CSV column header to the path of the script, like this:

 - field_description: /home/mark/Documents/hacking/workbench/scripts/

You must provide the absolute path to the script, and the script must be executable.

Each preprocessor script gets passed two arguments:

  1. the character used as the CSV subdelimiter (defined in the subdelimiter config setting, which defaults to |)
    • unlike bootstrap, shutdown, and post-action scripts, preprocessor scripts do not get passed the path to your Workbench configuration file; they only get passed the value of the subdelimiter config setting.
  2. the CSV field value

When executed, the script processes the string content of the CSV field, and then replaces the original version of the CSV field value with the version processed by the script. An example preprocessor script is available in scripts/

Post-action scripts

Post-action scripts execute after a node is created or updated, or after a media is created.

To register post-action scripts in your configuration file, add them to either the node_post_create, node_post_update, or media_post_create configuration setting:

node_post_create: ["/home/mark/Documents/hacking/workbench/"]
node_post_update: ["/home/mark/Documents/hacking/workbench/"]
media_post_create: ["/home/mark/Documents/hacking/workbench/"]

The arguments passed to each post-action hook are:

  1. the path to the Workbench config file that was specified in the --config argument
  2. the HTTP response code returned from the action (create, update), e.g. 201 or 403. Note that this response code is a string, not an integer.
  3. the entire HTTP response body; this will be raw JSON.

These arguments are passed to post-action scripts automatically. You don't specific them when you register your scripts in your config file. The scripts/ illustrates these arguments.

Your scripts can find the entity ID and other information within the (raw JSON) HTTP response body. Using the way Python decodes JSON as an example, if the entity is a node, its nid is in entity_json['nid'][0]['value']; if the entity is a media, the mid is in entity_json['mid'][0]['value']. The exact location of the nid and mid may differ if your script is written in a language that decodes JSON differently than Python (used in this example) does.


Not all Workbench configuration settings are available in post-action scripts. Only the settings are explicitly defined in the configuration YAML are available.

As with bootstrap and shutdown scripts, when using post-action scripts, it is good practice to include in your configuration file all configuration settings your script will need. The presence of a configuration setting set to its default value has no effect on Workbench.

Running multiple scripts in one hook

For all types of hooks, you can register multiple scripts, like this:

bootstrap: ["/home/mark/Documents/hacking/workbench/", "/home/mark/Documents/hacking/workbench/"]
shutdown: ["/home/mark/Documents/hacking/workbench/", "/home/mark/Documents/hacking/workbench/"]
node_post_create: ["/home/mark/scripts/", "/tmp/"]

They are executed in the order in which they are listed.