Download the PHP package aplia/content-tools without Composer

Improved API for working with content objects in eZ publish legacy

This packages contains classes which makes it easier to work with content objects when it comes to creating and updating content. It also has functionality for migrating content using Phinx.

Whenever something is missing or an action fails it will throw a proper Exception, see the namespace Aplia\Content\Exceptions for a list of possible errors.

ContentType

Wrapper around eZContentClass which defines a new class with chosen attributes.

Example, creating a new content class:

sortBy is a special property which should contain a string with the identifier of the field to sort-by as specified by eZContentObjectTreeNode, to reverse sorting prepend a dash. e.g. name and -name.

Each call to addAttribute creates a new ContentTypeAttribute instance which will be created in eZ publish when create(), update() or save() is called. An attribute mainly consists of the data-type, identifier and name. In addition a fourth parameter may be set which is an array with additional named parameters.

The following are common for all attributes.

• isRequired - Whether the attribute is required or not, defaults to false.
• isSearchable - Whether the attribute will be registered to the search index, defaults to true.
• canTranslate - Whether the attribute is allowed to be translated, defaults to true but also depends on the data-type in use.
• isInformationCollector - Whether the attribute is an information collector, defaults to false.
• value - Addtional value to send to the data-type, this is an associative array and the content depends on the data-type.
• language - The language of the attribute, defaults to the same language as the ContentType it belongs to.
• description - Description text for the attribute.
• placeAfter - Place attribute after the existing attribute, specified using the identifier
• placeBefore - Place attribute before the existing attribute, specified using the identifier

Similarly attributes may be removed with removeAttribute and supplying the identifier. This schedules the attribute to be removed on the next call to save().

There are also methods for checking for attribute existance and fetching attribute object. (hasAttribute() and classAttribute())

To remove a class use remove(), just check if it exists() first.

For easy access to content-class and attributs use the properties contentClass and attributes, these properties will lazy-load the values. Note they require that content-class exists.

ezselection2

Example

Groups

Class-groups can be managed using ContentType. To create or delete groups use ContentType::createGroup and ContentType::deleteGroup.

For instance:

Assigning a class to a group is done using addToGroup(), removeFromGroup() and setGroups(). These schedules changes which are written on the next sync. If the group assignment is present (added/removed) nothing changes.

For instance:

Group assignment may also be specified in the set() call by using groups key entry.

For instance:

Groups are either specified by ID, name or a eZContentClassGroup object. As there is no persistent ID for groups it is often the best to use names.

Translations

Some fields of content-classes and their attributs may be translated, this is done by using the addTranslation() and removeTranslation().

The following content-class fields may be translated:

• name
• description

The following content-class attribute fields may be translated:

• name
• description
• data_text

For instance:

For direct manipulation there is also createTranslation() and deleteTranslation().

There's currently no support for translating custom content of data-types, this could later be solved by using a plugin system.

Data-Types

While it is is possible to create an attribute with any data-type not all of them are supported when it comes to setting additional class properties. All supported types are set by supplying the value parameter to addAttribute, the value is always an array with entries specific to the datatype. e.g. setting max 50 characters for an ezstring attribute.

ezstring

The maxium length can be set with the max parameter in value, it defaults to 0 (no limit).

The default value can be set with the default parameter in value, it defaults to '' (empty string).

ezboolean

The default value can be set with the default parameter in value, it defaults to 0 (unchecked).

ezinteger

The minimum value can be set with the min parameter in value, it defaults to null (no limit).

The maximum value can be set with the max parameter in value, it defaults to null (no limit).

The default value can be set with the default parameter in value, it defaults to null.

eztext

The column size can be set with the columns parameter in value, it defaults to 10.

ezxmltext

The column size can be set with the columns parameter in value, it defaults to 10.

The tag preset can be set with the tag_preset parameter in value.

ezimage

The max file size (in MB) can be set with the max_file_size parameter in value, it defaults to 0 (no limit).

ezobjectrelation

The selection type can be set with the selection_type parameter in value. A value of 0 means to browse for relation, 1 means dropdown.

The default selection node can be set with the default_selection_node parameter in value.

The fuzzy match can be turned on or off with the fuzzy_match parameter in value.

ezurl

The default value can be set with the default parameter in value, it defaults to '' (empty string).

ContentObject

Wrapper around eZContentObject which defines a content object with chosen attributes. First create a ContentType instance either with an identifier when using an existing content-class or create() new content-class. Then use that type instance to instantiate a new object instance.

Example, instantiating a content class:

Data-Types and object attributes

Most attribute content is set by supplying the $content parameter to setAttribute, this value will then be sent to the data-type as-is using the fromString method. However some data-types require special content values, the following are suppored.

ezxmltext

Content must either be imported from HTML or using the internal XML format.

To import HTML content use the Aplia\Content\HtmlText class, instantiate it with the HTML text and pass it as content. The system will then take care of converting it to the internal format. Note: ezxmltext expects the HTML to contain <p> tags for the main text content, so setting simple text will require it to be wrapped in the <p> tags.

To import XML content use the Aplia\Content\RawXmlText class, instantiate it with the text and optionally url object links, related object IDs or linked object IDs.

Example with HTML text:

ezselection

This expects the selection value to be the integer value for the key which is selected, the first selection is 0, the next 1 and so on.

ezselection2

Indentifiers can be specified the same way the datatype support fromString.

Or as array

ezimage

Currently only supports uploaded HTTP files. Create an instance of Aplia\Content\HttpImage with the name of the file entry in $_FILES.

For instance if the HTML form contained an <input type="file" name="portrait_image"> then the image file will be available in the file entry portrait_image.

Content Export/Import

Content-tools has support for exporting/importing subtrees across sites, while also preserving ownership, relations, and embedded content. Remember to not have mismatching package versions of this when exporting from one project, and importing to another. It might require the use of a config file in the import, which maps content in the export to content in the destination. See section "Configuration for import" for more info regarding this.

Content Export

Exporting content is done with bin/dump_content. At the time of writing, these are the current available options:

(vendor/bin/dump_content --help)

Script echoes, so pipe the result to file.

Examples

Here is a configuration which will export all objects of the given content_class, under the given path (container_path, corresponding to --parent-node-option). This export assumes the parent directory already exists on the import destination (i.e. remove --exclude-parent if the entire directory is to be imported).

(Values in chevrons ('<>') should be replaced with your own options.)

When setting file storage path, this folder should be located in the same folder as the export files. For example:

Content Import

Importing content is done with bin/import_content. At the time of writing, these are the current available options:

(vendor/bin/import_content --help)

The import logs through stdout, so pipe to file to save log output. For example use | tee import_log.txt. (Piping to tee might not show interactive questions, so be sure that the script is not waiting for question response, if no update is showing.)

The script might ask whether to remove object relation, or reset ownership, if the object in question neither exists in the target database, or in the import. (Big imports might prompt for this A LOT, which is why the --yes-parameter was added.)

NB! The content import starts a database transaction, and does not commit before the entire import is finished. This way, in case of errors, an unsuccessful import will rollback the database transaction, so that no wrong data is committed to the database. This means that while the import is running, publishing new content on the destination will not work, and will give a database transaction error. See section "Batch Export/Import" for examples on how to make content publishing possible while importing content.

Examples

NB! Batch Export/Import

By dividing the exported content into multiple files, we reduce the number of operations before content is committed to the database, and thereby make it possible to publish content.

(This is the only current support for this because of all the actions required for import of a given object. I.e.: 1. Creating node/object skeleton, 2. Creating object, 3. Updating the object with data. It is done this way to be able to preserve locations, relations and ownerships when importing objects which depend on other objects)

Examples

• Export: Given an existing export <export_name>.ndjson, the following bash script will divide the export into a .head file, which will be prepended to the numbered files. It takes the first occurence of ez_contentobject, and splits the following on 5 lines, which corresponds to 5 content objects. PS: This can generate a lot of files.

• Import:

Configuration for import

The content importer supports using a configuration file to define mapping of data and defining transformation code to be run.

Transformation classes

The import system supports custom transformation classes, these are classes that are instantiated with the importer as the first parameter and the ini file as second.

They should then contain methods for transform import data into new data. The specifics are explained in subsequent sections.

NodeMap

NodeMap allows for remapping the uuid of imported nodes to a new uuid. It maps from the original UUID to a new UUID

e.g.

This is used for nodes of objects, for parent uuids and owner uuid. A typical use case is to relocate a parent node to a new location.

ObjectMap

ObjectMap allows for remapping the uuid of imported objects to a new uuid. It maps from the original UUID to a new UUID.

e.g.

Object transformation

Object import data may be transformed before the import system takes care of it. This can either be done on a specific object uuid, class identifier or for all objects.

To transform a single object specify TransformByUuid.

To transform objects of given class specify TransformByClass.

Transformation for all objects are specified with Transform

All object transformation classes must implement the ContentObjectTransformation interface.

Content input handler

Content-tools supports extending support for 3rd-party data-types by defining input handlers. The input handlers has the responsibility to take the input value and store it on the attribute in whatever manner makes sense. It may also transform the input value to the format that the data-type expects.

The handler is a class that implements the ContentInputHandler interface and must be defined in content.ini under DataTypeSettings and the variable ContentInputHandler. For instance lets say we want to support ezselection then add

The handler must then implement the storeContent method and store the value on the attribute or in the database. A simple example follows:

If the input value is simply text, integer, float or a string-like input then swap out class for the corresponding types text, int, float and string. To ignore the input value use type ignore. string means to pass the value to the fromString method on the data-type.

Migration

This package has support for the Phinx system for migrations of database content and other data. First install Phinx using composer.

If you get issues with symfony/console try adding symfony/console:^3.0 as a requirement.

Next sphinx requires a configuration file, the easiest way is to create the sphinx.php file and make use of the Aplia\Migration\Configuration class to get automatic setup based on the eZ publish site.

Create the file and add:

and add it to git.

The configuration assumes all migrations are located in extension/site/migrations. If you have an older site with a different extension name for the site the path must be configured in project.ini, add the following:

Or if you have migrated to using namespaces, add the following (multiple namespaced paths can be added if needed):

Then use vendor/bin/phinx to handle migrations, remember to add -c phinx.php to after all commands. For instance to see the current status.

To create a new migration file use create, for instance:

To run migration use migrate command, for instance:

To migrate eZ publish content the classes ContentType and ContentObject may be used inside the migrations.

Content class export

This package has support for dumping content classes by given identifiers to a runnable script. The script can be run as a migrate-style PHP snippet.

To run this script manually:

Output goes to stdout, so pipe to a file to save it.

Options

identifier

Specify which content class identifiers to export. It supports multiple values. E.g.: folder article.

php-style

Set to migrate, to use migrate style. E.g.: php-style=migrate

preamble

Set to include boilerplate script options before the class definitions. E.g.: preamble.

delete-existing

Include this to delete the existing content classes instead of updating existing.

update-class-group

Use this to reset the class group on all exported classes to something of your choice. E.g.: update-class-group=Seksjoner.

NB! Create this class group beforehand.

update-creator-id

Use this to set a user by object id as creator for the content classes. E.g.: update-creator-id=14 to set creator as admin. (This uses eZUser::setCurrentlyLoggedInUser().)

Example

Assuming location is project root:

Utilities

Utility classes using content-tools can be found in Aplia/Utilities and the scripts can be found in bin.

Search and Replace

This utility can be used to either search for all occurrences of the given search string in all objects, or also replace the occurrence with the given replace string. The script will print the path of the matched occurrences. It uses ContentObject class to update every object.

Search

Assuming project root, run php vendor/bin/ezp_search_and_replace <search_string> to search for objects containing the given search_string in their attributes data_text.

Replace

Assuming project root, run php vendor/bin/ezp_search_and_replace <search_string> <replace_string> --replace to replace the occurences of search_string with replace_string. It might be wise to first run search, and then append the replace string after verifying all objects that will be changed. Pass the --ignore-parameter, with object ids separated by comma, to ignore specific objects (e.g. --ignore=<objectId1>,<objectId2>,<objectId3>).

NB! Some attributes might fail to update. For example eZImage. If they do, the objects are simply not updated, and they are printed after running the script, for manual update.

Other parameters

• --print-urls: Whether to print urls of the objects. If not set, path_identification_string will be printed for every occurrence.
• --case-insensitive: Case insensitive search. (The sql uses REGEXP on BINARY. Setting this omits BINARY from the sql.)
• --new-version: Whether to publish a new version of updated object as admin. This might return template errors which can be ignored. This parameter presumes Admin user is object id 14.

Run php vendor/bin/ezp_search_and_replace --help for a complete overview.

Reindexing objects of migrated content class

The following is a SQL script for adding content objects of a given class to the index queue table. For example when using eZFind/Solr. (The script for running this is php runcronjobs.php -s <siteaccess> indexcontent) This can be added as a migration. (Note: A future version of this could replace this with a flag on the content class update, which queues the class for indexing.)