Overview of Paragraph Recognition
For Congree to know which contents need to be checked, not be checked, or be treated in a certain way, you need to configure Paragraph Recognition in the document settings of the Congree Control Center.
You can choose between automatic and manually configured Paragraph Recognition.
Important:
Congree paragraph recognition does not support character formats used for marking structural information.
Automatic Paragraph Recognition
Nothing can be configured here. Congree creates a standard DTD interpretation that recognizes paragraphs and the presence of inline elements in your text.
Manual configuration of Paragraph Recognition
Paragraph recognition can be configured in the Paragraph Recognition area of the document settings:
Treatment: What options are there?
The treatment determines how Congree is to handle a recognized element. All available options are explained below.
Check
All elements whose content is to be checked must be stored as check unit with Check treatment. The treatment will be passed down to child elements, except for child elements assigned to Skip treatment.
Skip
You can generally simply not list all elements that are to be skipped (i.e. that are not to be checked) in the table.
However, as Check treatment is inherited by child elements, child elements of Check elements that are to be skipped must be assigned to Skip treatment. The Skip treatment will be passed down to child elements and overwrite the Check treatment.
Check or skip depending on parent
If the treatment of an element is to depend on its parent element, you need to enter the parent element under Parents. If the parent element, in turn, is to have a certain parent element, use a forward slash (example: the parent element <description> is defined within its own parent element <cause>: specify cause/description).
Check or skip depending on attribute
If a check unit is to be checked or skipped only if it has a certain attribute, you need to enter the attribute under Attributes. Attributes can be combined with AND and OR operators. However, it is not possible to use both operators in one cell.
Instead of using the OR operator, you can also create a new row in which you can store the same element with a different attribute. This method is especially useful if you have many attributes.
Nested
If a check unit is embedded in another check unit, but is to be treated as a standalone element, you need to assign it the Nested treatment.
Note:
The Nested treatment only applies to the Congree Language Check.
Classes
You can assign each check unit to a particular Congree class. In this way, special rules will be applied to the check unit. For example, these special rules may include a specific Authoring Memory penalty or an activated or deactivated style rule.
Possible classes:
- List
- Title
- Table
- Procedure
- Hint
- Warning
- Literal
- Regular
If your system has further Congree classes that you need e.g. for the Authoring Memory (structure attributes), you can enter these as well. However, this will not affect your Congree Language Check. Congree will treat all unknown classes as "Regular".
Checking or not checking certain rules
If certain rules of your Style Guide are to be applied or not be applied to a certain check unit, you can configure these under Rules.
There are three different settings for the individual rules:
- Empty selection: The rule will not be applied.
- Checkmarked selection: The rule will be applied for this check unit. For this, it must be activated in the underlying Style Guide. This setting is only useful for customer-specific rules.
- Filled out selection: The standard behavior determined by the Style Guide will apply.
Important:
Not every behavior is possible with every rule. There are rules where you can only choose between an empty and a filled out selection.
There are also rules for which only checkmarked selection and filled out selection out are available. This means that you cannot configure that the rules are not applied here.
Note:
This setting applies to the Congree Language Check only.
Configuring inline elements
Treatment: What options are there?
Normal
Inline elements with Normal treatment will be interpreted as part of the sentence. This means that their content will be checked by Congree. Inline elements with Normal treatment will not segment the sentence.
Ignore content
Inline elements with Ignore content treatment are converted to empty elements and thus treated as placeholders. Congree knows that the inline element is write-protected and cannot be removed.
Ignore content - protected
Inline elements with Ignore content treatment and enabled Protected selection are converted to empty elements and thus treated as protected placeholders.
Ignore tag and content
Inline elements with Ignore tag and content treatment are treated by Congree as if they do not exist.
Ignore tag
For inline elements with the Ignore tag treatment, the inline element is ignored, but not its content.
Nested
If you want inline elements to be checked individually and also be interpreted as correct sentence elements, you need to assign them Nested treatment.
Note:
This setting applies to the Congree Language Check only.
Parent-dependent treatment
If the treatment of an inline element is to depend on its parent element, you need to enter the parent element under Parents.
If the parent element, in turn, is to have a certain parent element, use a backslash (example: cause\description).
Important:
Attribute-dependent treatment
If an inline element is to be treated specifically only if it has a certain attribute, you need to enter the attribute under Attributes. Attributes can be combined with the AND and OR operators. However, it is not possible to use both operators in one cell.
Instead of using the OR operator, you can also create a new row in which you can store the same element with a different attribute. This method is especially useful if you have many attributes.
Important:
Customizations for the Language Check
Important:
If Language Check specifics have to be defined, the respective column will be completed by Congree Language Technologies GmbH.
A use case for Language Check specifics is that you want different Paragraph Recognition behavior for Language Check and Authoring Memory.