Unsequenced In Document

Save PDF

Last Updated: May 13, 2026
3 minute read

Semaphore
Documentation

Summary of usage

The Unsequenced In Document Context looks for a fact and / or anchor in the same document. If they are, then the extractor fires and will return the fact or facts.

This is a very simple extractor as it applies practically no constraint on the child elements - apart from the fact of occurring in the same document - which, given how content is sent to CS, is always a given.

Exemplification

The TTL format of an example model can be found here: UnsequencedInDocument.excludeResultsOverlappedBy

This model has an anchor (a disease) and a fact (a person entity) that should appear in the same document, and where the person entity should not be returned if they are a doctor (an person entity with the phrase “Dr” before them).

Testing this model with:

this content:

          Dr Foster said Peter Smith had Paget Disease.

gives this response:

      <?xml version="1.0" encoding="UTF-8"?>
      <response>
       <STRUCTUREDDOCUMENT>
        <META name="Type" value="TEXT (4003)"/>
        <SYSTEM name="DeterminedLanguage" value="english"/>
        <SYSTEM name="LanguageGuessed" value="no"/>
        <META name="DocumentType" value="Doc" score="1.00"/>
        <META name="Mentioned patient" value="Peter Smith" score="1.00"/>
       </STRUCTUREDDOCUMENT>
      </response>

this content:

          Dr Foster examined Peter Smith.

          He concluded the patient had Paget Disease.

gives this response:

      <?xml version="1.0" encoding="UTF-8"?>
      <response>
       <STRUCTUREDDOCUMENT>
        <META name="Type" value="TEXT (4003)"/>
        <SYSTEM name="DeterminedLanguage" value="english"/>
        <SYSTEM name="LanguageGuessed" value="no"/>
        <META name="DocumentType" value="Doc" score="1.00"/>
        <META name="Mentioned patient" value="Peter Smith" score="1.00"/>
       </STRUCTUREDDOCUMENT>
      </response>

this content:

          Dr Foster said Peter Smith was sick.

          He had Paget Disease.

gives this response:

      <?xml version="1.0" encoding="UTF-8"?>
      <response>
       <STRUCTUREDDOCUMENT>
        <META name="Type" value="TEXT (4003)"/>
        <SYSTEM name="DeterminedLanguage" value="english"/>
        <SYSTEM name="LanguageGuessed" value="no"/>
        <META name="DocumentType" value="Doc" score="1.00"/>
        <META name="Mentioned patient" value="Peter Smith" score="1.00"/>
       </STRUCTUREDDOCUMENT>
      </response>

this content:

          Peter Smith had Paget Disease.

gives this response:

      <?xml version="1.0" encoding="UTF-8"?>
      <response>
       <STRUCTUREDDOCUMENT>
        <META name="Type" value="TEXT (4003)"/>
        <SYSTEM name="DeterminedLanguage" value="english"/>
        <SYSTEM name="LanguageGuessed" value="no"/>
        <META name="DocumentType" value="Doc" score="1.00"/>
        <META name="Mentioned patient" value="Peter Smith" score="1.00"/>
       </STRUCTUREDDOCUMENT>
      </response>

this content:

          Dr Peter Smith had Paget Disease.

gives this response:

      <?xml version="1.0" encoding="UTF-8"?>
      <response>
       <STRUCTUREDDOCUMENT>
        <META name="Type" value="TEXT (4003)"/>
        <SYSTEM name="DeterminedLanguage" value="english"/>
        <SYSTEM name="LanguageGuessed" value="no"/>
        <META name="DocumentType" value="Doc" score="1.00"/>
       </STRUCTUREDDOCUMENT>
      </response>

Full unit tests (models and content) for Unsequenced in Document

Find all models for Unsequenced in Document here: Unsequenced in Document unit test models

and the associated tests here:

If you wish to test, you will need to:

create a new model and then import the test TTL model file into it
copy and paste the content text from the model’s associated test into DA and classify