bottom

GEDIS Studio's Random Text Generator

Our random text generator produces text using your own document templates, with sentences made from real words and your own custom sentence templates.

The random text generator is one of the sample data generator projects available from our GEDIS Studio sandbox server. These sample projects are intended to demonstrate the features of this online test data generator and to help you to start designing your own generators.

Quick start :

  • If not already done, register on GEDIS Studio OnLine
  • Got ot the Workspace manager and select the Project tab
  • Locate the GenieLog's shared workspace
  • Click twice on the project named "Random Text" to import a copy into you own workspace

Now you have a copy of the our random text generator project into your own private workspace. The customization of generated text is acheived in various ways. These are essentialy by defining

  • the structure of the document to be generated
    (document template),
  • the structures of the sentences to be instanciated in the document
    (sentences templates file)
  • the list of words associated to each type of words referenced in a sentence template

Create a document template file named after the parameter INPUT_DOC. This is acheived by going to the Data tab in your workspace manager, click on the add file, put your template text here. Then change the default name of that file to what is defined for INPUT_DOC.

Create a sentence templates file named according the the parameter SENTENCES. Proceed like for the document template file and write a single sentence template per line.

Why should you look at this text generator?

It is appropriate for generating well formatted pages from a predefined structure (title, headings, etc.), using real words from selected dictionary files and a list of predefined sentence patterns.

Moreover, our random text generator allows you to control the density of specific keywords inserted in place of given tags in the location of your choice such as headings, titles, etc. We have designed this generator having in mind that the quality of a random text depends on:

  1. The choice of the words that make the text,
  2. The control you have on specific keywords usage frequency,
  3. The quality of the sentence structure,
  4. The control of the overall document structure,

Most text generators on the web are intended to fill pages for testing web sites. The best ones allow you to select a target language and produce text by assembling words from dictionary files. Our intent is to go further both in the direction of more realism and more control over the structure of the generated text.

Our random text generator enables you to

  • define your own document template (title, headings, paragraphs, ...)
  • define various sentence structures (based on article, noun, verbs, ...)
  • define words to be chosen amongst for each type word in a sentence balancing between regular words and special keywords with an augmented frequency
  • Chose which sentence structure to use in each document template areas

A Text Generator Based on Sentence and Document Templates

The principle of this text generator is to parse a template document including specific tags. Each tag is associated with a single word or a sequence of words. For example, the tag [NOUN] is associated with a single noun word, while [SENTENCE] or [PARAGRAPH] are associated with a sequence of words.

When the text generator parses the input document it replaces each occurrence of a tag with the appropriate word extracted from a source data file (nouns, verbs, adjectives, etc.)

You can provide several sentence templates that will be randomly selected each time you refer to the tag SENTENCE or PARAGRAPH. Below is an example of two sentence templates:

[Article] [Noun] [Verb] [Article] [Adjective] [Noun]
The [NOUN] [VERB] all the [NOUN]s

Which can produce sentences like

The cat look at all red book   (instantiates sentence template 1)
The dog spare all the books (instantiates sentence template 2)
All man paint each tall cat (instantiates sentence template 1)

Obviously, you can extend at will this text generator at will by adding more sentence structures, more words, or even more word types, to produce semantically more complex sentences.