String Containers

String containers essentially orchestrate the processing pipeline, cache results of expensive computations, and provide an interface for the user to engage with analysis results.

Given that we perform most analysis on a sentence-level while most texts consist of more than a single sentence, LIMES provides two types of string containers, the Text and the Sentence container.

Text

Python Implementation

Refer to the API documentation for the Text class for more information.

The Text container provides logic for generating Sentence containers for a given text. It uses the Parser to perform token-level analysis on the given text and to sentencize the text, creating Sentence objects for each sentence.

The user-provided string is processed lazily, i.e. the raw string is only operated on once the user calls a function that requires analysis results. This has the benefit that the objects don't take forever to create and don't take up too much memory but it has the downside that you won't be notified if the provided string has issues (e.g. if it is too unclean for the parser to produce meaningful results) on object creation.

Sentence

Python Implementation

Refer to the API documentation for the Sentence class for more information.

The Sentence container allows sentence-level evaluation of your text, as well as caching analysis results.

Caching

Both Barriers and text complexities are computed lazily, i.e. only once the user first requests information about them. Keep in mind that this means that - once they are computed - analysis results stay in memory until the whole Sentence object is garbage-collected.