perparagraph

Perparagraph is a term used in document processing and text analysis to describe an operation that is applied to each paragraph in a document independently. It is not a formal standard, but a descriptive shorthand used in software documentation and academic writing to contrast per-paragraph processing with global transformations applied to the whole text. It is sometimes written as per-paragraph or used as perparagraph in code and discussions.

Common uses include formatting per paragraph (such as indentation, spacing, or alignment), normalization (trimming whitespace, ensuring consistent paragraph breaks), metadata extraction (counting words per paragraph, detecting sentiment per paragraph), or applying transformations (translation, summarization) on a per-paragraph basis.

Implementation typically involves splitting the document into paragraphs, applying a function to each paragraph, and then recombining the results. In programming, this is a map-like operation over a list or sequence of paragraph strings. In HTML contexts, a perparagraph operation may map to applying styles or attributes to each paragraph element independently.

Considerations include accurately identifying paragraph boundaries, especially in formats where line breaks do not directly map to paragraphs (for example, markup languages or rich text). It also matters whether transformations should be idempotent, how to handle empty paragraphs, and performance on long documents.

Related ideas include per-sentence processing, per-word processing, and global text processing. The term perparagraph emphasizes per-unit handling, enabling granular control over formatting, analysis, and transformation tasks.