@startuml Plain Text Plugin Architecture
!include shared_interfaces.puml
!include shared_utilities.puml
package "Plain Text Plugin" {
class PlainTextPlugin {
- nlpProcessor: NLPProcessor
- paragraphDetector: ParagraphDetector
- topicExtractor: TopicExtractor
- sentenceSplitter: SentenceSplitter
- semanticChunker: SemanticChunker
+ indexFile(path, content): IndexShard
+ supports(path): bool
+ search(query, opts): SearchResult[]
+ processNaturalLanguage(content): ProcessedText
}
class NLPProcessor {
- tokenizer: Tokenizer
- posTagger: POSTagger
+ tokenize(text): Token[]
+ extractKeywords(text): Keyword[]
+ detectLanguage(text): string
+ analyzeComplexity(text): ComplexityScore
}
class ParagraphDetector {
- minParagraphLength: int
- paragraphSeparators: string[]
+ detectParagraphs(text): Paragraph[]
+ inferStructure(paragraphs): DocumentStructure
+ groupRelatedParagraphs(paragraphs): ParagraphGroup[]
}
class TopicExtractor {
- topicModel: TopicModel
- minTopicCoherence: float
+ extractTopics(text): Topic[]
+ findMainThemes(paragraphs): Theme[]
+ calculateTopicSimilarity(t1, t2): float
+ groupByTopic(paragraphs): TopicGroup[]
}
class SentenceSplitter {
- abbreviations: Set<string>
- sentenceEnders: string[]
+ splitSentences(text): Sentence[]
+ detectBoundaries(text): int[]
+ handleAbbreviations(text): string
+ preserveFormatting(sentences): Sentence[]
}
class SemanticChunker {
- targetChunkSize: int
- coherenceThreshold: float
+ createSemanticChunks(paragraphs): Chunk[]
+ measureCoherence(p1, p2): float
+ optimizeChunkBoundaries(chunks): Chunk[]
+ addOverlap(chunks): Chunk[]
}
class ProcessedText {
+ paragraphs: Paragraph[]
+ sentences: Sentence[]
+ topics: Topic[]
+ chunks: Chunk[]
+ metadata: TextMetadata
}
class Paragraph {
+ id: string
+ content: string
+ sentences: Sentence[]
+ topic: Topic
+ startPos: int
+ endPos: int
}
class TextChunk {
+ id: string
+ content: string
+ paragraphs: Paragraph[]
+ coherenceScore: float
+ metadata: ChunkMetadata
}
}
' Relationships
PlainTextPlugin --> NLPProcessor : uses
PlainTextPlugin --> ParagraphDetector : uses
PlainTextPlugin --> TopicExtractor : uses
PlainTextPlugin --> SentenceSplitter : uses
PlainTextPlugin --> SemanticChunker : uses
PlainTextPlugin --|> BaseDocumentPlugin
ParagraphDetector --> Paragraph : creates
SemanticChunker --> TextChunk : creates
PlainTextPlugin --> ProcessedText : produces
' External dependencies
PlainTextPlugin ..> IPlugin : implements
PlainTextPlugin ..> SemanticIndexer : integrates
NLPProcessor ..> "External NLP Library" : may use
@enduml
