Cache Sentence Tokenization in Co-Citations #2
Description
Cache Sentence Tokenization in Co-Citations (High Value, Low Risk)
File/Line: MyGraph.ts:220-235
Description: In the co-citations algorithm, cachedRead and tokenizer.sentences() are called for each backlink file. Cache tokenized sentences per file path in a Map within the MyGraph class, computed lazily on first access.
Why: Sentence tokenization is CPU-intensive and repeated for files with multiple backlinks. Caching reduces redundant parsing, speeding up co-citations (the most complex algorithm) for notes with many references. Low risk; uses existing Obsidian caching patterns.