A lightning intro to re-Isearch
summary
re-Isearch is a C++-based multimodal search and retrieval engine (with Python, Tcl, Java bindings) that uses mathematical models instead of an inverted index to permit unlimited term frequency/length, overlapping fields, and complex structured data. It focuses on flexible retrieval granularity by exploiting document structure (XML/visual groupings) so queries can return sub-document units (lines, paragraphs, chapters) rather than whole documents, and supports multiple object types (numerical, range, geospatial, ISO‑8601 date semantics) searchable both as their native types and as text.
AI-written summary · about
related talks
27mLightweight Semantics Engineering with Redex
15mREPTILE: Performant Tiling of Recurrences
16mIntrepydd: Performance, Productivity, and Portability for Data Science Application Kernels
22mEREQ: Regular Expressions with Quantifiers and Incremental Quantifier Elimination
05mInterview
46mWhat is PL Research?
47mReflection Techniques in C++
43mPriority Search Queues: 1.5 dimensional Tree Search