mirror of
https://github.com/MarginaliaSearch/MarginaliaSearch.git
synced 2025-02-23 21:18:58 +00:00
3101b74580
This is a system-wide change. The index used to have a lexicon, mapping words to wordIds using a large in-memory hash table. This made index-construction easier, but it also added a fairly significant RAM penalty to both the index service and the loader. The new design moves to 64 bit word identifiers calculated using the murmur hash of the keyword, and an index construction based on merging smaller indices. It also became necessary half-way through to upgrade guice as its error reporting wasn't *quite* compatible with JDK20. |
||
|---|---|---|
| .. | ||
| array | ||
| big-string | ||
| braille-block-punch-cards | ||
| btree | ||
| easy-lsh | ||
| guarded-regex | ||
| language-processing | ||
| message-queue | ||
| next-prime | ||
| random-write-funnel | ||
| term-frequency-dict | ||
| LICENSE.txt | ||
| readme.md | ||
Libraries
These are libraries that are not strongly coupled to the search engine's business logic. These libraries may not depend on features, services, processes, models, etc.
NOTE: These libraries are co-licensed under the MIT license.
Libraries
- The array library is for memory mapping large memory-areas, which Java has bad support for. It's designed to be able to easily replaced when Java's Foreign Function And Memory API is released.
- The btree library offers a static BTree implementation based on the array library.
- language-processing contains primitives for sentence extraction and POS-tagging.
- The message-queue library.
Micro libraries
- easy-lsh is a simple locality-sensitive hash for document deduplication
- guarded-regex makes predicated regular expressions clearer
- big-string offers seamless string compression
- random-write-funnel is a tool for reducing write amplification when constructing large files out of order.
- next-prime naive brute force prime sieve.
- braille-block-punch-cards renders bit masks into human-readable dot matrices using the braille block.