web search done right
Commits on main
[tantivy] kmerge posting lists to avoid iterating over documents that are not present in the posting list during segment merges, while preserving low memory footprint
54ac58f Mikkel Denker committed 2 years ago
[search_server] implement top key phrases as generic query
aba861e Mikkel Denker committed 2 years ago
[search_server] move some function bounnds into the traits for simplicity
ebacff4 Mikkel Denker committed 2 years ago
[search_server] add a 'GenericQuery' trait so the search index can handle queries that are not typical search queries
1a3a9cb Mikkel Denker committed 2 years ago
[search_server] use spawn_blocking for search requests to improve concurrency
b93afff Mikkel Denker committed 2 years ago
[webgraph] speedup queries by using warmed f64 column for scorer
cd40e27 Mikkel Denker committed 2 years ago
webgraph spawn_blocking to handle batch requests concurrently
8567b8e Mikkel Denker committed 2 years ago
take extra edges to ensure the remote has enough for deduplication
534ecd1 Mikkel Denker committed 2 years ago
take extra edges to ensure the remote has enough for deduplication
9083d60 Mikkel Denker committed 2 years ago
host group sketch query to compute the in/out degree somewhat accurately
fcf2b99 Mikkel Denker committed 2 years ago
convert RelFlags from a bitset to a vec of enums for public api
3763c51 Mikkel Denker committed 2 years ago
if the first edge in the postings was an edge from a host to itself, it would not get properly filtered
32c4a5f Mikkel Denker committed 2 years ago
deduplicate host edges across entire query, not just adjacent documents
ba01fc7 Mikkel Denker committed 2 years ago
edges get sorted by their segment id for performance when retrieving. assigning sort_scores with .zip would therefore be incorrect
91544e3 Mikkel Denker committed 2 years ago
re-use u64 column fields in webgraph queries as a significant fraction of the time was spend opening the columns
c9a1d6f Mikkel Denker committed 2 years ago