Commit Graph

139 Commits

Author SHA1 Message Date
Kerollmops
4e86ecf807
Retrieve the words before the intersect loops 2020-06-10 22:05:01 +02:00
Kerollmops
6ca3579cc0
Add more time debug measurements 2020-06-10 21:35:01 +02:00
Kerollmops
66a4b26811
Introduce a proximity based documents retriever 2020-06-10 16:54:28 +02:00
Kerollmops
78f27c0465
squash-me: Remove debugs 2020-06-10 16:29:46 +02:00
Kerollmops
3ad883d7c7
squash-me: Make the dijkstra work even with different attributes 2020-06-10 16:27:02 +02:00
Kerollmops
fecd8ca54a
squash-me: It works! we must remove the debug after having added more tests 2020-06-10 14:20:35 +02:00
Kerollmops
13977d9338
squash-me 2020-06-09 23:06:59 +02:00
Kerollmops
5d5b827f1a
Squash-me 2020-06-09 17:32:25 +02:00
Kerollmops
2a6d6a7f69
Introduce a first draft of the best_proximity algorithm 2020-06-09 10:11:43 +02:00
Kerollmops
dfdaceb410
Introduce a first basic working positions-based engine 2020-06-05 20:13:19 +02:00
Kerollmops
f51a63e4ef
Store documents ids under attribute ids 2020-06-05 16:32:14 +02:00
Kerollmops
ce86a43779
Make the query tokenizer a real Iterator 2020-06-05 09:49:28 +02:00
Kerollmops
f55f4cb02a
Not fetch the cached prefix postings when prefix is disabled 2020-06-04 21:22:45 +02:00
Kerollmops
06bf03f075
Add an help message on the front page
aaa
2020-06-04 21:22:45 +02:00
Kerollmops
eefc6d7c44
Add support for quoted query phrases 2020-06-04 20:25:51 +02:00
Kerollmops
1f7035f18f
Just do a little clean-up 2020-06-04 19:13:28 +02:00
Kerollmops
71dc6a3828
Disable prefix search when query is ended by a whitespace 2020-06-04 18:37:20 +02:00
Kerollmops
5d1c625b74
Change the page index texts 2020-06-04 18:20:57 +02:00
Kerollmops
c42d3c19e2
Merge the whole list of generated MTBL in one go 2020-06-04 17:38:43 +02:00
Kerollmops
3a23dc242e
More efficiently merge MTBLs, more than two at a time 2020-06-04 16:17:24 +02:00
Kerollmops
1df1f88fe1
Directly write to LMDB without intermediate final MTBL 2020-06-01 21:30:39 +02:00
Kerollmops
2174042994
Merge only 3 MTBL at the same time 2020-06-01 19:49:58 +02:00
Kerollmops
5cc81a0179
Merge many MTBL into one a the same time 2020-06-01 18:39:58 +02:00
Kerollmops
6a047519f6
Do a merge two by two 2020-06-01 18:27:26 +02:00
Kerollmops
5404776f7a
Add a little bit more debug 2020-06-01 17:52:43 +02:00
Kerollmops
dff68a339a
Use OnceCell to cache levenshtein builders 2020-05-31 19:27:11 +02:00
Kerollmops
dde3e01a59
Introduce prefix postings ids for better perfs 2020-05-31 18:20:49 +02:00
Kerollmops
a26553c90a
Reintroduce a simple HTTP server 2020-05-31 17:48:13 +02:00
Kerollmops
2a10b2275e
Support prefix typo tolerant search 2020-05-31 17:18:13 +02:00
Kerollmops
ba9527abc0
Support typos with a levenshtein automata 2020-05-31 17:01:11 +02:00
Kerollmops
6c726df9b9
Support multiple space seperated words 2020-05-31 16:09:34 +02:00
Kerollmops
24587148fd
Introduce MTBL parallel merging before LMDB writing 2020-05-31 14:22:57 +02:00
Kerollmops
6762c2d08f
Clean up a little bit 2020-05-31 14:22:57 +02:00
Kerollmops
3a998cf39c
Far better usage of rayon to fold indexed data 2020-05-31 14:22:57 +02:00
Kerollmops
1237306ca8
Introduce a thread that write to heed 2020-05-31 14:22:57 +02:00
Kerollmops
3668627e03
Use zerocopy without bitpacking as a first step 2020-05-31 14:22:07 +02:00
Kerollmops
a81f201fad
Inroduce the use of RocksDB instead of sled (RAM) 2020-05-31 14:22:06 +02:00
Kerollmops
91ba938953
Initial commit 2020-05-31 14:22:06 +02:00
Clément Renault
4573f00a0d
Initial commit 2020-05-31 14:21:56 +02:00