Commit Graph

202 Commits

Author SHA1 Message Date
Clément Renault
45d4361e7d
Correctly highlight when query string is longer 2019-10-30 17:49:50 +01:00
Clément Renault
b709a7a30a
Bump the meili-core/schema/tokenizer crates to 0.5.11 2019-10-30 17:40:31 +01:00
Clément Renault
c230f244be
Make the highlight system much better 2019-10-30 17:32:29 +01:00
Clément Renault
4dff8a215e
Reduce the ram consumption when re-indexing all the documents 2019-10-29 17:46:23 +01:00
Clément Renault
e9dce3ce81 Add a test to ensure that the indexer support stop words 2019-10-29 16:18:06 +01:00
Clément Renault
ff7dde7522 Make the RawIndexer support stop words 2019-10-29 16:18:06 +01:00
Clément Renault
a226fd23c3 Introduce the stop words deletion update type 2019-10-29 16:18:06 +01:00
Clément Renault
776673ebae
Introduce the stop words addition update type 2019-10-29 15:24:09 +01:00
Clément Renault
8a17fcdda5
Introduce a function to get all updates results 2019-10-29 11:37:40 +01:00
Clément Renault
ac12a4b9c9
Make documents additions accept only the last duplicate document 2019-10-28 20:40:33 +01:00
Clément Renault
a43b37dfc1
Send channel notification when clearing documents 2019-10-28 17:58:22 +01:00
Clément Renault
c08dcac1d4
Abort the update transaction before calling the update callback 2019-10-28 17:55:43 +01:00
Clément Renault
9a57cab3ee
Fix the exactness criterion algorithm 2019-10-26 18:34:40 +02:00
Clément Renault
4111b99a6d
Only highlight query words areas not the whole words 2019-10-26 15:56:34 +02:00
Clément Renault
cb5c77e536
Make sure that automatons group with more automatons are better 2019-10-24 15:18:53 +02:00
Clément Renault
047eba3ff3
Introduce a clear all documents update 2019-10-23 16:39:10 +02:00
Clément Renault
eb67195840
Commit an update only when it is Ok 2019-10-23 15:52:40 +02:00
Clément Renault
7d9cf8d713
Clean up the fetch algorithm 2019-10-23 12:06:21 +02:00
Clément Renault
03eb7898e7
Introduce a basic working version of phrase query for splitting words 2019-10-23 11:40:13 +02:00
Clément Renault
858bf359b8
Make possible to convert recursive object into strings 2019-10-22 16:02:02 +02:00
Clément Renault
0f30a221fa
Introduce the reindex_all_documents indexing function 2019-10-22 14:07:27 +02:00
Clément Renault
78535b3e33
Move to heed 0.1.0 2019-10-21 12:05:53 +02:00
Clément Renault
9dce41ed6b
Cargo clippy pass 2019-10-18 13:30:06 +02:00
Clément Renault
ca26a0f2e4
Cargo fmt pass 2019-10-18 13:30:06 +02:00
Clément Renault
2ef51f7df9
Introduce a function to copy and compact a database env 2019-10-18 12:56:56 +02:00
Clément Renault
526202ec8b
Make the levenshtein algorithm consider transpositions to cost 1 2019-10-17 18:07:15 +02:00
Clément Renault
dd74af4c70
Use an RoTxn to retrieve the number of documents 2019-10-17 17:30:54 +02:00
Clément Renault
d941c512db
Rework the QueryBuilder to make it easier to construct and use 2019-10-17 14:45:21 +02:00
Clément Renault
2ea3e9b081 Improve the automaton producer quality by changing the production order 2019-10-17 13:19:08 +02:00
Clément Renault
da71821204 Make the example take the fetch-timeout-ms argument into account 2019-10-17 13:19:08 +02:00
Clément Renault
1cf6afad9a
Fix updates serialization to use serde_json instead of bincode 2019-10-17 12:31:46 +02:00
Clément Renault
1667e1b32f
Move to zerocopy-lmdb 2019-10-16 17:12:55 +02:00
Clément Renault
5e8d432614
Introduce an Iterator to visit all documents attributes counts 2019-10-15 15:27:18 +02:00
Clément Renault
3278d22279
Introduce an Index mathod to retrieve the currently processed update 2019-10-15 14:54:52 +02:00
Clément Renault
0de37819b4
Simplify the document fields counts deletion 2019-10-15 11:17:23 +02:00
Clément Renault
9ff92c5d15
Update the exact criterion to use the documents fields counts 2019-10-14 18:48:54 +02:00
Clément Renault
e629f51af4
Use the documents_fileds_count store in the QueryBuilder 2019-10-14 18:48:32 +02:00
Clément Renault
b377003192
Compute and store the number of words in documents fields 2019-10-14 14:07:10 +02:00
Clément Renault
a7e40a78c1
Introduce the DocumentsFieldsCounts store 2019-10-14 14:06:34 +02:00
Clément Renault
9cdda8c46a
Make the RawIndexer index_text method return the number of words 2019-10-14 13:56:52 +02:00
Clément Renault
710ab2386c
Get ride of rust messagepack (rmp) 2019-10-11 16:17:37 +02:00
Clément Renault
02575a2ef6
Introduce customs updates 2019-10-11 15:33:35 +02:00
Clément Renault
da6ab2753e
Rename Update/Type SchemaUpdate into Schema 2019-10-11 13:49:17 +02:00
Clément Renault
12b80e08be
Do not commit updates, let the user do 2019-10-11 11:29:47 +02:00
Clément Renault
9dca18f966
Fix the DocumentsIdsIter and do not iter on an Option 2019-10-10 18:32:22 +02:00
Clément Renault
9eb27811b1
Remove the useless generic documents_deletion parameter 2019-10-10 16:16:53 +02:00
Clément Renault
485480560a
Add method to create a query builder along with criterion 2019-10-10 15:32:08 +02:00
Clément Renault
e09d3b654d
Rename main_store into common_store 2019-10-10 15:22:23 +02:00
Clément Renault
19c22a8c5e
Create two open and create index functions 2019-10-10 13:48:30 +02:00
Clément Renault
7b26bd88c0
Improve the DocumentsIdsIter internal 2019-10-10 13:40:18 +02:00
Clément Renault
d1e59be46b
Introduce the DocumentsIds iterator 2019-10-10 10:35:57 +02:00
Clément Renault
46c19dfc5a
Expose the UpdateType 2019-10-10 10:24:41 +02:00
Clément Renault
d8fdad1455
Make the timeout QueryBuilder setting optional to and pass the tests 2019-10-09 17:59:31 +02:00
Clément Renault
f56636e1e9
Merge branch 'moving-to-lmdb' 2019-10-09 17:23:48 +02:00
Clément Renault
8f63ec39da
Unrestrict static lifetime of Criterion names 2019-10-09 16:15:31 +02:00
Clément Renault
5a1c1aeb02
Reintroduce the sort-by-attr criterion 2019-10-09 16:08:30 +02:00
Clément Renault
6ec575f8de
Use a buffered sync channel to avoid blocking the update system 2019-10-09 15:49:35 +02:00
Clément Renault
683b6afbfb
Introduce a way to filter documents with a basic syntax 2019-10-09 14:20:37 +02:00
Clément Renault
663714bb6d
Make the example return documents field in a consistent order 2019-10-09 13:48:33 +02:00
Clément Renault
bb35ca0d40
Reintroduce the distinct and filtering of documents 2019-10-09 13:44:18 +02:00
Clément Renault
5f3072e67e
Support a basic update callback system 2019-10-09 11:45:19 +02:00
Clément Renault
2a4707d51e
Expose a function to be able to now the status of an update 2019-10-08 17:35:47 +02:00
Clément Renault
6534a9ec1d
Clean up many warning messages 2019-10-08 17:31:07 +02:00
Clément Renault
0a5ad4db06
Move the push update functions to their related modules 2019-10-08 17:24:11 +02:00
Clément Renault
6ee0d72c7b
Expose the synonyms operation updates on the Index 2019-10-08 17:18:22 +02:00
Clément Renault
ba32ce21d0
Introduce synonyms deletions updates 2019-10-08 17:16:48 +02:00
Clément Renault
0e224efa46
Introduce synonyms additions updates 2019-10-08 17:06:56 +02:00
Clément Renault
175461c13a
Port all tests to the TempDatabase struct 2019-10-08 16:16:30 +02:00
Clément Renault
c514692233
Introduce the TempDatabase in the QueryBuilder tests 2019-10-08 15:22:36 +02:00
Clément Renault
d8d0442d63
Fix many indexing and searching related bugs 2019-10-08 14:56:14 +02:00
Clément Renault
2236ebbd42
Introduce an example file to test indexing and searching csv documents 2019-10-08 14:48:48 +02:00
Clément Renault
0bfba3e4ba
Introduce a query_builder method on Index 2019-10-07 17:55:46 +02:00
Clément Renault
a57a64823e
Make possible to create an index and add a schema later on 2019-10-07 17:48:26 +02:00
Clément Renault
aa05459e4f
Introduce a background thread that manage updates to do 2019-10-07 16:16:04 +02:00
Clément Renault
0615c5c52d
Consume updates in the order of insertion 2019-10-07 15:00:28 +02:00
Clément Renault
487411340a
Prefix all the store names to avoid colliding with main stores 2019-10-07 10:56:55 +02:00
Clément Renault
5139dc7f3e
Let the caller commit/abort the operation 2019-10-07 10:52:45 +02:00
Clément Renault
88d0d3931c Store the schema in the main index 2019-10-04 17:49:13 +02:00
Clément Renault
df2ef8d2e1 Introduce update_task, popping an update and pushing the result of it 2019-10-04 17:49:13 +02:00
Clément Renault
29229b2137
Remove the update from the database when popped out 2019-10-04 17:16:34 +02:00
Clément Renault
851cc38216
Introduce the Database struct to manage indexes 2019-10-04 16:49:17 +02:00
Clément Renault
effbbc7370
Load the indexes at startup 2019-10-04 13:26:33 +02:00
Clément Renault
08e3f23408
Add the meilidb-schema/tokenizer projects 2019-10-04 10:29:44 +02:00
Clément Renault
62a0aefe44
Make the project be a workspace 2019-10-04 10:26:32 +02:00
qdequele
3fe346101b
chore: change logs in query_builder from info! to trace! 2019-09-24 13:35:46 +02:00
Clément Renault
69e0bae75e
chore: Bump dependencies 2019-09-18 14:42:23 +02:00
qdequele
59ae6458dc chore: bump dependencies 2019-09-17 18:50:44 +02:00
Clément Renault
c8ee21f227
feat: Replace the rayon::scope by always checking time 2019-09-01 18:52:38 +02:00
Clément Renault
9c5ec110e5
feat: Introduce a way to enable or disable query timeouts 2019-08-28 13:24:34 +02:00
Clément Renault
67302d09f3
feat: Multiword rewrite while there is time 2019-08-19 11:12:23 +02:00
Clément Renault
7dc9ea78fa
feat: Make the automaton DFA construction lazy 2019-08-19 11:12:23 +02:00
Clément Renault
0ee56314fb
feat: Try to simplify Store trait bound with a rayon scope 2019-08-19 11:10:54 +02:00
Clément Renault
b7b60b5fe5
feat: Introduce a new thread to avoid waiting on doc indexes fetchs 2019-08-16 16:35:19 +02:00
Clément Renault
d9c9fafd78
feat: Fetch doc indexes while there is time 2019-08-16 15:01:25 +02:00
Clément Renault
bb0a79c577
feat: Process automatons in the order they were sort 2019-08-16 12:25:35 +02:00
Clément Renault
81d44a0854
feat: Order automatons by importance 2019-08-16 12:19:34 +02:00
Clément Renault
bf3c2c3725 feat: Move the multi-word rewriting algorithm into its own function 2019-08-16 11:25:42 +02:00
Clément Renault
89df496f0c feat: Separate highlights from matches to make the code easier to follow 2019-08-16 11:25:42 +02:00
Clément Renault
9959f2e952 feat: Move the RawDocument type to its own module 2019-08-16 11:25:42 +02:00
Clément Renault
795557c046 feat: Remove query splitting from the automaton generation 2019-08-16 11:25:42 +02:00