Breaking changes
-
Improve error handler (#1848) @ManyTheFish
errorCode
,errorType
anderrorLink
renamed tocode
,type
andlink
- New error
type
s:invalid_request
,auth
andinternal
- Failed updates now return an error object containing the following fields:
message
,code
,type
, andlink
. Failed responses no longer includemessage
,errorCode
,errorType
, anderrorLink
- New errors:
dump_not_found
,unretrievable_document
,dump_already_processing
,no_space_left_on_device
,invalid_store_file
,index_primary_key_already_exists
,database_size_limit_reached
,primary_key_inference_failed
,document_fields_limit_reached
,invalid_api_key
- Removed and renamed errors:
primary_key_already_present
,missing_primary_key
,max_fields_limit_exceeded
,dump_already_in_progress
,max_database_size_limit_reached
,invalid_token
- You can find an exhaustive list of new errors as well as updated error
code
s andmessage
s in the API spec
-
attributesToHighlight
now works on both numerical and string values (#1813) @Jhnbrn90
The_formatted
object in the search response does not contain any numerical value anymore but only strings (and array/object of strings). The numerical values of the original document are converted into strings and returned in the_formatted
object. The original document (containing the numerical values) is still available at the root of the search response. -
Using the docker image, in the container, MeiliSearch runs using the
meili
user and notroot
anymore. This involves the path where thedata.ms
is stored in the container is not/data.ms
anymore but/home/meili/data.ms
.
Improvements
- Increase limit of words per attribute from 1000 to 65535 (meilisearch/milli#368 on milli's side) @ManyTheFish
- Improve download-latest.sh file: add option to increase GitHub rate limit (#1760) @nav1s
- Reduce build size (#1781) @MarinPostma
Fixes
- Prevent error when importing an empty dump (#1787) @irevoire
- Optimize document addition (#1847) @MarinPostma
- Fix indexation performance on mounted disks (#1824) @ManyTheFish
- Fix CSV Content-Type error message (#1806) @sanders41
Misc
- Use Segment for data collection and telemetry (#1800) @irevoire
- Remove memmap dependency (#1793) @palfrey
- Add tests for mini-dashboard status and assets (#1852) @CuriousCorrelation
- Reduce ArmV8 binary build time (#1811) @patrickdung
- Remove
movies.json
sample dataset from the repository (#1693) @curquiza - Improve RUNs in Dockerfile (#1817) @nfsec
- Add Milestone Check status to bors.toml (#1808) @curquiza
- Add SECURITY.md file (#1853) @CaroFG
- Update CONTRIBUTING.md and README.md
❤️ Thanks again to our external contributors:
- MeiliSearch: @Jhnbrn90, @nav1s, @nfsec, @palfrey, @patrickdung, and @sanders41
- Milli: @CNLHC, @fumblehool, and @palfrey