51cf44d6fd
456: Remove useless grenad merging r=Kerollmops a=Kerollmops This PR must be merged after #454. This PR removes the part of code that was merging all of the grenad Readers merging that we don't need as the indexer should have merged them and, therefore, we should only have one final grenad Reader. We reduce the amount of CPU usage and memory pressure we were doing uselessly. `@ManyTheFish` are you sure I can skip merging the `word_docids` database? Here is the benchmark comparison with the previously merged PR #454: ``` group indexing_reintroduce-appending-sorted-values_c05e42a8 indexing_remove-useless-grenad-merging_d5b8b5a2 ----- ----------------------------------------------------- ----------------------------------------------- indexing/Indexing movies with default settings 1.06 16.6±1.04s ? ?/sec 1.00 15.7±0.93s ? ?/sec indexing/Indexing songs with default settings 1.16 60.1±7.07s ? ?/sec 1.00 51.7±5.98s ? ?/sec indexing/Indexing songs without faceted numbers 1.06 55.4±6.14s ? ?/sec 1.00 52.2±4.13s ? ?/sec ``` And the comparison with multi-batch indexing before #436, we can see that we gain time for benchmarks that index datasets in multiple batches but there is _so much_ variance that it's not clear. ``` group indexing_benchmark-multi-batch-indexing-before-speed-up_45f52620 indexing_remove-useless-grenad-merging_d5b8b5a2 ----- ---------------------------------------------------------------- ----------------------------------------------- indexing/Indexing geo_point 1.07 6.6±0.08s ? ?/sec 1.00 6.2±0.11s ? ?/sec indexing/Indexing songs in three batches with default settings 1.12 57.7±2.14s ? ?/sec 1.00 51.5±3.80s ? ?/sec indexing/Indexing songs with default settings 1.00 47.5±2.52s ? ?/sec 1.09 51.7±5.98s ? ?/sec indexing/Indexing songs without any facets 1.00 43.5±1.43s ? ?/sec 1.12 48.8±3.73s ? ?/sec indexing/Indexing songs without faceted numbers 1.00 47.1±2.23s ? ?/sec 1.11 52.2±4.13s ? ?/sec indexing/Indexing wiki 1.00 917.3±30.01s ? ?/sec 1.09 998.7±38.92s ? ?/sec indexing/Indexing wiki in three batches 1.09 1091.2±32.73s ? ?/sec 1.00 996.5±15.70s ? ?/sec ``` What do you think `@irevoire?` Should we change the benchmarks to make them do more runs? Co-authored-by: Kerollmops <clement@meilisearch.com>
a concurrent indexer combined with fast and relevant search algorithms
Introduction
This repository contains the core engine used in Meilisearch.
It contains a library that can manage one and only one index. Meilisearch manages the multi-index itself. Milli is unable to store updates in a store: it is the job of something else above and this is why it is only able to process one update at a time.
This repository contains crates to quickly debug the engine:
- There are benchmarks located in the
benchmarkscrate. - The
http-uicrate is a simple HTTP dashboard to tests the features like for real! - The
infoscrate is used to dump the internal data-structure and ensure correctness. - The
searchcrate is a simple command-line that helps run flamegraph on top of it. - The
helperscrate is only used to modify the database inplace, sometimes.
Compile and run the HTTP debug server
You can specify the number of threads to use to index documents and many other settings too.
cd http-ui
cargo run --release -- --db my-database.mdb -vvv --indexing-jobs 8
Index your documents
It can index a massive amount of documents in not much time, I already achieved to index:
- 115m songs (song and artist name) in ~48min and take 81GiB on disk.
- 12m cities (name, timezone and country ID) in ~4min and take 6GiB on disk.
These metrics are done on a MacBook Pro with the M1 processor.
You can feed the engine with your CSV (comma-separated, yes) data like this:
printf "id,name,age\n1,hello,32\n2,kiki,24\n" | http POST 127.0.0.1:9700/documents content-type:text/csv
Don't forget to specify the id of the documents. Also, note that it supports JSON and JSON
streaming: you can send them to the engine by using the content-type:application/json and
content-type:application/x-ndjson headers respectively.
Querying the engine via the website
You can query the engine by going to the HTML page itself.
Contributing
You can setup a git-hook to stop you from making a commit too fast. It'll stop you if:
- Any of the workspaces does not build
- Your code is not well-formatted
These two things are also checked in the CI, so ignoring the hook won't help you merge your code.
But if you need to, you can still add --no-verify when creating your commit to ignore the hook.
To enable the hook, run the following command from the root of the project:
cp script/pre-commit .git/hooks/pre-commit