Loïc Lecrenier

Batch document additions and deletions together

4

When someone sends many document addition tasks interspersed with document deletion tasks, meilisearch is currently forced to process these tasks serially. For example, if the following tasks are sent in...

I'd like to understand how the use of [DigitalOcean's Volume Block Storage](https://www.digitalocean.com/products/block-storage) and similar cloud offerings impact the performance of a Meilisearch instance, for all kinds of indexing tasks as...

Implement more term matching strategies: frequency, first, etc.

3

Linked to the following product discussion: https://github.com/meilisearch/product/discussions/255. It's possible to implement two additional matching strategies ([v0.30.0 documentation link](https://github.com/meilisearch/documentation/blob/v0.30/reference/api/search.md#matching-strategy)): - `frequency` will drop the query words that are the most frequent...

This issue is linked to the following product discussion: https://github.com/meilisearch/product/discussions/593 . More will be added to it depending on the course of the discussion.

Remove the "iterative" versions of the search algorithms

2

The ranking rules `proximity`, `sort`, and `attribute` have two different implementation strategies. The first one (`set-based`) queries milli's databases and performs set operations on roaring bitmaps to find buckets of...

Meilisearch allocates too much memory when processing a settings update

3

**Describe the bug** When processing a `settingsUpdate` task, meilisearch allocates twice the amount of memory given by `--max-indexing-memory`. **To Reproduce** 0. Launch meilisearch with the flag `--max-indexing-memory 2GB` 1. Index...

We use a lot of `grenad::Reader/Writer/Sorter` throughout `milli`, and it can be a challenge to ensure that the data we put into them is encoded and decoded consistently. Sometimes we...

Write snapshot tests

1

When refactoring code in milli, it can be hard to make sure that the changes we make are correct. I experienced this when I worked on https://github.com/meilisearch/milli/pull/587 1. I refactored...

As explained in the [Meilisearch documentation](https://docs.meilisearch.com/learn/core_concepts/relevancy.html#behavior) and in [this issue](https://github.com/meilisearch/meilisearch/issues/3123), the role of a ranking rule is to sort search results by their relevancy. Each ranking rule narrows down a...

Meilisearch works by performing a [bucket sort](https://en.wikipedia.org/wiki/Bucket_sort) of the documents that match a search query ([as explained shortly in the Meilisearch documentation](https://docs.meilisearch.com/learn/core_concepts/relevancy.html#behavior)). For example, with the following ranking rules: ```...