Sync Fork

.env.example

@@ -0,0 +1,12 @@
+# Test
+export API_MOCK_SERVER_PORT=8002
+export SLACK_MOCK_API_PORT=8002
+export POSTGRES_HOST=localhost
+export POSTGRES_PORT=5432
+export DB_NAME=algoliax_test
+export DB_USERNAME=postgres
+export DB_PASSWORD=postgres
+
+# Run
+export ALGOLIA_API_KEY=
+export ALGOLIA_APPLICATION_ID=

.github/CODEOWNERS

@@ -0,0 +1 @@
+* @WTTJ/mission-platform-backend

.github/dependabot.yml

@@ -0,0 +1,10 @@
+version: 2
+updates:
+  - package-ecosystem: mix
+    directory: "/"
+    schedule:
+      interval: daily
+      time: "06:00"
+      timezone: "Europe/Paris"
+    open-pull-requests-limit: 100
+    rebase-strategy: "disabled"

.github/workflows/publish_public_package.yml

@@ -0,0 +1,44 @@
+name: Publish package to Hex.pm
+
+on:
+  release:
+    types: [released]
+
+jobs:
+  publish:
+    runs-on: ubuntu-latest
+    if: "!github.event.release.prerelease"
+
+    steps:
+    - uses: actions/checkout@v4
+
+    - name: Set up Elixir
+      uses: erlef/setup-beam@v1
+      with:
+        elixir-version: '1.17.2'
+        otp-version: '26.2.5.2'
+
+    - name: Install dependencies
+      run: |
+        mix deps.get
+
+    - name: Compile
+      run: |
+        mix compile
+
+    - name: Check version consistency
+      run: |
+          MIX_VERSION=$(mix run -e 'IO.puts Mix.Project.config[:version]')
+          RELEASE_VERSION=${{ github.event.release.tag_name }}
+          RELEASE_VERSION=${RELEASE_VERSION#v}
+
+          if [ "$MIX_VERSION" != "$RELEASE_VERSION" ]; then
+            echo "Version mismatch: mix.exs version ($MIX_VERSION) != release version ($RELEASE_VERSION)"
+            exit 1
+          fi
+
+    - name: Publish to Hex.pm
+      env:
+        HEX_API_KEY: ${{ secrets.HEXPM_PUBLIC_PACKAGES_KEY }}
+      run: |
+        mix hex.publish --yes

.gitignore

@@ -1,3 +1,10 @@
+# IDE
+.idea/
+.vscode/
+
+# Env
+.env
+
 # The directory Mix will write compiled artifacts to.
 /_build/
 

CHANGELOG.md

@@ -1,6 +1,167 @@
 # Changelog
 
-## DEV
+## v0.10.0 - 2025-03-10
+
+### New
+
+Added the `synonyms` settings to the `Algoliax.Indexer`.
+It allows you to automatically setup synonyms for your indexes when calling `MyIndexer.configure_index()`
+
+If not specified or set to `nil`, the synonyms will not be configured.
+Otherwise, the following keywords are expected, with each key having a default value:
+
+- `synonyms`: a list of synonym groups, as expected by Algolia. Default `[]`
+- `replace_existing_synonyms`: Whether to replace existing synonyms. Default `true`
+- `forward_to_replicas`: Whether to forward synonyms to replicas. Default `true`
+
+You can also provide an arity-1 function (that takes the `index_name`) that returns the same keyword list.
+
+If using `forward_to_replicas: true`, make sure not to specify synonyms on the replicas themselves to avoid conflicts/overwrites.
+
+```elixir
+  defmodule Global do
+    use Algoliax.Indexer,
+      index_name: :people,
+      object_id: :reference,
+      schemas: People,
+      algolia: [
+        attribute_for_faceting: ["age"],
+        custom_ranking: ["desc(updated_at)"]
+      ]
+      synonyms: [
+        synonyms: [
+          %{
+            objectID: "synonym1",
+            type: "synonym",
+            synonyms: ["pants", "trousers", "slacks"],
+            ...
+          },
+          %{
+            objectID: "synonym2",
+            ...
+          }
+        ],
+        forward_to_replicas: false,
+        replace_existing_synonyms: false
+      ]
+  end
+```
+
+### Other changes
+
+- Updated the list of supported algolia settings.
+
+## v0.9.1 - 2024-12-12
+
+### New
+
+The `Algoliax.Indexer` now supports dynamic definition for the `:algolia` settings. It can supports 2 types of configurations:
+
+- (Existing) Keyword list
+- (New) Name of a 0-arity function that returns a keyword list
+
+```elixir
+defmodule People do
+  use Algoliax.Indexer,
+    index_name: :people,
+    object_id: :reference,
+    schemas: [People],
+    algolia: :runtime_algolia
+
+  def runtime_algolia do
+    [
+      attribute_for_faceting: ["age"],
+      custom_ranking: ["desc(updated_at)"]
+    ]
+  end
+end
+```
+
+Also added 2 new exceptions for the `:algolia` configuration: `InvalidAlgoliaSettingsFunctionError` and `InvalidAlgoliaSettingsConfigurationError`
+
+### Other changes
+
+Existing direct calls to `Algoliax.Settings.replica_settings/2` will still work but will not benefit from
+the dynamic `:algolia` configuration. Please use `Algoliax.Settings.replica_settings/3` instead.
+
+## v0.9.0 - 2024-11-26
+
+`ALGOLIA_API_KEY` and `ALGOLIA_APPLICATION_ID` aren't read anymore from system env variables
+inside the code. Only application config is now used (as documented).
+
+If you used these env vars, you should now read them inside the config:
+
+```elixir
+config :algoliax,
+  api_key: System.get_env("ALGOLIA_API_KEY"),
+  application_id: System.get_env("ALGOLIA_APPLICATION_ID)
+```
+
+## v0.8.3 - 2024-10-18
+
+New `if` option for replicas which decides if they should be updated or not.
+
+- **If not provided**, the replica will be updated (so no impact on existing configurations)
+- Must be `nil|true|false` or the name (atom) of a arity-0 func which returns a boolean
+- If provided, the replica will be updated only if the value is `true` or the function returns `true`
+
+Useful for Algolia's A/B testing which requires replicas and only deploy them in production.
+
+## v0.8.2 - 2024-09-16
+
+- Upgrading all dependencies
+- Added `dependabot` to the repository
+
+## v0.8.1 - 2024-09-11
+
+#### New
+
+- Added new **optional** settings `default_filters` to be applied automatically when calling `reindex` without query
+  or `reindex_atomic`. Defaults to `%{}` which was the previous behavior.
+
+```elixir
+defmodule BlondeBeerIndexer do
+  use Algoliax.Indexer,
+    index_name: :blonde_beers,
+    object_id: :name,
+    schemas: [Beer],
+    default_filters: %{where: [kind: "blonde"]} # <--- can be a map
+end
+
+defmodule BeerIndexer do
+  use Algoliax.Indexer,
+    index_name: :various_beers,
+    object_id: :name,
+    schemas: [Beer1, Beer2, Beer3],
+    default_filters: :get_filters # <--- can be a function
+
+  def get_filters do
+    %{
+      Beer1 => %{where: [kind: "blonde"]},  # <--- custom filter for Beer1
+      :where => [kind: "brune"] # <--- Will be used for Beer2 and Beer3
+    }
+  end
+end
+```
+
+#### Contributing
+
+- New `CONTRIBUTING.md` file
+- Simplified the `config/test.exs` file
+- Provide a `.env.example` file to help contributors to setup their environment
+
+## v0.8.0 - 2024-09-11
+
+#### Breaking changes
+
+- Errors for `get_object` and `get_objects` are now of arity 4 and returns the original request as last attribute
+
+#### New
+
+- Added possibility to have multiple indexes for a model (applies for replicas as well)
+- Added `build_object/2` with the index name as second parameter (useful for translations for example)
+
+## v0.7.2 - 2023-09-18
 
 #### Breaking changes
 
@@ -10,7 +171,9 @@
 
 #### New
 
-- Add wait_task/1 (https://hexdocs.pm/algoliax/Algoliax.html#wait_task/1)
+- Add wait_task/1 (<https://hexdocs.pm/algoliax/Algoliax.html#wait_task/1>)
+
+## v0.7.0 - 2022-03-29
 
 #### Breaking changes
 

CONTRIBUTING.md

@@ -0,0 +1,33 @@
+# Contributing
+
+## Install
+
+If you are not using `asdf`, you will need to manually install the program versions listed in `.tool-versions`.
+Otherwise, simply run `asdf install` and then:
+
+```shell
+mix deps.get
+mix deps.compile
+mix compile
+```
+
+## Run tests
+
+Before running `mix test`, ensure you have setup your environment variables.
+Look at the `.env.example` file for the required variables.
+Then run `mix test` to run the tests.
+
+## Quality
+
+- Run `mix format` to format the code.
+- Run `mix credo` to run the linter.
+
+## CI/CD
+
+We use CircleCI to:
+
+- Run the code_analysis (format/credo)
+- Check for vulnerabilities
+- Run the tests
+
+See the `.circleci/config.yml` file for more details.

README.md

@@ -16,7 +16,7 @@ The package can be installed by adding `:algoliax` to your list of dependencies
 ```elixir
 def deps do
   [
-    {:algoliax, "~> 0.7.0"}
+    {:algoliax, "~> 0.10.0"}
   ]
 end
 ```
@@ -89,6 +89,22 @@ defmodule People do
 end
 ```
 
+- `build_object/2` does the same but provides the current index name as a second parameter. Can be useful when indexing the same model on multiple indexes (ie. for translations).
+
+```elixir
+defmodule Article do
+  ...
+
+  @impl Algoliax
+  def build_object(author, "article_index_" <> locale) do
+    %{
+      author: article.author,
+      content: article.content[locale]
+    }
+  end
+end
+```
+
 #### Index name at runtime
 
 It's possible to define an index name at runtime, useful if `index_name` depends on environment or comes from an environment variable.
@@ -108,6 +124,21 @@ defmodule People do
 end
 ```
 
+#### Multiple indexes
+
+It's possible to define multiple indexes for a same model.
+
+To achieve this, just specify an array of index names, or simply return an array in your `index_name/0` runtime function
+
+```elixir
+defmodule Article do
+  use Algoliax.Indexer,
+    index_name: [:algoliax_article_fr, :algoliax_article_en],
+    object_id: :reference,
+    algolia: [...]
+end
+```
+
 #### Index functions
 
 ```elixir
@@ -238,8 +269,42 @@ use Algoliax.Indexer,
   ]
 ```
 
+If the main index holds multiple indexes (e.g for an index per language usecase), replicas need to hold the same amount of names.
+The order is important to be associated to the correct main index.
+
+```elixir
+use Algoliax.Indexer,
+  index_name: [:algoliax_article_en, :algoliax_article_fr],
+  object_id: :reference,
+  repo: MyApp.Repo,
+  algolia: [
+    attributes_for_faceting: ["published_at"],
+    searchable_attributes: ["content"],
+  ],
+  replicas: [
+    [index_name: [:algoliax_article_by_publication_asc_en, :algoliax_article_by_publication_asc_fr], inherit: true, algolia: [ranking: ["asc(published_at)"]]],
+    [index_name: [:algoliax_article_by_publication_desc_en, :algoliax_article_by_publication_desc_fr], inherit: false, algolia: [ranking: ["desc(published_at)"]]]
+  ]
+```
+
+## Configure index name at runtime
+
+To support code for multiple environments, you can also define the index name at runtime. To achieve this, create a function within your indexer module and reference it using its atom in the Indexer configuration.
+
+```elixir
+defmodule People do
+  use Algoliax.Indexer,
+    index_name: :runtime_index_name,
+    #....
+
+  def runtime_index_name do
+    System.get_env("INDEX_NAME")
+  end
+end
+```
+
 ## Copyright and License
 
-Copyright (c) 2020 CORUSCANT (welcome to the jungle) - https://www.welcometothejungle.com
+Copyright (c) 2020 CORUSCANT (welcome to the jungle) - <https://www.welcometothejungle.com>
 
 This library is licensed under the [BSD-2-Clause](./LICENSE.md).