Tuning

Tune important Document attributes.

In the "Tuning" section, you can add a linear boost to your fields. Boosting a field plays an important role when calculating a Document's relevancy.

Imagine we have those 3 Documents.

[
  {
    "name": "Lilo & Stitch",
    "category": ["Family", "Comedy"],
    "release_year": "2002",
    "duration_minutes": 85
  },
  {
    "name": "Peter Pan",
    "category": ["Fantasy", "Adventure"],
    "release_year": "1953",
    "duration_minutes": 76
  },
  {
    "name": "Treasure Planet",
    "category": ["Adventure", "Sci-fi"],
    "release_year": "2002",
    "duration_minutes": 95
  }
]

We now set the realease_year boost to 3.0 and the category boost to 1.0.

Now your user searches for

adventure 2002

Of course, the first result will be the "Treasure Planet" movie that contains both tokens.

But the second will be the movie "Lilo & Stitch" that was released in 2002 because compared to "Peter Pan" it's release_date attribute is more important the the category attribute in the "Peter Pan" movie.

Here is an overview of the sorted returned results, and their matches

1. Treasure Planet

   • Category -> Adventure
   • Release Year -> 2002

2. Lilo & Stitch

   • Release Year -> 2002

3. Peter Pan

   • Category -> Adventure

Attributes

Tune your index attributes on a Search-Level in your Search settings.

The Tuning section you can add a linear boost to your fields. Boosting a field plays an important role when calculating the relevancy for a Document.

Image we have those 3 documents

[
  {
    "name": "Lilo & Stitch",
    "category": ["Family", "Comedy"],
    "release_year": "2002",
    "duration_minutes": 85
  },
  {
    "name": "Peter Pan",
    "category": ["Fantasy", "Adventure"],
    "release_year": "1953",
    "duration_minutes": 76
  },
  {
    "name": "Treasure Planet",
    "category": ["Adventure", "Sci-fi"],
    "release_year": "2002",
    "duration_minutes": 95
  }
]

we now set the realease_year boost to 3.0 and the category boost to 1.0.

Now your user searches for

adventure 2002

Of course the first result will be the "Treasure Planet" movie that contains both tokens.

But the second will be the movie "Lilo & Stitch" that was released in 2002 because compared to "Peter Pan" it's release_date attribute is more important the the category attribute in the "Peter Pan" movie.

Here is an overview with the sorted returned results, and their matches

1. Treasure Planet

   • Category -> Adventure
   • Release Year -> 2002

2. Lilo & Stitch

   • Release Year -> 2002

3. Peter Pan

   • Category -> Adventure

Attributes

Tune your index attributes on a Search-Level in your Search settings.