Faceted navigation available in Filerobot products (Widget, Hub and Portal)
At its core, faceted navigation utilizes a faceted classification system to organize information. This system categorizes items based on multiple, independent attributes called facets. Each added facet acts as new filtering layer.
Concept: Faceted navigation falls under the umbrella of parametric search techniques. Instead of relying solely on keyword matching in a traditional search, it allows users to filter based on pre-defined parameters (facets) associated with the asset. This enables a more nuanced, smarter search experience.
Technology: Faceted navigation relies on efficient data structures and algorithms to handle large datasets and filter combinations. In essence, the system needs to quickly identify items that match all selected facet criteria. Search engines like Solr or Elastic Search are popular choices for implementing faceted navigation due to their ability to handle complex filtering operations.
Benefits: The technical aspects of faceted navigation translate into several benefits:
Drill-down refinement: Users can progressively narrow down results by applying multiple facet filters, leading to a precise subset of assets.
Faceted exploration: Faceted browsing allows users to explore different categories and discover new assets they might not have found through a simple keyword search.
Improved efficiency: By filtering irrelevant options, faceted navigation reduces the number of results a user needs to sift through, saving them time and effort.
Faceted navigation is applied for 4 types of attributes (as of May 8th, 2024)
Date (Uploaded and Created date)
Metadata Single-select
Metadata Multi-select
Metadata Boolean
→ Faceted navigation is only available in the "Assets" view (ie. not in the Folder view)
→ Faceted navigation section is built as a side bar at the left side of the UI, replacing entirely the filter bar (shown below, that will then be hidden) in the Assets view.
This type of navigation is activated at a {token} level, and is a global setting for all users, that can be found at the Preference configuration page.
Within a metadata, the operator between 2 values or more is set to OR
Between 2 metadata or more, the operator is set to AND
In details:
Within a metadata
-> For multi-select, we provide 2 operators: AND
and OR
Let assume we are calling query HAS
(equal to IS
but for multi-select we call it HAS
just for gramatical purpose)
We will have 2 options
Metadata HAS
a AND
b -> expected results are files which have both values a and b, file that contains only a or b won't be returned
Metadata HAS
a OR
b -> expected results are files which has either a or b, file that contains both of the values will be returned also
➡️ This option cover the 1st option -> within a multi-select metadata, we will choose HAS-OR
operator to cover more file results
-> For single-select, it will always be OR
when users click in more than one answers as one file can not contain more than one value.
For 2 metadata or more
For faceted search, it is an AND operator between multiple metadata as the answers will be more specific and this is the main idea for faceted navigation: trimming down the most correct answers
a. Multi-select metadata "colour" with value: “red”, “white”, “black”
When "red" and "white" are searched for -> expected answers are:
Files that contain value red only
Files that contain value white only
Files that contain both value red and white
b. Single-select metadata "size" with value: “35”, “36”, “37”
When “35” and “36” are searched for -> expected answers are:
files that contain 35 value only
files that contain 36 value only
c. Now, combining the two previous examples
When “red”, “white”, “35”, “36” are searched for -> expected answers have to use AND
operator to encompass the 2 metadata constraints, so answers will be assets that contain:
red and 35,
red and 36,
white and 35,
white and 36,
red and white and 35,
red and white and 36
→ Refining searches further is leading to more accurate and relevant results.
AI powered search is a hub feature that allows the user to search assets by simply typing prompt in the search (like "photo of villa with swimming pool) It is an alternative way to find assets in the hub without using the metadata & standard search
AI powered search is available on specific tokens (PT_ADVANCED) and require the post process "Vector / Embed" to be enabled: only assets that have been vectorized using this post process are visible to the feature
Assets supported are images (PNG, JPG, WEBP)
In the asset page, simply switch the AI toggle on to activate the AI Search and enter the prompt:
After validation, you will be presented with a list of assets matching the prompt and sorted by relevance: most relevant assets are displayed first
Nota
The results generated largely depends on the assets available in the hub and the prompt used
Find Similar assets is a hub feature that allows the user to select an asset and generate with AI a list of existing assets sorted by similarity to the original asset
It is an alternative way to find assets in the hub without using the metadata & standard search
Find similar assets are available on specific tokens (PT_ADVANCED) and require the post process "Vector / Embed" to be enabled: only assets that have been vectorized using this post process are visible to the feature
Assets supported are images (PNG, JPG, WEBP)
In any of the hub page, simply select an asset (image: png, jpg, webp) et choose the find similar option in the contextual menu
After computing, you will be presented with a list of assets sorted by similarity with your original asset (indicated with a purple tag): most similar assets are displayed first
Nota
The results generated largely depends on the assets available in the hub (ie if the asset selected is unique, the similar assets suggested will be less relevant
Advanced search and filtering
The search bar allows user to use simple search with keywords
Advanced search is activated by typing the @
sign and selecting one of the dedicated search operator (filename,...). This allows user to search in specific fields - file, metadata, asset properties, etc.
User can also use the Context dropdown to narrow down the search scope to a specific folder or to Favourites.
Assets can be filtered by their creation or last modification date using pre-defined or custom time frames
With the format filter, user can choose from a list of asset types or file formats depending on the content of the Filerobot container
Filtering by filesize by entering a size range
This option allows user to filter by selecting one or more tags in the container
The metadata filter allows user to select multiple metadata fields as well as multiple values for each field and apply them as filters
The image group of filters lets user filter by image dimensions/resolution, contents, color or aspect ratio
A new filter called "Products" has been added to the search bar, allowing users to filter results based on the product_ref field. This feature lets users search for multiple product values at once and easily combine product references with any other metadata filter or grouping.Users must go to the Preferences section under the Project tab and toggle the Products Filter setting to enable this filter. Once enabled, the "Products" filter will be available for use in the search bar.