First, select a template from the list, which correspond to the types of resources that can be described in ATLAS. Press Save.

You can contribute to ATLAS with any information you know, no need to fill in all fields, just pay attention to those fields that are flagged with a red *, which are mandatory. Once you submit a record, if you do not have an account, you'll not be allowed to modify it. But don't worry, our editorial board will review your suggestions before publishing the record.

Please, use Chrome, Firefox or Safari!

Some fields have autocomplete functionalities! You recognise them because they have an icon next to the field name. A dropdown pops up while typing, including best candidate matches with the term you typed (powered by Wikidata, VIAF and ATLAS).

If you found a match in the list, you can click on the icon to be redirected to the page of the term. To select a term from the list, click on the label, and it will appear below the text area.

If you don't find any term that matches your search, just click on return to add a new one to ATLAS. It will appear below the text area. Other contributors will be able to use it in new records.

You can always remove values by clicking on the X.

The same functionalities are also extended to the fields associated with SKOS vocabularies. A dropdown is displayed, including suggestions of terms gathered from the available SKOS vocabularies. A handy link to each vocabulary's web page allows you to preview the option.

Suggestions are retrieved on-the-fly. The search starts on keyup (after few milliseconds) and can take few seconds if the connection is unstable (so be patient!). Secondly, results may reload slowly, meaning that refinements of the search might take time. If no relevant results are found at the first attempt, please try to remove the last character and re-type it. Lastly, when looking for two or more words (e.g. people names, place names, types of objects) be aware that you may have to try several combinations of words (e.g. firstname+lastname rather than lastname+firstname).

Some fields are not expected to be filled as usual. Instead, they provide an advanced functionality for extracting Key Entities from various sources. A dropdown allows you to select the option you prefer to run your query (a SPARQL endpoint, an API service, a static file, a website).

To properly use this Knowledge Extraction tool, you must always retrieve uri-label pairs. In case a URI is not available, you should use the uri variable to retrieve the best suited label for entity reconciliation. The system will automatically retrieve the best URI from the configured service. Conversely, if a label is missing, it will also be automatically retrieved from the configured service.

API

Required parameters to send a query to an API service include:

• API access point (mandatory): the URL of the required service.
• Query parameters (optional): a set of key,value pairs providing query parameters. You can optional omit this information in case query parameters are encoded within the query URL.
• Result keys (mandatory): to properly handle retrieved results, you must provide the keys pointing to the desired values (a uri and its label) within the query output. Results are usually returned in the form of an array of objects. Consequently, three paths are needed to reach the array of results and to access URIs and labels. In case multiple keys should be accessed to retrieve a value, they must be separated by a dot. Example:

  array,results.results
  uri,uri
  label,prefLabel
  

• Filter results (optional): allows you to refine the retrieved data by applying one or more filters to specific variables returned by the query. Each filter must specify: the filter type (Regex, Greater than, Less than), the target variable (the field to which the filter applies), and the filter value (the threshold or regex pattern). For example: ("Greater than", "Mentions counter", "10") retrieves only items where the "Mentions counter" value exceeds 10.

SPARQL

Required parameters to send a query to an API service include:

• SPARQL endpoint (mandatory): the URL of the required service.
• Query (mandatory): a SPARQL query to retrieve ?uri and ?label. Remember to include necessary prefixes at the beginning of the query. Example: PREFIX skos: <http://www.w3.org/2004/02/skos/core#>

Static files

This option allows knowledge extraction from .json, .csv, and .xml files and provides two query methodes:

• Manual query: provide the URL of a compatible file. Its content will be parsed to extract json keys, csv column labels, or xml tags. You can use these outputs to retrieve meaningful entities. However, variables disambiguation is not available, so you should get labels and let the system automatically retrieve URIs.
• SPARQL query: provide the URL of a compatible file and a SPARQL query to retrieve ?uri and ?label. Please follow SPARQL Anything recomendations to properly interrogate different file formats.

Websites

Required parameters to extract information from websites include:

• Website URL (mandatory): the URL of the target website to be processed.
• HTML Selector (mandatory): a valid selector used to isolate meaningful HTML elements within the loaded DOM structure.
• HTML Attribute (optional): by default, the inner text of the selected HTML element is retrieved and used as the label for Entity Reconciliation to obtain a URI. Alternatively, you can specify the name of an attribute (e.g., href, src, title) to extract its corresponding values.
• Regex (optional): a regular expression used to transform or clean the extracted text. You must provide two patterns — one for search and one for replacement — separated by a comma (e.g., "pattern_to_find","replacement_pattern"). The modified label will then be available for entity reconciliation.