Bidirectional linking

Linking tables

It is now possible (from 22nd August 2023) to enable bidirectional links between HEPData tables possibly in different records. Applications might include linking theoretical predictions or covariance matrices to their related measurements. The YAML document specifying the table metadata in the input submission.yaml file should contain a new field related_to_table_dois comprising a list of table DOIs:

related_to_table_dois:
- 10.17182/hepdata.12345.v1/t2
- 10.17182/hepdata.67890.v3/t4

Please note: This field should not be used for self-referencing, the DOIs inserted should be for OTHER related tables.

You can find the desired table DOI displayed next to the table name on the relevant HEPData record page. It is built from a common prefix (10.17182/hepdata), the HEPData record identifier (for example, 12345 or 67890), the HEPData record version (v1, v2, etc.), and the table number within the HEPData record (t1, t2, etc.).

After upload, the related_to_table_dois field is persisted to the database and rendered above the table description with links to the related tables. The name of the linked tables will be displayed instead of the DOI, with a tooltip containing their description displayed when the user hovers over the table name.

The linked tables will also display links back to the referring tables, but only after the records containing the referring tables have been finalised. The backward link is only displayed if the referring record is the most recent finalised version. The related_to_table_dois field is ignored for uploads to the HEPData Sandbox.

An example record is https://www.hepdata.net/record/ins2729396?version=1. However, note that the links between tables are automatically bidirectional so that, for example, if specifying the related_to_table_dois field in Table 1 (linked to Table 11), then the related_to_table_dois field does not also need to be specified in Table 11 (linked to Table 1). Doing so will result in duplicate links. Note that a HEPData software bug currently means that the links will only be rendered after a record has been finalised.

The hepdata_lib tool can be used to write the related_to_table_dois field (see Adding links to related tables).

Linking records

In a similar way, bidirectional links can be made between HEPData records, rather than individual tables, via a new field related_to_hepdata_records added to the first YAML document of the submission.yaml file, for example, alongside the comment and additional_resources. It should be given as a list of HEPData record identifiers:

related_to_hepdata_records:
- 12345
- 67890

Please note: This field should not be used for self-referencing, the IDs inserted should be for OTHER related records.

Note that the HEPData record identifier must be given as an integer (not a string) and it differs from the INSPIRE record identifier often found after ins in a HEPData record URL. You can find the HEPData record identifier from the integer after the prefix 10.17182/hepdata in the HEPData record DOI displayed alongside the publication information in the left-hand panel of a HEPData record.

Again, after upload, the related_to_hepdata_records field is persisted to the database and rendered in the left-hand panel of a HEPData record with links to the related records. A tooltip containing the title of the linked record is displayed when the user hovers over the record identifier. The linked records will also display links back to the referring records, but only after the referring records have been finalised and if the referring record is the most recent version. The related_to_hepdata_records field is ignored for uploads to the HEPData Sandbox.

The hepdata_lib tool can be used to write the related_to_hepdata_records field (see Adding links to related records).