Under Community Review

Export number of TUs from TMs and Sequences

(Raised on behalf of the customer) It would be  very useful to be able to do the following:

- Export the number of Translation units per language per TM

- Export the number of Translation units per language per TM sequence

The ideal format would be Excel.

Sandra Franco
  • in reply to Luis Lopes

    It would be good to have this data available on the SSRM, but not all customers have it - it would be good to have at least a basic ability to export such data outside of the reporting module.

  • in reply to Sandra Franco

    Thanks Sandra for getting this additional information. I wonder if it would not make more sense to feed this data into Datahub and then allow the customer to run such reports, rather than have some explicit Excel download just for this TM info. From Reporting they could then still download to Excel if they wanted to, but it seems more consistent if we approached it like that. What do you think?   FYI

  • Hi Luis, this is the information the customer provided on what data they would like and what it would be used for:

    To monitor performance and throughput: The time it takes to process content depends on the size of TM sequence. The report would provide this information.

    Optimisation of TM sequence: The TM sequence size report can be used to understand TM leverage (e.g. low leverage for specific lang pair) and used to optimize the sequence itself. And by re-running the report, you can validate the optimisation.

    TM growth: By comparing the report data to a baseline version, it is possible to see which TMs are growing or not.

    Compare language growth: For a specific TM, it can be interesting to compare number of TU and growth of TUs per language pair.

    Machine Translation: The TM and TM sequence size reports are useful to get insight information on what TMs could be used to train an MT model.

  • Thanks Sandra. Could you elaborate on the use case? What would happen with this exported data?