Problem

LLM usage is cost intensive. Therefore a statistic is needed, how many tokens are used per task/user/client/language resource/time period

Solution

Extend the current language resource usage statistic. to count tokens for AI usage.

Make the statistic available in the same xlsx export. Can be on separate sheets, if that make sense.

Record every request, as we do it in the "Resource usage log" sheet of the current xlsx.

In addition to the currently logged columns also log the request type, where the tokens stem from (pre-translation, quality estimation on pre-translation, user saving of segment, InstantTranslate file pre-translation, InstantTranslate file quality estimation, InstantTranslate text field pre-translation, InstantTranslate text field quality estimation). 

Also log the task ID and task name in case of tasks.

Offer to export the task-related values as a task statistic in the analysis tab of a task next to the "Export as Excel" button.

Provide a