When a job is launched, the details of its execution are available. In this log you can see traces of the different processes, including the requests sent to the MeaningCloud server:
MeaningCloud will log the beginning of the process with the main parameters, and then a line for each one of the texts processed. If any of these requests are not successful, details on the error returned will be output there, as well as shown in the results:
Every error that may be returned by the APIs is explained in the error codes catalog.
When an error that's considered blocking (that is, all API requests are going to fail) is returned, an exception is raised:
In this case, we are trying to classify texts with a model included in a vertical pack we don't have access to.