More expressive primary service init failure #476
Conversation
t-ober
changed the title
# Conflicts: # CHANGELOG.md # src/main/scala/edu/ie3/simona/service/primary/PrimaryServiceWorker.scala
| @@ -28,6 +28,7 @@ simona.input.primary.csvParams = { | |||
| directoryPath: "input/samples/vn_simona/fullGrid" | |||
| csvSep: "," | |||
| isHierarchic: false | |||
| timePattern: "yyyy-MM-dd'T'HH:mm:ss[.S[S][S]]'Z'" | |||
There was a problem hiding this comment.
That's the default time pattern, so this line effectively does not change any configuraiton. Does it make sense to include it?
| case Failure(exc: DateTimeParseException) => | ||
| throw new IOException( | ||
| s"Could not initiate the time series source. Parsing ${exc.getParsedString} failed with timepattern: $timePattern. The time pattern can be configured in the configuration.", | ||
| exc | ||
| ) |
There was a problem hiding this comment.
If we want this kind of distinction, would'nt it rather make sense to have it in PSDM?
So, everything that can go wrong within *TimeSeriesSource is fully described by an exception that its constructor throws. Then, when we arrive at the level of PrimaryServiceWorker, we can then wrap the exception once more and say: "During creation of the timeseries source, an error occured." or something like that.
There was a problem hiding this comment.
Possibly solved here (or with an earlier PR). We should still check if DateTimeParseException is also wrapped in a SourceException within PSDM
| valueClass, | ||
| factory | ||
| ) | ||
| ) match { |
There was a problem hiding this comment.
This error matching would also be nice for SqlTimeSeriesSource as well, no? So maybe we can move this an abstraction level upwards, i.e. checking for Failure after *TimeSeriesSource has been created.
Resolves #475