Added additional information to exceptions

[roll]

• fixes #126

[coveralls]

Pull Request Test Coverage Report for Build 462

• 13 of 13 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.06%) to 85.423%

---

Totals
Change from base Build 458:	0.06%
Covered Lines:	1758
Relevant Lines:	2058

---

💛 - Coveralls

[coveralls]

Pull Request Test Coverage Report for Build 453

• 13 of 13 (100.0%) changed or added relevant lines in 2 files are covered.
• No unchanged relevant lines lost coverage.
• Overall coverage increased (+0.07%) to 85.119%

---

Totals
Change from base Build 451:	0.07%
Covered Lines:	1716
Relevant Lines:	2016

---

💛 - Coveralls

[roll]

Hi @akariv @cschloer,

I played a little bit with additional information on exceptions. It partially works but I don't fully understand how flows are composed so this implementation is really limited.

Can be a starting point for a better implementation

[cschloer]

This looks good, and pretty much exactly what I was getting at, but I see that it's only limited to processors that inherit the DataStreamProcessor class. Is there any way to extrapolate this out to the flow.py file? So it works by default for all processors, even custom ones that don't inherit any datahq code? I also am not familiar enough with how flows work to get a sense of whether that is possible.

[roll]

hi @akariv WDYT?

[roll]

As decided on the prev meeting - it's parked for now to get @akariv input first

[akariv]

@roll - sorry for missing this, it's really great!

Regarding the implementation -

• The added information to the exceptions is neat, but most users won't notice them. I wonder if we could create a 'DataflowsException' which will have these attributes and will be chained to the original exception (using the raise ... from e syntax)
• Flow composition can be done using chaining processors or Flows. The stored position is the processor in the flow, but it might get confusing when combining a number of flows together. Perhaps we could have the position to be a combination of Flow + position in flow, but for now I think the implementation is fine.
• row_processor, rows_processor, datapackage_processor, iterable_loader - all return an instance of DataStreamProcessor so all should get the position parameter as well

[roll]

Hi @akariv,

I've added the exception objects you mentioned. Also, updated the implementation to make it work with different combinations of processors. TBH not sure how it works but it seems to be working