Fix error checking on hex2int

[ajfriend]

Change from

cpdef H3int hex2int(H3str h) except 0

to

cpdef H3int hex2int(H3str h) except? 0

The previous version considered any return of 0 an error, which led to confusing error messages for the user. This function simply converts; check for a valid cell happens after conversion.

Changing except 0 to except? 0 makes it so that Python only checks if there is an error when 0 is returned. We need some C-compatible sentinel value to flag actual/potential errors for pure C/Cython functions.

Old error message:

New error message:

This PR also adds a test to verify that we get a helpful error message in this example.

[codecov]

Codecov Report

Merging #127 into master will not change coverage.
The diff coverage is n/a.

@@            Coverage Diff            @@
##            master      #127   +/-   ##
=========================================
  Coverage   100.00%   100.00%           
=========================================
  Files            9         9           
  Lines          181       181           
=========================================
  Hits           181       181           

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 944d917...8a6bdd8. Read the comment docs.

[isaacbrodsky]

I wonder how other Python libraries handle ensuring the user doesn't pass in incorrect types. I seem to recall Pandas giving me an error about types that was more specific, but I don't recall exactly what function call for that right now.

[ajfriend]

I'll land this for now, but more than happy to revisit if we have ideas around better error reporting.

Pandas might have it a little easier because it does have typed arrays for columns, in the form of numpy arrays, but I'm not sure if that's the error message you were thinking of.