Look more closely into `number_format` in both writing and reading situations

[bartkl]

A few observations and remarks:

• The number_format parameter is implemented in the Write to cell keyword, but data types are similarly important in situations where Read from cell is used. What consequences does this have possibly?
  • If you read number or date-time values, there's a good chance types will be inferred by OpenPyXL, leading to integers, floats, DateTime objects, andsoforth. If this is the case, the trim functionality will cause problems in those cases. Look into it!
• Should number_format be implemented as a parameter for the Read from cell keyword as well?
• Should we simplify things and convert everything to strings? I think not, but it's worth discussing. We could also consider adding a parameter as_string=True to both write and read keywords, which could optionally simplify things (especially in regards to the trimming caveats mentioned before).

If someone has an opinion about this: let's get this discussion going. In the meantime I won't be spending too much time documenting or even understanding the exact consequences of number_format.