Implement commits missing in refactored code

[17o2]

After merging #447 (based on code from #251), the following features were not yet implemented in the refactored code.
This issue preserves the list, and a new PR should bring these issues into the new codebase.

• 557122c Look-up mated connectors before mate processing (Fix bugs in mate processing #358)
• 82751e4 Move mates processing above tweak processing (Fix bugs in mate processing #358)
• 668ba72 Avoid decimal point and trailing zero for integer BOM quantities (Avoid decimal point and trailing zero for integer BOM quantities #374)
• 6488eb5 Avoid Graphviz error when hiding all pins (Avoid Graphviz error when hiding all pins #375)

[kvid]

• 557122c Look-up mated connectors before mate processing (Fix bugs in mate processing #358)

The related code has been implemented differently, and this fix doesn't fit as is. It seems not needed with the refactored code. Test use cases from #355 and #359 seem to pass without any exceptions when processed by the refactored code.

---

• 82751e4 Move mates processing above tweak processing (Fix bugs in mate processing #358)

The related code has been implemented differently, and this fix doesn't fit as is. It seems not needed with the refactored code. TODO: Keep the comments added below tweak code in #358 to avoid future similar bugs.

---

• 668ba72 Avoid decimal point and trailing zero for integer BOM quantities (Avoid decimal point and trailing zero for integer BOM quantities #374)

There seems to be another bug (see #490) in the refactored code that generates BOM with wrong Qty values in some cases, and that makes it hard to check the original issue with the decimal point.

TODO: Investigate more after this other bug has been fixed.

---

• 6488eb5 Avoid Graphviz error when hiding all pins (Avoid Graphviz error when hiding all pins #375)

The problem is still triggered with the refactored code when executed with the suggested input described in #375, but this fix doesn't fit as is with the refactored code.

There is currently a TODO about handling the issue at https://github.com/wireviz/WireViz/blob/dev/src/wireviz/wv_graphviz.py#L268

However, it is not described how to handle it. One way could be to raise an exception describing the problem (and how to avoid it). Another way could be to use tbl = Table(pin_rows, delete_if_empty=True, border=0, ...) to silently (or preferably with a warning output) delete the table when empty. The latter way is similar to the fix from PR #375.

As a table with zero rows could trigger the same problem also in other contexts, we probably also should handle this generic issue inside the Table class and inform the user somehow unless already handled at a higher level as described above.

I suggest we first create a generic fix in the Table class and test it with the #375 use case. Then, we can add extra information to the user about the root cause in this context at the TODO in the gv_pin_table() function if needed.