[Python][HW] hw.WireOp inner_sym is inconvenient to use

Nobody has claimed this yet.

The move to using `InnerSymAttr` in https://github.com/llvm/circt/pull/5703 broke a somewhat obscure usage of the Python API for creating `hw.WireOp`s directly.

The old code used to look like this:

```python
op = hw.ConstantOp(ir.BoolAttr.get(True))
wire_name = "Example"
wire = hw.WireOp(op, name=wire_name, inner_sym=wire_name)
```
Now we have:
```python
op = hw.ConstantOp(ir.BoolAttr.get(True))
wire_name = "Example"
wire = hw.WireOp(op, name=wire_name,
inner_sym=hw.InnerSymAttr.get(ir.StringAttr.get(wire_name)))
```
Which doesn't seem better.

`sv.WireOp()` got a `sym_name` parameter that does the right thing:
```python
if sym_name is not None:
attributes["inner_sym"] = hw.InnerSymAttr.get(StringAttr.get(sym_name))
```
The same fix should probably be applied to `WireOp`, and maybe just generally to all `hw` operations

Start by comparing the Python API entry points for hw.WireOp and sv.WireOp, focusing on their inner_sym and sym_name parameters and the InnerSymAttr construction shown in the issue. Check whether other hw operations expose the same inconvenient pattern; done means WireOp accepts the simpler string form and any broader scope is addressed consistently.

Written by the indexing model from the issue text.