[WIP]Processing Jumpers and Loops, Plus Processing Wires inside the Table as Edges based on #251

[tobiasfalk]

Now, based on #251. Thanks, @kvid, for your help with git(https://github.com/wireviz/WireViz/pull/251#issuecomment-2214979990) I will implement a fix for the problem that was discussed in #369 with @martinrieder and I will look into the implementation of the discussion in #286 by @formatc1702 and @martinrieder.

[SnowMB]

Hey I tried your branch and ran into the following issues:

• Examples do not build. Somehow in the html tables a pair of brackets is emited outside of a element
  • It is even present in the already commitid example files e.g. examples/ex01.gv:

      ...
    <td>
     <table border="0" cellborder="1" cellpadding="3" cellspacing="0">
      <tr>
       <td>GND</td>
       []
       <td port="p1r">1</td>
      </tr>
      <tr>
       <td>VCC</td>
       []
       <td port="p2r">2</td>
      </tr>
      <tr>
       <td>RX</td>
       []
       <td port="p3r">3</td>
      </tr>
      <tr>
       <td>TX</td>
       []
       <td port="p4r">4</td>
      </tr>
     </table>
    </td>
    ...
  

  • I traced this to this lines src/wireviz/wv_graphviz.py:293

   def gv_short_row_part(pin, connector) -> List:
       short_row = []# Td("ADA"), Td("DAD")
       for short, shPins in connector.shorts.items():
          if pin.index + 1 in shPins:
               short_row.append(Td("", port=f"p{pin.index+1}j"))
           else:
               short_row.append(Td(""))
  -    return short_row
  +    return short_row if len(short_row) > 0 else None
  

  • The old wire table still exists even though you now draw the edge ontop of the html table. There seems to be a small bug, that the inner html of the old wiretable defines columnspan=5 even though this table has only one column:

  <tr>
       <td border="0" cellspacing="0" cellpadding="0" colspan="5" height="6" port="w1">
        <table border="0" cellborder="0" cellspacing="0">
         <tr>
          <td bgcolor="#FFFFFF" border="0" cellpadding="0" colspan="5" height="2"></td>
         </tr>
         <tr>
          <td bgcolor="#FFFFFF" border="0" cellpadding="0" colspan="5" height="2"></td>
         </tr>
         <tr>
          <td bgcolor="#FFFFFF" border="0" cellpadding="0" colspan="5" height="2"></td>
         </tr>
        </table>
       </td>
      </tr>
  

  • I guess the internal wire table can just be removed:

   def gv_wire_cell(wire: Union[WireClass, ShieldClass], colspan: int) -> Td:
  -    if wire.color:
  -        color_list = ["#000000"] + wire.color.html_padded_list + ["#000000"]
  -    else:
  -        color_list = ["#000000"]
  -
  -    wire_inner_rows = []
  -    for j, bgcolor in enumerate(color_list[::-1]):
  -        wire_inner_cell_attribs = {
  -            "bgcolor": "#FFFFFF", # bgcolor if bgcolor != "" else "#000000", # TODO: More elegent solution for making black/whit space needed, since the wire is drawn as an actual edge
  -            "border": 0,
  -            "cellpadding": 0,
  -            "colspan": colspan,
  -            "height": 2,
  -        }
  -        wire_inner_rows.append(Tr(Td("", **wire_inner_cell_attribs)))
  -    wire_inner_table = Table(wire_inner_rows, border=0, cellborder=0, cellspacing=0)
       wire_outer_cell_attribs = {
           "border": 0,
           "cellspacing": 0,
           "cellpadding": 0,
           "colspan": colspan,
  -        "height": 2 * len(color_list),
  +        "height": 6, # is 6 correct here? before it was 6 for single color wires
           "port": f"w{wire.index+1}",
       }
       # ports in GraphViz are 1-indexed for more natural maping to pin/wire numbers
  -    wire_outer_cell = Td(wire_inner_table, **wire_outer_cell_attribs)
  +    wire_outer_cell = Td(None, **wire_outer_cell_attribs)
   
       return wire_outer_cell
  

  Should I create a merge request against your branch?

[tobiasfalk]

@SnowMB Yes please, did not see that there is a problem with the html output.

[tobiasfalk]

The old wire table still exists even though you now draw the edge ontop of the html table. There seems to be a small bug, that the inner html of the old wiretable defines columnspan=5 even though this table has only one column:

This was intentionally, since I was a bit leasy and it was the fastest🤦‍♂️

[ssweber]

Interested in this. (To anyone else you can see changes more clearly by comparing to refactor/big-refactor

[tobiasfalk]

See #455