[KiPart BUG]: Bundle doesn't work as expected when stacking power and passive pins as per KLC S4.3

[cnieves1]

Kicad library rule:

https://klc.kicad.org/symbol/s4/s4.3/

Specifies that invisible pins when stacking should be have passive pin type.

However, KiPart only bundles pins when those pins are power type.

Changing:

def is_pwr(pin, fuzzy_match):
    """Return true if this is a power input pin."""
    return (
        find_closest_match(name=pin.type, name_dict=PIN_TYPES, fuzzy_match=fuzzy_match)
        == "W"
    )

to

def is_pwr(pin, fuzzy_match):
    """Return true if this is a power input pin."""
    return (
        (find_closest_match(name=pin.type, name_dict=PIN_TYPES, fuzzy_match=fuzzy_match)
        == "W") or 
        (find_closest_match(name=pin.type, name_dict=PIN_TYPES, fuzzy_match=fuzzy_match)
        == "P")
    )

does the trick.

[dnschneid]

Does #63 do what you expect? Note that it takes care of changing the type to "passive" and making it invisible for you.

[cnieves1]

Thanks for your answer!

It almost does it. The only missing feature is that it appends "[n]" string, where n is the number of bundled pins.

Official Kicad library symbols don't do that. Any chance to remove that?

[cnieves1]

As some people would like to keep the current behaviour, what about adding an option to do that?

diff --git a/kipart/kipart.py b/kipart/kipart.py
index 4b6ccef..9cc06b8 100644
--- a/kipart/kipart.py
+++ b/kipart/kipart.py
@@ -502,6 +511,8 @@ def draw_symbol(
     fill,
     box_line_width,
     push,
+    annotate_pins,
 ):
     """Add a symbol for a part to the library."""
 
@@ -618,13 +629,14 @@ def draw_symbol(
         transform = {}
 
         # Annotate the pins for each side of the symbol.
-        for side_pins in list(unit.values()):
-            annotate_pins(list(side_pins.items()))
+        if annotate_pins:
+            for side_pins in list(unit.values()):
+                annotate_pins(list(side_pins.items()))
 
         # Determine the actual bounding box for each side.
         bbox = {}
         for side, side_pins in list(unit.items()):
@@ -805,6 +818,8 @@ def kipart(
     reverse=False,
     fuzzy_match=False,
     bundle=False,
+    annotate_pins=True,
     debug_level=0,
 ):
     """Read part pin data from a CSV/text/Excel file and write or append it to a library file."""
@@ -845,6 +860,8 @@ def kipart(
             fill=fill,
             box_line_width=box_line_width,
             push=push,
+            annotate_pins=annotate_pins,
         )
 
 
@@ -891,6 +908,8 @@ def call_kipart(args, part_reader, part_data_file, file_name, file_type, parts_l
         reverse=args.reverse,
         fuzzy_match=args.fuzzy_match,
         bundle=args.bundle,
+        annotate_pins=args.annotate-pins,
         debug_level=args.debug,
     )
 
@@ -985,6 +1004,18 @@ def main():
         action="store_true",
         help="Bundle multiple, identically-named power and ground pins each into a single schematic pin.",
     )
+    parser.add_argument(
+        "-t",
+        "--annotate-pins",
+        action="store_true",
+        help="Annotate pin names if there when stacking pins'",
+    )
     parser.add_argument(
         "-a",
         "--append",

Thoughts?

[dnschneid]

Yeah, I personally prefer [n-1:0] rather than [n], so something like that where you can pick 0, 1, or 2 would mean I wouldn't have to maintain my own patchset to get the format I want :)

[cnieves1]

Mmm... I'm curious... what features where added or modified with your patchset? I noticed another deviation from KLC, which is that symbols should be centered... do you have any patch to do this?

[dnschneid]

Nope. Mostly just adding additional fields and formatting/positioning them differently.

[dnschneid]

PR #64

[cnieves1]

Great!! Thanks a lot!! I will use both! Looking forward to getting both MR merged!

[devbisme]

kipart currently only works with version 5 symbol libraries. What's your current solution to using it for version 6? Just import it using KiCad6? (Or are you not using version 6?)

[dnschneid]

Yeah, I just import into KiCad6. So far the only thing I've missed (at least, among the new features that would make sense for csv-based symbol generation) is support for the "free" pin type, which is relatively rare and easy to apply afterwards.

[devbisme]

You can try the updated kipart using this command:

pip install git+https://github.com/devbisme/kipart

Let me know if there are any problems. Then I can upload it to PyPi.

[cnieves1]

It works. Yet, when there are bundled pins, pin number of the visible pin is not shown.

There are parts with stacked pins in Kicad library, where the pin number of the visible pin is shown (CC430F5137xRGZ for example).

I think kipart behaviour should stick with existing rules...

[dnschneid]

Yeah, my first patch hid the pin numbers, since I figured showing one random pin out of a stack was a bit arbitrary. I suppose, though, it makes sense to show one of the pin numbers for stacked power nets (if you need to probe one pin, you've probed them all), but I don't think it makes sense to show any pin numbers for stacked NC pins.

[dnschneid]

#65 implements that.

[dnschneid]

I think this can be closed now?

[cnieves1]

Yes! Thanks!