Add support for Multicolored Wires #17

aakatz3 · 2020-06-25T01:28:16Z

Closed #12, adding support for multicolored wires, along with example. Also extends the DIN color code, and adds the following color codes:

T568A
T568B
even-count color code/25 pair color code/Telco color code

Also add in the 25 pair code, and the full DIN code

kimmoli

No solution to provide, but seems the Wire table shows a gradient of two colors, not stripes as the wire

I tried to make the wires look like as real wires in these DIN cables, but seems graphviz edge can not do it easily.

aakatz3 · 2020-06-25T13:35:39Z

GraphViz can make the wires banded in the stripes, however it causes half of the wire to be one color, the other half to be the other, and would require removing the black borders. I ended up adopting the twisted stripe commonly seen in CAT5/CAT6A cables.

For the gradient, It may be possible to have it add multiple "cells" that have the additional wire colors (this may be the best option so the thickness matches), but I thought it was likely better to have the gradient than to just have the primary color, so I left that default behavior in.

The gradient behavior happens due to line 152 of wireviz.py:
bgcolor = ":".join(wv_colors.translate_color(x, 'hex').split(':')[:2])

This line suppresses the warning about additional colors being ignored. To just have the major/first color, you can either change the 2 to a 1, or replace the line with the following:

bgcolor = wv_colors.translate_color(x[:2], 'hex')

kvid · 2020-06-25T13:46:46Z

src/wireviz.py

@@ -141,7 +142,7 @@ def create_graph(self):
                for bla in p:
                    html = html + '<td>{}</td>'.format(bla)
                html = html + '</tr>'
-                bgcolor = wv_colors.translate_color(x, 'hex')
+                bgcolor = ":".join(wv_colors.translate_color(x, 'hex').split(':')[:2])
                html = html + '<tr><td colspan="{colspan}" cellpadding="0" height="6" bgcolor="{bgcolor}" border="2" sides="tb" port="{port}"></td></tr>'.format(colspan=len(p), bgcolor=bgcolor if bgcolor != '' else '#ffffff', port='w{}'.format(i))


I suggest changing height="6" to height="10" for two-colored wires to match the extra thickness of the connecting curved wires. The thickness seems to be 2x the number of color stripes (black edge + base color + stripe color + base color + black edge).

An even better improvement (to avoid the color gradient) is to build the striped wire using multiple HTML rows with the port attribute only present in the middle row for each wire. See this excerpt from the upper wire in ex07.gv where the black edges are separated rows:

<tr><td colspan="3" cellpadding="0" height="2" bgcolor="#000000" border="0" sides="tb"></td></tr> <tr><td colspan="3" cellpadding="0" height="6" color="#ffffff" bgcolor="#00ff00" border="2" sides="tb" port="w1"> <tr><td colspan="3" cellpadding="0" height="2" bgcolor="#000000" border="0" sides="tb"></td></tr>

It might be easier to to use one row (with no border) for each color in the generic case with multiple stripes.

SVG provides dashes ( https://developer.mozilla.org/en-US/docs/Web/SVG/Attribute/stroke-dasharray ) which provide a simple and clean way to add this. Draw a solid wire then a dashed wire.

I just saw your changes, if you are able to, please check out my solution and let me knnow if you have any comments. I will try to review your pull request shortly.

I just created this PR: aakatz3#4 with my suggested implementation.

@njhurst While it would be nice to do this, there is a problem: Graphviz does not support this properly. When drawing dashed lines, you cannot specify an alternate color, and you can't seem to overlay them properly. One COULD modify the SVG after the fact, but it breaks compatibility with GraphViz

… wires

# Conflicts: # src/wireviz.py

aakatz3 · 2020-06-25T17:38:07Z

@kvid I merged your pull request in, please let me know if it looks good

kvid · 2020-06-25T18:12:16Z

Thank you. It looks good, except I found a minor issue in the odd case of a wire with more than two colors. See aakatz3#5

Reverse colors to match the curved wires when more than 2 colors

aakatz3 · 2020-06-25T21:34:12Z

Fixed and merged to my branch. Thank you!

X-Ryl669 · 2020-06-26T19:37:41Z

I think once there is a multicolored wire, all wires should be drawn with extended thickness, as it does not look very good with mixed thickness wires.

kvid · 2020-06-26T20:11:17Z

@X-Ryl669 Do you prefer equal wire thickness within the same cable or globally in all cables? If we go in this direction, then maybe a difference in cable gauge also optionally could be visualized as a difference in wire thickness between different cables. Equal wire thickness can be implemented using the same number of color entries for these wires internally and for Graphviz, i.e. several equal color entries for unicolored wires.

X-Ryl669 · 2020-06-27T07:25:04Z

Yes, it makes sense to have wire with same gauge using the same thickness (whether they are multi-colored or not).
Having wire with higher thickness for higher gauge is even better (but as I understand it currently, the gauge value is not interpreted, only used for mapping to AWG).

17o2 · 2020-06-27T08:16:41Z

This looks very promising!

@njhurst While it would be nice to do this, there is a problem: Graphviz does not support this properly. When drawing dashed lines, you cannot specify an alternate color, and you can't seem to overlay them properly. One COULD modify the SVG after the fact, but it breaks compatibility with GraphViz

It is indeed possible to define multi-colored dashed lines in GraphViz (see example here); however, it might be hard to make it look good, especially making it look even between the GraphViz edges and the HTML table cell (which I guess could be subdivided to simulate stripes?).

While it would be cool, it might make the code unnecessarily complicated. What do you think?

kvid · 2020-06-27T11:07:37Z

AFAIK, the wire gauge value is not currently inerpreted, but one approach could be to sort all unique gauge values, use the minimum thickness for the smallest mm^2 value (or highest AWG), add a constant to the thickness for the next gauge value, etc. If we need a maximum thicness limit, then close gauge values can be grouped and use equal thickness. Alternatively, a less functional, but simpler implementation could be to add an optional wire thickness attribute in the YAML file.

17o2 · 2020-06-27T12:22:23Z

Perhaps it makes sense to split the wire gauge visualization into a separate issue, so as not to further complicate things. It would be a nice addition but is not necessarily related to banded/striped wires.

It will probably be necessary to allow for both lengthwise stripes (like Ethernet or PE) and color rings (like DIN 47100)... although lengthwise stripes seem to be easier to implement.

X-Ryl669 · 2020-06-27T12:53:24Z

Agreed. It's just that this current implementation enlarge the wire if it's multicolor but doesn't do that if it's monochrome.

aakatz3 · 2020-06-27T19:37:18Z

In order to make the dashed lines work properly in graphviz, they would need to lose the black outlines, and we would need a way to count and calculate the number of stripes that should be placed. While you can put two or more colors into multicolored lines, I couldn't find a way to pattern them without just repeating, over and over, the color code.

A possible solution to enlarging single color wires instead of just tripling up the color may be to change the stroke width, but I do not know if this can be done for only the center spline/stroke.

17o2 · 2020-06-28T11:05:22Z

I've recently made some rather big refactoring changes (#9), is it possible to rebase these changes onto the latest version to make merging easier?

aakatz3 · 2020-06-30T18:48:18Z

@formatc1702 I will try my best to do this ASAP. I will have to check out the new code, and will provide an update when I can.

# Conflicts: # .gitignore # src/batch.py # src/wireviz.py # src/wv_colors.py

17o2 · 2020-07-15T17:39:17Z

There hasn't been any activity in the last 10 days, but I really would like to merge this ASAP.

@aakatz3 any news?
@kvid do you have the time to take over this issue, look over the requested changes and get this ready to merge into the current dev branch with no merge conflicts?

Only addendum to the code comments: Ideally please revert any changes to the example output files, keeping only the input .yml files, so the diff of the commit is not so big. I will rebuild and publish all examples when we merge into master.

kvid · 2020-07-15T22:31:46Z

I'm a bit busy at day-time these days, but I can put in some effort at evenings to get this PR ready. Unless @aakatz3 suddenly gets back, I suggest I clone his branch aakatz3:feature-multicolor-wires and address the change requests and merge conflicts. Then I can create a new PR with my version of the branch referring to this one. This way @aakatz3 or someone else can continue this PR in case I fail or use too much time trying.

17o2 · 2020-07-16T06:24:23Z

Thanks, appreciate it. See if you can rebase your feature branch onto your dev, I want to avoid having a tangle in the commit tree in the end :)

aakatz3 · 2020-07-16T12:32:08Z

I've been working on trying to rebase my commit tree, but I keep running into issues within the HTML, so I was trying to look into refactoring the HTML. I don't quite know the best way forward. I may do a squash rebase and submit a new pull request. Do you have any thoughts on the best way to tackle this?

On a separate note, I found the problem with UTF-8, encodings weren't hardcoded or detected, so on windows, it would write utf-8 but read as codepage2512, which caused the issue (or vice versa). My personal git tree is a major mess at this point trying to figure this out.

# Conflicts: # examples/ex01.bom.tsv # examples/ex01.gv # examples/ex01.html # examples/ex01.png # examples/ex01.svg # examples/ex05.gv # examples/ex05.html # examples/ex05.png # examples/ex05.svg # src/wireviz/Harness.py # src/wireviz/build_examples.py

17o2 · 2020-07-16T13:01:31Z

@aakatz3 glad you're back!

Honestly, so much has changed since work on this issue started, my naïve approach would be to start from the current state of the dev branch, and manually build in the changes there.
No sense in desperately trying to untangle the commit tree! Not sure if the other co-authors take issue with removing the development history of this feature, but pragmatically speaking, it would be the easiest; just act as if you were starting from scratch from the current state of things.

17o2 · 2020-07-16T13:04:57Z

This might mean starting a clean branch from dev, porting your changes to there, and submitting a new PR. Let me know if you want to move forward with this.

kvid · 2020-07-16T13:30:31Z

What are the major differences between your last pushed commit and your personal commits? If you push your last commits in a separate temporary branch, then I can have a look and perhaps suggest a way forward.

kvid · 2020-07-16T16:01:01Z

I am thinking it's better to merge than to rebase in this case to avoid a lot of non-trivial conflicts because files have been moved in both branches. I sugggest creating a branch before Merge branch 'dev' into feature-multicolor-wires to reduce the tangles, and we also avoid the two commits after this point that was not wanted by @formatc1702 as far as I understand (if any of them is wanted, they can be cherry picked afterwards):

Add Jetbrains IDE config
Update all examples

I tried this in my personal repo:

git branch aakatz3-multicolor e6317536f75ef5ed45dde0133c973a13145d704f
git checkout dev
git merge aakatz3-multicolor

I had to resolve these conflicts:

.gitignore (keeping changes from both branches)
src/wireviz.py (deleting file)
src/wireviz/build_examples.py (keeping the highest number of examples and tutorials)
src/wv_colors.py (deleting file)

After this merger, I could build all examples without errors, but I have not yet checked if all changes was merged in correctly.

17o2 · 2020-07-16T18:16:41Z

I am thinking it's better to merge than to rebase in this case to avoid a lot of non-trivial conflicts because files have been moved in both branches.

That's precisely why I am advocating for a fresh start from dev, cleanly re-implementing the desired changes and getting rid of the messy history.
Choosing rebase over merge (or vice-versa) does nothing to avoid conflicts, they would have to be resolved either way.

aakatz3 · 2020-07-16T23:20:16Z

Dev-reset complete, please see #96

kvid · 2020-07-17T03:37:16Z

I did rebase aakatz3-multicolor described above onto dev and adapted the commits to use f-strings:
https://github.com/kvid/WireViz/tree/aakatz3-multicolor-rebased
Then I discovered the new PR #96 from @aakatz3 and stopped. It was a nice excercise though. Each commit is verified by running the build script and inspecting some critical diagrams. It would be fun to compare this with the commits in PR #96 if I have the time. :-)

17o2 · 2020-07-17T05:09:58Z

I've merged #96 into my own feature branch, will tweak and merge into dev soon. Closing this one. Thanks everyone for the hard work.

aakatz3 added 3 commits June 24, 2020 20:57

Ignore IDE settings and windows stuff

10e747c

Add support for Multicolored (banded) wires

21bc9b1

Also add in the 25 pair code, and the full DIN code

Add example for banded wires

dccdfa7

kimmoli reviewed Jun 25, 2020

View reviewed changes

kvid reviewed Jun 25, 2020

View reviewed changes

kvid and others added 2 commits June 25, 2020 18:09

Create multicolored wires with multiple HTML rows to match the curved…

f23c186

… wires

Fix table color issue

89048d0

kvid mentioned this pull request Jun 25, 2020

Create multicolored wires with multiple HTML rows to match curved wires aakatz3/WireViz#4

Merged

Merge #1

3562c0e

# Conflicts: # src/wireviz.py

Reverse colors to match the curved wires when more than 2 colors

b253d0a

Merge pull request #5 from kvid/kvid/feature-multicolor-wires

9cd46e4

Reverse colors to match the curved wires when more than 2 colors

17o2 changed the base branch from master to dev June 27, 2020 10:13

kvid mentioned this pull request Jun 28, 2020

[feature] Visualize wire gauge differences as differences in wire thickness #29

Open

17o2 mentioned this pull request Jun 29, 2020

Cleanup/Refactor the code #39

Merged

aakatz3 added 2 commits July 1, 2020 12:39

Rename examples to allow merge

e631753

Merge branch 'dev' into feature-multicolor-wires

f0d10bf

# Conflicts: # .gitignore # src/batch.py # src/wireviz.py # src/wv_colors.py

17o2 force-pushed the dev branch 4 times, most recently from 7534300 to 0cef5e3 Compare July 11, 2020 21:23

Prepare for dev reset and re-merge

e4a034c

aakatz3 mentioned this pull request Jul 16, 2020

Allow cleaning of examples #94

Merged

aakatz3 added a commit to aakatz3/WireViz that referenced this pull request Jul 16, 2020

Change colors as per wireviz#17 review

7ab4971

aakatz3 force-pushed the feature-multicolor-wires branch 2 times, most recently from 2c063da to e4a034c Compare July 16, 2020 23:16

aakatz3 marked this pull request as draft July 16, 2020 23:17

aakatz3 mentioned this pull request Jul 16, 2020

Implement Feature/multicolor wires on refactored code #96

Merged

17o2 pushed a commit that referenced this pull request Jul 17, 2020

Change colors as per #17 review

eecf925

17o2 closed this Jul 17, 2020

kvid mentioned this pull request Jul 18, 2020

[bug] GitHub Actions build_examples.py output not showing up #100

Closed

17o2 pushed a commit that referenced this pull request Jul 19, 2020

Change colors as per #17 review

ea948a2

martonmiklos mentioned this pull request Sep 12, 2020

Remove HTML links from the input attributes #164

Merged

kvid mentioned this pull request Sep 13, 2020

Syntax documentation + new /docs directory #111

Merged

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add support for Multicolored Wires #17

Add support for Multicolored Wires #17

aakatz3 commented Jun 25, 2020

kimmoli left a comment

aakatz3 commented Jun 25, 2020

kvid Jun 25, 2020

kvid Jun 25, 2020 •

edited

Loading

njhurst Jun 25, 2020

aakatz3 Jun 25, 2020

kvid Jun 25, 2020

aakatz3 Jun 25, 2020

aakatz3 commented Jun 25, 2020

kvid commented Jun 25, 2020

aakatz3 commented Jun 25, 2020

X-Ryl669 commented Jun 26, 2020

kvid commented Jun 26, 2020 •

edited

Loading

X-Ryl669 commented Jun 27, 2020

17o2 commented Jun 27, 2020 •

edited

Loading

kvid commented Jun 27, 2020 •

edited

Loading

17o2 commented Jun 27, 2020

X-Ryl669 commented Jun 27, 2020

aakatz3 commented Jun 27, 2020

17o2 commented Jun 28, 2020

aakatz3 commented Jun 30, 2020

17o2 commented Jul 15, 2020

kvid commented Jul 15, 2020

17o2 commented Jul 16, 2020

aakatz3 commented Jul 16, 2020

17o2 commented Jul 16, 2020 •

edited

Loading

17o2 commented Jul 16, 2020

kvid commented Jul 16, 2020

kvid commented Jul 16, 2020 •

edited

Loading

17o2 commented Jul 16, 2020 •

edited

Loading

aakatz3 commented Jul 16, 2020

kvid commented Jul 17, 2020 •

edited

Loading

17o2 commented Jul 17, 2020

Add support for Multicolored Wires #17

Add support for Multicolored Wires #17

Conversation

aakatz3 commented Jun 25, 2020

kimmoli left a comment

Choose a reason for hiding this comment

aakatz3 commented Jun 25, 2020

kvid Jun 25, 2020

Choose a reason for hiding this comment

kvid Jun 25, 2020 • edited Loading

Choose a reason for hiding this comment

njhurst Jun 25, 2020

Choose a reason for hiding this comment

aakatz3 Jun 25, 2020

Choose a reason for hiding this comment

kvid Jun 25, 2020

Choose a reason for hiding this comment

aakatz3 Jun 25, 2020

Choose a reason for hiding this comment

aakatz3 commented Jun 25, 2020

kvid commented Jun 25, 2020

aakatz3 commented Jun 25, 2020

X-Ryl669 commented Jun 26, 2020

kvid commented Jun 26, 2020 • edited Loading

X-Ryl669 commented Jun 27, 2020

17o2 commented Jun 27, 2020 • edited Loading

kvid commented Jun 27, 2020 • edited Loading

17o2 commented Jun 27, 2020

X-Ryl669 commented Jun 27, 2020

aakatz3 commented Jun 27, 2020

17o2 commented Jun 28, 2020

aakatz3 commented Jun 30, 2020

17o2 commented Jul 15, 2020

kvid commented Jul 15, 2020

17o2 commented Jul 16, 2020

aakatz3 commented Jul 16, 2020

17o2 commented Jul 16, 2020 • edited Loading

17o2 commented Jul 16, 2020

kvid commented Jul 16, 2020

kvid commented Jul 16, 2020 • edited Loading

17o2 commented Jul 16, 2020 • edited Loading

aakatz3 commented Jul 16, 2020

kvid commented Jul 17, 2020 • edited Loading

17o2 commented Jul 17, 2020

kvid Jun 25, 2020 •

edited

Loading

kvid commented Jun 26, 2020 •

edited

Loading

17o2 commented Jun 27, 2020 •

edited

Loading

kvid commented Jun 27, 2020 •

edited

Loading

17o2 commented Jul 16, 2020 •

edited

Loading

kvid commented Jul 16, 2020 •

edited

Loading

17o2 commented Jul 16, 2020 •

edited

Loading

kvid commented Jul 17, 2020 •

edited

Loading