Rename wires

Well, I'm going to show my Unix bias here, but there are a *lot* of tools available to do things like this built in to Linux and OS X and available for Windows.  It depends how deep you want to go down the rabbit hole!

My first reaction would be to use "sed" the stream editor.  (Do a search for "sed for windows".)  Sed is used at the command line, and takes some input, modifies it, and sends it to some output.  For example,

sed 's/W500/W512' schematic.sch > newschematic.sch

which would change occurrences of W500 to W512 and create a new file called newschematic.sch without changing the original schematic.sch.  To make sure things look good, you could use the "diff" command on the two files to see exactly what was changed.  Of course, you wouldn't want to have to change that command line and re-run it for every substitution you want to make, so it's better to create a sed script file with all the commands you want (somewhat analogous to that CSV file in the referenced article), and then run the command:

sed -f script.file schematic.sch > newschematic.sch

Sed is a great tool and can be used over and over again without having to modify and recompile.  Now, if you want to learn something a little more sophisticated, there is a simple, yet amazing text processing language called AWK.  AWK knows how to read in a text file line by line, split the line into words and do things to those words.  You can do more testing and manipulating than you can in sed.  AWK is very much like the C programming language, so you can do things like printf to create new output.  Then, if you want to really go down the rabbit hole and get into more powerful scripting languages, the big three in my book are PERL, Python and Ruby.  Each has strengths and weaknesses and each has devout followers and detractors.  Let me know if you're interested or do a search to learn more...

BTW, if you don't want to download and install any of these unix tools, you might consider using Windows PowerShell too...

Thank you for your help.

after some tests I found out that Win doesn't like the following command:

sed 's/W500/W5000' 30c.sch 30d.sch

But the following seems to work so far (thank you google):

sed "s/W500/W5000/" 30c.sch 30d.sch

Well, first everything seems to work fine, it reads and schows all data in the 30c file, but after everything has finished, the 30d.sch is still empty. Any idea what I am doing wrong? Sorry I have to admit that I'm not the shell-guy...

As for the advice with the inconsistent board: Thank you for your advice, but there is no board, only the schematic. It is the shematic of my newly build camper. I know, eagle may not be the best solution for doing that, but works so far ok for me...

small thing, big consequences...thank you very much, I missed that one. Now everything works perfectly. In combination with the script file it is very powerfull.

For a more complex rename I'm now looking for a solution as well. Lets asume we have the following net names:

W100/1.5RD

W101/1.5RD

W102/50RD

W103/50BK

W104/1.5OR

W105/xxxx

...

W150/1.5RD

W151/1.5OR

These names are in no order somewhere in the schematic. Is it possible to sort these so that the first W1xx is on page 1 and so on. Important; the numbers and letters after the slash should not be changed. So for example if W150/1.5RD is changed to W103 it should still have the /1.5RD at the end. The "old" W103 may then become W104/50BK, W104 is changed to W105/1.5OR and so on...Hope you understand what I'd like to do.

Thank you very much for the great help.

My suggestions so far have been based on treating the schematic file as one big stream of text.  (That is what Unix was originally designed for, to process streams of text.). When we need to start thinking about the structures in a file, we need to use a different type of tool.

There are two categories of solution that come to mind.  We can either try to use the tools that are built in to EAGLE, or we can use one of those scripting languages I mentioned.  It's a trade off between specific knowledge and general applicability.  Are you more interested in learning more about EAGLE or learning how to write scripts that can apply to other programming problems?

Let's assume that you just want to get this problem solved, so we'll use available tools for this task and not reinvent anything.  EAGLE has a command prompt right there on the screen.  You can type commands instead of selecting them with a mouse.  Just for grins, when you have your schematic open, type "win fit" and press the enter key.  Just like a sed file, or an operating system batch file, you can put a series of these commands into an EAGLE script file and run it.  If you use a particular script a lot, you can assign it to a key or add it to a menu.  This is what makes EAGLE so appealing to an old Unix guy like me!

More in a bit.  I can't stand using this iPad to type...

Now interestingly enough, The EAGLE scripting language doesn't have any control structures.  That means there is no way to test things, or repeat things, other than to duplicate a line x number of times.  EAGLE scripts are just brute-force lists of commands to run.

So, for whatever reason, the Cadsoft folks decided to leave the scripting alone and add a different programming language to create scripts.  This is called "User Language" and you create User Language Programs (ULPs).  The interesting thing about ULPs is that they don't get to change things while they run, they have to generate EAGLE scripts that you run after the ULP is complete.  In some ways, this is a very Unix-like thing to do.  And I'm part German, (born in Frankfurt actually), so I have the in-group privilege to say it's a very German thing too!  LOL

SIDE NOTE: It would be amazingly cool if Cadsoft made all of the EAGLE commands available as library functions that could be called by other languages.  Personally, I'd love it if I could start up a Python program, import an EAGLE library and directly manipulate a schematic or board! 

So, anyway, User Language is very powerful, it knows about the structures in EAGLE files.  This saves *lots* of time!  You can write a ULP that works with the schematic, the board, a library, or parts of any of them.

I'm not sure I know *exactly* what you're trying to do, but my first interpretation is that you want to renumber wires based upon which page of the schematic they are on.  So, in a ULP, we would start by using the schematic, then looping through each page of the schematic.  The code to do that would be:

schematic(S){

  S.sheets(SH){

     // do something here

  }

}

The first line says we're going to work with the schematic and call it "S".  The next line says, take S, and go through each sheet in it one at a time and call it "SH" while we're using it.  I took a quick look at the EAGLE help, and I found that you can go through each wire of the sheet by adding:

schematic(S){

  S.sheets(SH){

    SH.wires(W){

      // do something with each wire here

    }

  }

}

But the problem seems to be that the name of each wire is not listed in the data members of the UL_WIRE structure.  I'm guessing we'll have to loop through something different, like nets() maybe:

schematic(S){

  S.sheets(SH){

    SH.nets(N){

      // do something with each net here

    }

  }

}

This illustrates the toughest part of writing a ULP, figuring out the parent-child relationships of all the structures.  Once you get the structure correct, then at the "do something" point, we output the command that does the actual rename.

I apologize, but I must depart for the day, my son turned 19 today and we have a party to prepare for.  I hope I've given you some hints as to how you might proceed.  I'm not sure, but I think I'd probably look at looping through segments() and labels() next.  Everything is documented in EAGLE help.  If you want to directly access the file with your web browser the way I do, open the file eagle_en.htm in the bin directory of your EAGLE installation, or you can get to the help files on my web site at http://analognotes.com/eagle/helpfiles/

If you learn from reading other people's code, go to the Cadsoft ULP download site and search for "renumber"...

Your suggestions so far (with sed) have been just great to change the names of roundabout 60% of the names I'd like to change (W3xx up toW9xx). To change the W1xx names are so to say a new problem.

You assume right, at the moment I just need to get this problem solved in Eagle. Going deeper into scripting and programming would need time I don't really have right now...

Win fit? Hmm, not sure how this special command would help me on that right now...for that I use Ctrl+F2...But while I'm writing about; is it just to tell me there are powerfull commands in Eagle? If so, I know that, but I have no Idea how to use them correctly to solve my problem. This is my first project in Eagle and I am learning by doing. That's also why I have to rename a lot. It startet small and back then the net names where not really an issue, but as things grew (up to 50 pages right now) I got more and more messed up on that...

I think the rename command would be my best friend, but to solve my problem in a reasonable amount of time and manual work, that one should probably bring some other friends along with. As for using scripts in Eagle, I played a little bit around with the scr and ulp functions in Eagle, but I have not really a clue how to use them correctly to solve my problem. Said that, i'd really apreciate if you could give me some advice how to do that.

By the way, if that script could also reorder the terminals who are aranged "chaotic" all over the shematic would be nice. I promess, this is the last "new" labelling issue I'll ask about ;-)

The terminals are labelled 11X01, 11X02..., 12X01, 12X02..., 14X01...

11X, 12X 14X should never change, the numbers after the X would be nice if they start on page one left  with 01 and then continue on. Mainly the same problem as with the W1xx but no numbers/letters at the end who schouldn't be changed.

Thank you very much

I think my last and your last posts just crossed each other. Anyway.

I think you understood me a bit wrong. I don't want to rename based on the page number.

Right now I have the following (just an example):

                           Now                       Should be

W100/1.5RD  on page 27                  1. W1xx wire on page 01

W101/1.5RD  on page 32                  2. W1xx wire on page 01

W102/50RD  on page 01                   3. W1xx wire on page 01

W103/50BK  on page 45                   1. W1xx wire on page 02

W104/1.5OR  on page 26                 2. W1xx wire on page 02

W105/xxxx  on page 14                    1. W1xx wire on page 03

...                                                     ...

W149/1.5OR  on page 06                  5. W1xx wire on page 26

W150/1.5OR  on page 01                  1. W1xx wire on page 27

Now the old W100 (on page 27) will be renamed to, W150 (assumed it is the 150iest W1xx wire in the shematic). But the last part of the name should be kept as it was before. The last part of the name is the dimension and the color of the wire, so it should stick to that wire where it is right now. So new we would have:

W100/1.5OR  on page 01 (because the wire on Page 01 is a 1.5OR)

W150/1.5RD  on page 27 (because the wire on page 27 is a 1.5RD)

I'll have a look at your sources...and try around a little bit. Happy birthday to your son and have a nice party.

Btw. we could also write in german ;-) I've just chosen English because I tought the chances to find an Eagle crack would be bigger. Not sure if that would be apreciated if we go on in german. German would be easier, but at the end is also a foreign language for me. So its up to you...

While writing I just got another idea: I use sed to change all W1xx to W999 (or something else). I would then get W9900, W99901, W99902... Then I would only need a solution how to rename all the W999xx in the script one after the other to W100, W101... Then I get W100xx, W101xx, after that just delete the xx part. Not sure if sed can do that, and if it can how. Probably to simple? I may run in trouble on that wires who start on one page and continue on diferent other pages...?

Thank you

Ursicin Caminada wrote:

Win fit? Hmm, not sure how this special command would help me on that right now...for that I use Ctrl+F2...But while I'm writing about; is it just to tell me there are powerfull commands in Eagle?

Yes, that is just to point out that there are many powerful commands and learning them is a good thing.  Every command can be started from the command line instead of just the mouse...

I think you understood me a bit wrong. I don't want to rename based on the page number.

Right now I have the following (just an example):

                           Now                       Should be

W100/1.5RD  on page 27                  1. W1xx wire on page 01

W101/1.5RD  on page 32                  2. W1xx wire on page 01

W102/50RD  on page 01                   3. W1xx wire on page 01

W103/50BK  on page 45                   1. W1xx wire on page 02

W104/1.5OR  on page 26                 2. W1xx wire on page 02

W105/xxxx  on page 14                    1. W1xx wire on page 03

...                                                     ...

W149/1.5OR  on page 06                  5. W1xx wire on page 26

W150/1.5OR  on page 01                  1. W1xx wire on page 27

Yes, I must be misunderstanding, because it is looking to me like you rename all the wires starting on page one, then going to page two, then three...

Now the old W100 (on page 27) will be renamed to, W150 (assumed it is the 150iest W1xx wire in the shematic). But the last part of the name should be kept as it was before. The last part of the name is the dimension and the color of the wire, so it should stick to that wire where it is right now. So new we would have:

W100/1.5OR  on page 01 (because the wire on Page 01 is a 1.5OR)

W150/1.5RD  on page 27 (because the wire on page 27 is a 1.5RD)

To keep the last part of the name the same, you will have to use the string functions of EAGLE.  First, use the strchr() function to find the location of the "/" character, then use the strsub() function to extract the last part of the string, so etwas:

string oldname = "W154/1.5OR";

char c = '/';

int position = strchr(oldname, c);

string lastpart = strsub(oldname, position);

Now, lastpart should be "/1.5OR".  You can test this by going to the EAGLE control panel, then File->New->ULP and copy this and save it:

string oldname = "W154/1.5OR";

char c = '/';

int position = strchr(oldname, c);

string lastpart = strsub(oldname, position);

dlgMessageBox(lastpart);

Now open your schematic and then click the ULP button and choose the file you just saved.  the dlgMessageBox command will open a dialog box that should say "/1.5OR" and have an OK button.

To generate the full new name, use the "sprintf" command, so:

int number = 100;

string newname;

sprintf( newname, "W%d", number);

Now, newname is "W100".  Then, add the last part to the wire name with "+=", so:

newname += lastpart;

Test:

int number = 100;

string oldname = "W154/1.5OR";

char c = '/';

int position = strchr(oldname, c);

string lastpart = strsub(oldname, position);

string newname;

sprintf(newname, "W%d", number);

newname += lastpart;

dlgMessageBox(newname);

OK, lunch time.  I will post later about how to use this in a loop to process all wires...

OK, let's look at how we would loop through everything.  In a previous message, I talked about how EAGLE can loop through every sheet of a schematic.  Let's make a test to see how that works.  Create a new ULP or you can use the one we created in the last message and delete everything in it.  Let's start with this code, so copy and paste this into the ULP:

string sheetname;

string message;

schematic(S){

  S.sheets(SH){

    sprintf(sheetname, "Sheet: %d\n", SH.number);

    message += sheetname;

  }

}

dlgMessageBox(message);

Save it and then run it by clicking the ULP button, or you can type "run" and the filename at the command line in the schematic.  (I called my ULP "rename.ulp", so I type "run rename.ulp" and press the enter key.)  You should get a dialog box with a line for every page of the schematic.  EAGLE dialog boxes are not very smart, so it may extend below the bottom of the screen.  Press the escape key or the space bar or click on the close button of the window.

Now, let's make a list of all the nets on each page of the schematic.  Add in these lines after "message += sheetname;":

SH.nets(N){

  message += N.name;

  message += " ";

}

message += "\n";

The important things to know here are that we are going through all of the nets on the schematic page, and when we are using each one, it is called "N".  We add the name of the net to the message, then to make it readable, we are adding a space (" ").  After each page, we are adding a new line ("\n") to separate the pages.  So the whole thing should look like:

string sheetname;

string message;

schematic(S){

  S.sheets(SH){

    sprintf(sheetname, "Sheet: %d\n", SH.number);

    message += sheetname;

    SH.nets(N){

      message += N.name;

      message += " ";

    }

    message += "\n";

  }

}

dlgMessageBox(message);

Save that and run it.  Do you see all of the names of your wires?  If not, we are looking in the wrong place.  Let me know and we will try something else.  If you do see the names that you want to change, then we can proceed.

Now it's time to add in the name changing code from my last message.  Let's remove the space and newline entries first, then we need to define some more variables:

string oldname;

string newname;

string lastpart;

char c = '/';

int number = 100;

int position;

Put those lines below the sheetname and message variables.  Now the name changing lines:

oldname = N.name;

position = strchr(oldname, c);

lastpart = strsub(oldname, position);

sprintf(newname, "W%d", number);

newname += lastpart;

number += 1;

Do you see how I got oldname from the name of the net this time?  The last line here is the magic bit that makes the wire number count up.  Now, let's put in the lines that show what we have done so far in the dialog box.  The whole thing should be:

string sheetname;

string message;

string oldname;

string newname;

string lastpart;

char c = '/';

int number = 100;

int position;

schematic(S){

  S.sheets(SH){

    sprintf(sheetname, "Sheet: %d\n", SH.number);

    message += sheetname;

    SH.nets(N){

      oldname = N.name;

      position = strchr(oldname, c);

      lastpart = strsub(oldname, position);

      sprintf(newname, "W%d", number);

      newname += lastpart;

      number += 1;

      message += "Rename old: ";

      message += oldname;

      message += " to new: ";

      message += newname;

      message += "\n";

    }

  }

}

dlgMessageBox(message);

Run that and see if that looks like what you are trying to accomplish.  In the next message, I'll show you how to generate the script that actually renames everything...

Wow, thank you for your huge work and patience.

Unfortunately it doesn't look like a 5-minute-job to try out I'll try to go through that step by step, but that has to wait till tomorrow evening. You know, there are such unimportant things like work, sleeping and so on who have to be done as well

I really apreciate your help and how you explain it step by step. Like that I have a small chance in understanding what I'm doing... Thank you very much.

Good night

Doug Wellington wrote:

Ursicin Caminada wrote:

 

To keep the last part of the name the same, you will have to use the

string functions of EAGLE.  First, use the strchr() function to find the

location of the "/" character, then use the strsub() function to extract

the last part of the string, so etwas:

also have a look at strxstr, strrchr and strsplit / strjoin  could be

an alternative

--

Lorenz

CadSoft Guest wrote:

also have a look at strxstr, strrchr and strsplit / strjoin  could be an alternative

strxstr?  Do we want to give this poor guy a brain cramp already?  LOL!!!

Some people, when confronted with a problem, think “I know, I'll use regular expressions.” Now they have two problems.  -Jamie Zawinski*

For the unknowing/unsuspecting, strxstr() is a very powerful function that uses a thing called a "regular expression" to search for a pattern of characters.  I think the line I quoted from Jamie is hilarious, but I actually *LOVE* regular expressions.  We use them in text processing, for example, building a compiler, which is one of the most advanced things you can do in computer science.  I taught a class on regular expressions and there are entire books devoted to just regular expressions. 

*There is a humorous web site devoted to this kind of quote: Two Problems

EDIT: Here's an example of a regular expression ("regex"):

^[ \t]+|[ \t]+$

quoted from Wikipedia: Regular expression 

CadSoft Guest wrote:

also have a look at strxstr, strrchr and strsplit / strjoin  could be an alternative

strxstr?  Do we want to give this poor guy a brain cramp already?  LOL!!!

Some people, when confronted with a problem, think “I know, I'll use regular expressions.” Now they have two problems.  -Jamie Zawinski*

For the unknowing/unsuspecting, strxstr() is a very powerful function that uses a thing called a "regular expression" to search for a pattern of characters.  I think the line I quoted from Jamie is hilarious, but I actually *LOVE* regular expressions.  We use them in text processing, for example, building a compiler, which is one of the most advanced things you can do in computer science.  I taught a class on regular expressions and there are entire books devoted to just regular expressions. 

*There is a humorous web site devoted to this kind of quote: Two Problems

EDIT: Here's an example of a regular expression ("regex"):

^[ \t]+|[ \t]+$

quoted from Wikipedia: Regular expression 

Doug Wellington wrote:

CadSoft Guest wrote:

since when is element14 suppressing the sender of newsgroup post?

also have a look at strxstr, strrchr and strsplit / strjoin  could be

an alternative

 

strxstr?  Do we want to give this poor guy a brain cramp already? 

LOL!!! [...]

I know 8-)

I only mentioned it for completeness sake while pointing out that

there are more possibilities beside strchr and substr

--

Lorenz

CadSoft Guest wrote:

since when is element14 suppressing the sender of newsgroup post?

Yeah, this site is a bit wonky, eh?  Do you post from the Stratford Digital EagleCentral site or do you post directly via NNTP?  I finally gave up on those and use Element14 because I can edit posts.  The down side is that the Jive software Element14 uses really just seems to suck.  For example, when I click on the "CadSoft Guest" title at the top of your message, it is linked to Drew Fustini, and I'm guessing you aren't Drew!  As far as I can tell, post edits don't propagate off this site, and in general, things are hard to find, which may be more of an organization thing than a Jive thing, but it doesn't reflect well on either.  There's a webinar about the site in less than an hour, so maybe things will clear up a bit...

strxstr?  Do we want to give this poor guy a brain cramp already?

LOL!!! [...]

 

I know 8-)

 

I only mentioned it for completeness sake while pointing out that

there are more possibilities beside strchr and substr

Thanks.  Yeah, TIMTOWTDI as the Perl guys say...

The fun thing is, when we come full circle in this discussion, we'll get back to sed, which is actually a great regular expression tool... 

Doug Wellington wrote:

CadSoft Guest (aka Lorenz) wrote:

 

since when is element14 suppressing the sender of newsgroup post?

 

Yeah, this site is a bit wonky, eh?  Do you post from the Stratford

Digital EagleCentral site or do you post directly via NNTP?  I finally

gave up on those and use Element14 because I can edit posts.  The down

side is that the Jive software Element14 uses really just seems to

suck.  For example, when I click on the "CadSoft Guest" title at the top

of your message, it is linked to Drew Fustini, and I'm guessing you

aren't Drew!  As far as I can tell, post edits don't propagate off this

site, and in general, things are hard to find, which may be more of an

organization thing than a Jive thing, but it doesn't reflect well on

either.  There's a webinar about the site in less than an hour, so maybe

things will clear up a bit...

I'm posting directly via nntp (since I started using eagle, long long

before farnell bought cadsoft).

There is a nice quote regarding nntp vs. web bases forums:

Web based forums are like subscribing to 10 different newspapers and

having to visit 10 different news stands to pickup each one. Email

list-server groups and USENET are like having all of those newspapers

delivered to your door every morning.

I follow about 30-40 news groups and mailing list (using gmane.org as

mailing list to nntp bridge) regularly.

That wouldn't be possible using web based forums.

Using my new client I can skim over the new posts during my morning

coffee break.

--

Lorenz

Sorry to continue the off-topic hijack, but that's how these forum things seem to work... 

Lorenz wrote:

There is a nice quote regarding nntp vs. web bases forums:

Web based forums are like subscribing to 10 different newspapers and

having to visit 10 different news stands to pickup each one. Email

list-server groups and USENET are like having all of those newspapers

delivered to your door every morning.

I follow about 30-40 news groups and mailing list (using gmane.org as

mailing list to nntp bridge) regularly.

 

That wouldn't be possible using web based forums.

Using my new client I can skim over the new posts during my morning

coffee break.

[curmudgeon]

Indeed.  I've always hated forums, but it's the way of the world now for most of the topics I seem to be interested in.  I used to, and still do, love listservs (remember Bitnet?) and newsgroups because, as you say, they come to you instead of you having to remember to go to them.  It all seemed so simple back then.  I'm an old MH user too, so I have a *huge* directory structure of messages from both email and nntp.  Of course, for me at least, email isn't the best answer any more, as most people don't edit much, they just pile answers on top of the previous message(s) and most of them are RTF or HTML (in other words, much larger than they need to be).  How many times have you gotten a 90k message that only adds "+1" to all the previous quotes?    Forums like this seem to be nicer in that regard, as you have to explicitly click a button to quote the previous message...

[/curmudgeon]

On 04/03/14 20:02, Doug Wellington wrote:

That's English for "wise man who speaks truth"