Extracting X Y values from a .txt file

I am working on a small project using data for a 3D printer. My problem is I am having a hard time extracting the movements of the printer to a usable array. In my text file I have things like:

;TYPE:SKIRT

G1 F1500 X18.188 Y1.255 E0.36904

G1 X18.188 Y20.575 E1.33292

G1 X26.144 Y20.575 E1.72984

G1 X26.144 Y26.455 E2.02320

;TYPE:WALL-INNER

G1 F1500 X16.988 Y2.455 E9.77394

G1 X16.988 Y21.775 E10.73782

G1 X24.944 Y21.775 E11.13474

;TYPE:WALL-OUTTER
...

The code then continues with this same format for each layer. So if there are 50 layers there are 50 wall outer sections.

If all I want is each x and y numbers for all the wall outer values (while making note of which layer the xs and ys were for) how would I do that? I have been searching the net and have found nothing for something I would think would be very simple.

[jonas]

Can you provide part of the txt file?

[Paolo]

Is the expected output for the example:

 16.988 2.455 

 16.988 21.775 

 24.944 21.775

?

[James Andrew]

Is the expected output for the example:

 16.988 2.455 
 16.988 21.775 
 24.944 21.775
?

If I were looking for the x and Ys of wall inner yes. Those would be the values. But like I said there are multiple blocks of wall inner and I would like to get the values of each block thought the whole file and making note which block they belong to.

https://imgur.com/41vaVKv So you can see at layer 0 these are the commands and I want all of the x and Ys that have a G1 in the front inside the block between Wall inner and wall outer (AKA getting all the moves that make the inner wall) That being said there are OTHER wall inners thought the text for each layer https://imgur.com/41vaVKv Here is further down in the code where a new block of commands are made for wall inner.

I really hope I am not asking for too much I just have no idea how to do this. I am currently trying to solve my problem myself by first converting the txt file into a table but it is having problems of its own.

[Paolo]

Not a problem. Can you upload your text file please?

[James Andrew]

This is it

[James Andrew]

Still trying to get the table to work but having just as much trouble.

Answer 1

[Paolo]

I think the best solution in this situation is to use regular expressions. Matlab's regexp supports named capture groups, which automatically creates a structure for you, with the names of the capture groups as fields and the captured text as data. You can read more about it here.

There are three steps in the solution below.

1. Initially the text file is parsed and each layer is extracted.

2. A loop is used to iterate over every layer to extract the types.

3. A inner loop is used to iterate over every type in every layer, and X Y data is extracted.

Here's the code:

 alldata = fileread('TGC.txt');
 layers = regexp(alldata,'(?:LAYER:\d)(?<layer>[\w\s.;:-]*?)(?=;LAYER:|;End)','names');
 n = size(layers);
 for i=1:n(2)
     [layers(i).types pos] = regexp(layers(i).layer,'(?:;TYPE:)(?<TYPE>[\w-]*)','names');
     pos = [pos length(layers(i).layer)];
     m = size(layers(i).types);
     for j=1:m(2)
         layers(i).types(j).data = regexp(layers(i).layer(pos(j):pos(j+1)),'(?:G1[\w\s]*X)(? <XDATA>\d*\.\d*)(?:\sY)(?<YDATA>\d*\.\d*)','names');
     end
 end

The output is a structure where you can easily access your data. I am attaching screenshots to clarify what the code does.

1. There are 58 layers in your code (your index starts from 0).

2. Each of your layers has several types:

3. Each type contains X and Y data:

To access your data you can use the dot notation. If you want the X and Y data for the first layer and the first type:

 layer(1).types(1).data

Note that for numerical calculations you will need to convert your data from char to double.

[dpb]

Ah! Someone came along that does know regular expressions! :)

Altho the new(ish) string package has functions that allow for higher-level abstraction for many cases...here the big problem was that last record is missing the 'E' field and there are inconsistent numbers of fields to skip per record.

[Paolo]

Always learning :) Questions like this one are great for practicing.

Yes, I found OP's instructions slightly confusing so I figured I might as well get all of the data for him to access.

[dpb]

I've decided I'm too old a dog for that new trick...no longer the patience needed! :)

Answer 2

[dpb]

You'll have to parse that file to locate the desired sections...none of the builtin reading tools are able to do such a file structure on their own.

fid=fopen('TGC.txt');
while ~feof(fid)  % begin a loop
  l=fgetl(fid);   % read a line
  if contains(l,';Layer count:')  % found the number layers
    nLay=sscanf(l,';Layer count: %d');   % read the # layers
    break                                   % now can go counted loop
  end
end
XY=cell(nLay,1);                          % preallocate a cell arrray that size
for i=1:nLay                              % go until get 'em all...
  l=fgetl(fid);                           % read record
  while ~contains(l,';TYPE:WALL-OUTER'),l=fgetl(fid);end       % find the desired section
  g={fgetl(fid)};                         % get first record in section
  while ~contains(g,';')                  % and build cellstr array
    g=[g;{fgetl(fid)}];
  end
  g(end)=[];                              % remove the trailing header line
  g(end)=strcat(g(end),{' E'});           % normalize the row delimiters
  XY{i}=str2double([extractBetween(g,"X","Y") extractBetween(g,"Y","E")]);
end
fid=fclose(fid);

ADDENDUM Above seems to work with a sorta' klunky fixup...

>> james   % original name for script!!!  :)
>> whos XY
  Name       Size            Bytes  Class    Attributes

    XY        58x1             14848  cell               

>> XY{1}
ans =
                    17.388                     2.055
                    17.388                    21.375
                    25.344                    21.375
                    25.344                    25.655
                     3.616                    25.655
                     3.616                    21.375
                    11.591                    21.375
                    11.591                     2.055
                    12.129                     2.795
>>

The layer isn't saved explicitly; it's inherent in the cell index(*); it appears for this file the size of each array is the same; not sure if that were to be a requirement, hence the cell array as opposed to, say, a 3D array by layer.

(*) Of course, it's 1-based instead of 0-based since ML arrays begin (only, unfortunately) at 1.