How to best possibly protect .m file? - matlab

I have an .m file which I wish to share with my friends but I am not interested in giving .m file. Could someone help me with best possible ways to convert it to a file that is not decodable?
I tried converting it to .p file by simply typing pcode example.m
however I don't believe it is really protecting it. I was able to convert my .p file back to .m file with the following link. https://sites.google.com/site/sippeyfunlabs/matlab-hacks/please-do-not-secure-your-password-in-matlab-p-code
This actually confirms that my code is not protected.
It'll be nice if someone shares the best methodology to protect .m file and sharing.
Thanks

The link you provided yourself already indicates that it is very difficult to obfuscate MATLAB code:
In fact, MATLAB language is very difficult to be secured or even obfuscated.
This is due to the late binding (or dynamic binding) feature of
MATLAB. [...] The amount of meta information associated with this
feature basically forbid any attempt of adding code level security.
Simply put, if there is a MATLAB file, and it calls a function foo
inside it. Until the runtime, the MATLAB interpreter do not even know
if foo is a function stored in M file or a built-in function or a mex
function or even a workspace function handle. Thus, it must store foo
as is somewhere inside the generated P-code.
Also to best solution is already mentioned on that page:
If there is really a need to do this, using the good old binary is a
much better solution. Or you can put critical code on a server, away
from the user.
Create a binary, i.e. mex file (see the answer of Wolfie): Note that to some extent also binary code can be decompiled especially small ones.
Host your code on a server (and run the code server-side): This is the best method to protect your code. You should let the user upload the input for your script and return the result. You can also automatize this process using a matlab script to make this process transparent to the user.

You could build a mex file.
This will completely obfuscate your actual MATLAB code, since it will be written in C/C++/FORTRAN, but the algorithms will still be there if your friends are determined enough to look for them.

Related

Read / run a MATLAB script from a URL

Fairly simply question, but I cannot find an answer via Google or on Stack.
I have a use-case where it would be highly-preferable to simply read a .m MATLAB script from a URL.
How should I do this correctly?
<disclaimer>
Clearly only do this with files you have complete control of (and/or find a solution with better validation). This is a "dangerous" method as there is no check that you're not about to run a file which, for example, copies your entire harddrive to Bob's computer before corrupting it all. Bob and Alice might be spending the whole evening laughing at your embarrassing holiday photos.
Treat this more as a proof of concept than a how-to, it addresses your problem but by no means should be used in production code.
</disclaimer>
You'll need to use eval to evaluate code. Every time I mention eval I feel compelled to point out it's not recommended, in particular in this case because you could be evaluating whatever random code is living in that file on the web! In this case your only alternative is to save the file locally and call run.
Then you can use
eval(urlread('http://myscript.m'))
or, since urlread is not recommended (from the docs), you can use webread and specify that the output should be text in the options
eval(webread('http://myscript.m', weboptions('ContentType', 'text')))
Using webread appears to be really slow, not sure why when it's the recommended function. In light of this, urlread might be preferable.
There is a note in the webread docs which suggests you wouldn't even need to specify the weboptions
If a web service returns a MATLAB® file with a .m extension, the function returns its content as a character vector.
Although you suggested that webread returned a uint8 variable which didn't work.
If you'd like to save a file from a URL then run it, you can use websave and str2func like so:
fcnName = 'newscript'; % Name for the function/script file
websave([fcnName '.m'], 'http://myscript.m'); % Download and save it
fcn = str2func(fcnName); % Create function handle
fcn(); % Evaluate function/script
It should of course go without saying that you want to be really sure you can trust the source of the file, otherwise you're gonna have a bad time.

running compiled matlab from matlab

I am trying to run compiled MATLAB code (by mcc) from inside MATLAB in a way that I can avoid using another license that is required by the compiled code. We need this because we run this same specific code part again and again and execution is stuck due to license waiting. We don't want to buy tons of this specific license just to mass run the same part. Is there any way to do this? tutorial?
Is it possible to compile a .m file to dll/so and wrap it like a mex and call it from MATLAB on the fly? How would I pass and retrieve complex arguments?
According to
http://www.mathworks.de/products/compiler/description3.html
creating shared libraries should well be possible.
Concerning passing and retrieving complex arguments:
If you plan to use mex, I'd assume you should be able to call the shared-library "main"-function with any arguments you'd like, using the mxArray type that you'll have to use anyway.
To run the MATLAB-compiled code in MATLAB, you want codegen, part of the MATLAB Coder. See this blog post on generating C code from MATLAB. The alternative, deploying code with mcc/mbuild and then reloading it into MATLAB with loadlibrary is rather contorted, and I wouldn't advice it.

How to include a c-header with constants in Matlab Simulink

I'm developing a Simulink modell with many C-s-functions. For an easier handling I want to use constants in the c-s-function as in the simulink-modell. So I have a c-header with preprocesser constants like:
#define THIS_IS_A_CONSANT 10
And there is the question:
How it is possible to include this in Simulink in this way I can use THIS_IS_A_CONSANT for example in a constant source like a workspace-variable?
Thanks and regards
Alex
There is functionality in Simulink that will allow you to include custom C header files that define constants, variables, etc.; however, as far as I know (and as one might expect) this really is only pertinent in cases where code is being generated and compiled.
So, for the most part, this particular functionality is only relevant when you are using Simulink Coder to generate a stand-alone executable from your model. For example, this link shows how to include parameters stored in an external header file during code generation through the use of Simulink.Parameter objects with Custom Storage Classes and the Code Generation - Custom Code Pane under the model's Configuration Parameters.
This link from the Simulink doc shows how to use the #define custom storage class to achieve similar results.
However, it sounds like neither of these really solve your issue, as you want to make use of the code in the header file during simulation.
That said, considering that there are elements in Simulink, such as Stateflow Charts and MATLAB Function blocks, that generate and build code "under the hood" during simulation, it's (at least hypothetically) possible that you might be able to use some of the concepts described above to access the values in your header file from one of those elements during simulation. For example, I was successfully able to access preprocessor macros in a Stateflow chart just by going to the Simulation Target->Custom Code pane under Configuration Parameters and including the text #include "header.h" under Include custom C code in generated: Header file. (In this case, header.h contained the line of C code that you included in your post)
Although it seems like you should be able to extend this functionality further, this really was the limit of what I was able to achieve as far as accessing the header file during simulation was concerned. For example, I know that running a model in Rapid Accelerator mode actually generates and builds code under the hood, so seemingly you should be able to use some combination of the techniques I described above to be able to access values from the header file during simulation. It looks like the code that Rapid Accelerator mode generates doesn't respect all of the settings defined by those techniques in the same way that Simulink/Embedded Coder do, though, so I just kept running into compilation errors. (Although maybe I'm just missing some creative combination of settings that could make that work).
Hopefully that helps explain Simulink's abilities (and limitations) regarding the inclusion of C header files. So to summarize, according to the links included above, what you are asking for is almost barely possible, but in practice... not really.
So if really all you want is to be able to create workspace variables out of the preprocessor #define's in your header file, it probably is just easiest to manually parse the file with a MATLAB script, as had previously been suggested in the comments. Here is a quick-and-dirty script that loads in a header file, iterates over each line, uses a regular expression (which you can improve upon if needed) to parse #define statements, and then calls eval to create variables from the parsed input.
filename = 'header.h';
pattern = '^\s*#define\s*(\w*)\s*(\d*\.?\d+)';
fid = fopen(filename);
tline = fgetl(fid);
while ischar(tline)
tokens = regexp(tline, pattern,'tokens','once');
if(numel(tokens) == 2)
eval([tokens{1} ' = ' tokens{2}]);
end
tline = fgetl(fid);
end
fclose(fid);
You could put this code in a callback so that it will execute every time you load your model. Just goto File->Model Properties->Model Properties, click on the Callbacks tab, and then place the code under whichever callback you desire (such as PreLoadFcn if you want it to run immediately before the model loads).

hierarchy of functions in MatLab

I have been reading someone else's matlab code and I don't know how the code structured. I mean I would like to know the hierarchy of functions, which function uses which function. I am reading the code to figure that out but its taking a lot of time.
So is there any other way I can see this hierarchy without reading the whole thing? To be honest it is starting to get confusing. Maybe MatLab has a built in function for that! I found this:
How can I generate a list of function dependencies in MATLAB?
but this doesn't seem to be helpful!
The MATLAB profiler will show you what functions are called by your code (and much more information to boot) and allow you to click through the hierarchy of function calls. You can either call profile on and then run your code, then call profile off and profile viewer, or you can simply call profile viewer and type a single line of code to run in the edit box at the top.
Use the dependency report provided in MATLAB:
http://www.mathworks.co.uk/help/matlab/matlab_prog/identify-dependencies.html
There are also some tools on the File Exchange, such as fdep.
No idea about a function to show visible or depended-upon functions. However the basic rules are:
1) Only the first function in a .m file (normally has to have the same name as the file itself) is visible outside that file.
2) Any function can see any top level (see 1.) function if the file is in the matlab path. Matlab can show you the path so you know where it's hunting.
3) The order of the path is important, the first instance of a function called foo found in the path will be called. Obviously the current directory is at the top of the path.
3) All functions in a given file can see all other functions in that file.
That's the basics. No doubt there are other rules, and possibly exceptions to this. But that understanding generally serves me well.
Obviously the easiest way to work out which function is being called is to click on it in the editor and open it.
One thing I do is simply place in each function at the beginning fprintf("inside function <name>/n"); and at the end of the function fprintf("leaving function <name>/n"); where <name> is the name of the function.
This will give you a very specific list of which function is being called by which function (based on the order that they appear). Another thing like this would be to place fprintf("function <name1> calling function <name2>/n"); so you can be more explicit about which function is being called by which one.

Accessing variable by string name

I need to load experimental data into scicoslab, a (pretty badly designed) clone fork of scilab which happens to support graphical modeling. The documentation on the web is pretty poor, but it's reasonably similar to scilab and octave.
The data I need to process is contained into a certain number of text files: Data_005, Data_010, …, Data_100. Each of them can be loaded using the -ascii flag for the loadmatfile command.
The problem comes from the fact that loadmatfile("foo", "-ascii") loads the file foo.mat into a variable named foo. In order to to cycle on the data files, I would need to do something like:
for i = [5:5:100]
name = sprintf("Data_%02d", i);
loadmatfile(name, "-ascii");
x = read_var_from_name(name);
do_something(x);
end
where what I search for is a builtin read_var_from_name which would allow me to access the internal symbol table by string.
Do you know if there exist a similar function?
Notes:
There's no way of overriding this behavior if your file is in ascii format;
In this phase I could also use octave (no graphical modelling is involved), although it behaves in the same way.
>> foo = 3.14; name = 'foo'; eval(name)
foo =
3.1400
The above works in MATLAB, and Scilab's documentation says it also has an eval function. Not sure if I understood you correctly, though.
#arne.b has a good answer.
In your case you can also do that in matlab:
a=load('filename.mat')
x=a.('variable_name')
lets go through your points one by one:
"ScicosLab, a (pretty badly designed) clone of Scilab" This in my opinion is an inaccurate way of introducing the software. ScicosLab is not a clone of Scilab but a fork of it. The team behind ScicosLab (INRIA) are the ones who made scocos (now called xcos in Scilab development line). At some point (from Scilab v4) the Scilab team decided to move away from Tcl/tk towards Java, but the SciccosLab/scicos team departed, keep using the language (Tcl) and it's graphical user interface design package (tk). Giving the ScocosLab community the credit that the whole Scilab documentation and support is not very good in general. :) (more about Scilab and the forks here)
Regarding the technical question I'm not sure what you are trying to achieve here, Scilab/ScicosLab still have the eval function which basically does what you want. However this function is to be deprecated in favor of evstr. There is also the execstr function which worth studying.
The loadmatfile, as far as I have understood, "tries" to load the variables defined in a MATLAB .mat file (MATLAB's proprietary tabular format) into the Scilab workspace. For example if there is a variable foo it will "try" to create the variable foo and loads its value from the MATLAB script. Check this example. I would create a variable x(i) = foo in the for loop. again your question is not completely clear.
As a side note maybe you could consider exporting your data as CSV instead of .mat files.