-------- Original Message -------- Subject: general help with R Date: Fri, 20 Jan 2012 18:24:35 -0500 From: Ruth Flatscher <ruth.flatsc...@univie.ac.at> To: morphmet@morphometrics.org Dear Anna Maria, I understand there are many problems if you start using a statistics package that requires some programming logics and script writing, if you have never done such a thing before - that's how I felt in the beginning with R and I still do very often. I don't know anything really quick, but there are some small things that helped me a lot. The introductory manuals are certainly very good to understand the principle how R works - how you read stuff into R, create objects from your data and how you apply functions to these objects. I find it especially important to study the variety of ways how to select certain parts of your data matrices, because in one way or the other you'll need that all the time (in case of GM, you might for example want to leave some individuals out, or use just a subset of landmark coordinates to do certain things etc.) For this and other purposes, R project has a short reference card with a selection of important functions: http://cran.r-project.org/doc/contrib/Short-refcard.pdf. - I found it very useful to print it out and have it next to me to look up things. I also use the R documentation a lot, especially if I want to apply a new function and don't know how the synthax should be. You just call the help/documentation by typing "?" followed by the function you want to use (e.g. ?prcomp for a principal components analysis). The exact application of functions is very important: R can drive you crazy by constantly giving you error messages, until you find out that you have missed out a single damn comma or bracket, or misspelled something... so better check all that simple stuff twice before you throw away everything. You'll see after some time you'll get better at guessing where the mistake could be... Personally, I find it very helpful to use R with some script editor/graphic user interface - this is the one I'm using (http://sourceforge.net/projects/tinn-r/), but there are several out there (an overview is given here: http://www.sciviews.org/_rgui/projects/overview.html), maybe you have a look and see which one you like best. The advantage is that it will show some part of your script in colour, so that you can see which parts are R functions (in TinnR, they will be shown in blue), which terms are logicals, and which are your annotations or comments. In TinnR, you also see which brackets belong together: if you type a bracket or touch it with the cursor, it will be highlighted in red, and so will "the other half". I found this very handy in tracing those wicked little mistakes mentioned above. Hope this helps - and may you not be deterred by the strange world of R language... Best wishes, Ruth On 20 January 2012 20:01, morphmet <morphmet_modera...@morphometrics.org <mailto:morphmet_modera...@morphometrics.org>> wrote: -------- Original Message -------- Subject: Re: R-package for for shape analysis and mesh operations [under development] Date: Fri, 20 Jan 2012 11:25:46 -0500 From: Anna Maria Vettraino <vettr...@unitus.it <mailto:vettr...@unitus.it>> To: <morphmet@morphometrics.org <mailto:morphmet@morphometrics.org>> Dear all, I would take advantage from the discussion…I tried to use R and I felt really stupid…do you know a quick guide to learn how use it? Thank you in advance regards AM Il giorno 20/01/12 17:13, "morphmet" <morphmet_moderator@__morphometrics.org <mailto:morphmet_modera...@morphometrics.org>> ha scritto: -------- Original Message -------- Subject: Re: R-package for for shape analysis and mesh operations [under development] Date: Fri, 20 Jan 2012 10:38:07 -0500 From: ppi...@uniroma3.it <mailto:ppi...@uniroma3.it> To: morphmet@morphometrics.org <mailto:morphmet@morphometrics.org> CC: morphmet <morphmet@morphometrics.org <mailto:morphmet@morphometrics.org>> Dear Stefan That seems great! There is any help documentation with sample files and sample scripts? It would facilitate the use of Morpho package... Thanks in advance Paolo -------- Original Message -------- Subject: R-package for for shape analysis and mesh operations [under development] Date: Fri, 20 Jan 2012 04:33:23 -0500 From: Stefan Schlager <stefan.schlager@uniklinik-__freiburg.de <mailto:stefan.schla...@uniklinik-freiburg.de>> To: morphmet@morphometrics.org <mailto:morphmet@morphometrics.org> Dear all, as it seems that more and more people are using R and also are working with triangular meshes, I also take the opportunity to share some scripts with you: I am developing an R package (which are basically functions coded for my dissertation) called Morpho which allows a lot of mesh operations. Including: * importing all kinds of mesh formats (e.g. obj,ply,stl,...) (file2mesh) and * exporting objects of class "mesh3d" to ascii-ply format (mesh2ply) * operations like TPS deformation, and affine transformations based on landmark informations (warp.mesh, rotmesh.onto, scalemesh), including the updating of faces’ normals. * the imported objects are of class "mesh3d" so that they can be rendered by the functions shade3d and wire3d from the rgl-package * a function (called slider3d) allowing sliding of points, surfaces and curves (using the bending energy criterion) and projecting the slidden points back onto the 3D-mesh. * There is also a function for semi-automatic placement of surface patches (place.patch - yet undocumented) onto a whole sample based on a initially defined atlas configuration. * I’ve included functions for im- and exporting landmark-data from landmark-IDAV (read.pts,write.pts,read.__lmdta) and extracting the information, if the coordinate belongs to a curve or surface from imported pts files (c.extract), as I used this program to place the curves and the atlas surface. * There is also a GPA algorithm optimised for speed (procSym), useful for big sample sizes and many (semi-)landmarks, that also allows analysis of symmetry/asymmetry. * CVA (thanks to M. Rufino from this mailing list for providing nice examples in the documentation) * some permutation tests * graphical interface to find outliers (similar to MorphoJ) And here comes the bad part: I’m a bit behind with the documentation, so a lot code remains unexplained :( If you're interested, you can download precompiled packages or the source code from https://sourceforge.net/__projects/morpho-rpackage/__files/ <https://sourceforge.net/projects/morpho-rpackage/files/> It is optimised for Linux but also works on OSX and Windows (with some restrictions)!!! For the full functionality you must also download some command line tools which I coded using the header files of the meshlab/vcglib project (http://meshlab.sourceforge.__net/ <http://meshlab.sourceforge.net/> , http://vcg.sourceforge.net/__index.php/Main_Page <http://vcg.sourceforge.net/index.php/Main_Page>) and install them (add them to your system’s path). I recently managed to compile those not only for Linux but also for OSX and Windows. Read the README!! It’s all licensed under GPL, so you also can rip it apart and improve or adapt it, if necessary. Best regards Stefan Stefan Schlager M.A. Anthropologie Medizinische Fakultät der der Albert Ludwigs- Universität Freiburg Hebelstr. 29 79104 Freiburg Anthropology Faculty of Medicine, Albert-Ludwigs-University Freiburg Hebelstr. 29 D- 79104 Freiburg phone +49 (0)761 203-5522 <tel:%2B49%20%280%29761%20203-5522> fax +49 (0)761 203-6898 <tel:%2B49%20%280%29761%20203-6898> -- Paolo Piras Center for Evolutionary Ecology and Dipartimento di Scienze Geologiche, Università Roma Tre Largo San Leonardo Murialdo, 1, 00146 Roma Tel: +390657338000 <tel:%2B390657338000> email: ppi...@uniroma3.it <mailto:ppi...@uniroma3.it>
