Adventures & experiences in contemporary technology
My response to changes in the technology stack used in clinical research is to develop some familiarity with programming and scripting platforms that complement and compete with SAS, a system with which I have been programming since 2000. One of these has been R but Python is another that has taken up my attention and I now also have Julia in my sights as well. There may be others to assess in the fullness of time.
While I first started to explore the Data Science world in the autumn of 2017, it was in the autumn of 2019 that I began to complete LinkedIn training courses on the subject. Good though they were, I find that I need to actually use a tool in order to better understand it. At that time, I did get to hear about Python packages like Pandas, NumPy, SciPy, Scikit-learn, Matplotlib, Seaborn and Beautiful Soup though it took until of spring of this year for me to start gaining some hands-on experience with using any of these.
During the summer of 2020, I attended a BCS webinar on the CodeGrades initiative, a programming mentoring scheme inspired by the way classical musicianship is assessed. In fact, one of the main progenitors is a trained classical musician and teacher of classical music who turned to Python programming when starting a family so as to have a more stable income. The approach is that a student selects a project and works their way through it with mentoring and periodic assessments carried out in a gentle and discursive manner. Of course, the project has to be engaging for the learning experience to stay the course and that point came through in the webinar.
That is one lesson that resonates with me with subjects as diverse as web server performance and the ongoing pandemic pandemic supplying data and there are other sources of public data to examine as well before looking through my own personal archive gathered over the decades. Some subjects are uplifting while others are more foreboding but the key thing is that they sustain interest and offer opportunities for new learning. Without being able to dream up new things to try, my knowledge of R and Python would not be as extensive as it is and I hope that it will help with learning Julia too.
In the main, my own learning has been a solo effort with consultation of documentation along with web searches that have brought me to the likes of Real Python, Stack Abuse, Data Viz with Python and R and others for longer tutorials as well as threads on Stack Overflow. Usually, the web searching begins when I need a steer on a particular or a way to resolve a particular error or warning message but books always are worth reading even if that is the slower route. Those from the Dummies series or from O’Reilly have proved must useful so far but I do need to read them more completely than I already have; it is all too tempting to go with the try the “programming and search for solutions as you go” approach instead.
To get going, many choose the Anaconda distribution to get Jupyter notebook functionality but I prefer a more traditional editor so Spyder has been my tool of choice for Python programming and there are others like PyCharm as well. Spyder itself is written in Python so it can be installed using pip from PyPi like other Python packages. It has other dependencies like Pylint for code management activities but these get installed behind the scenes.
The packages that I first met in 2019 may be the mainstays for doing data science but I have discovered others since then. It also seems that there is porosity between the worlds of R an Python so you get some Python packages aping R packages and R has the Reticulate package for executing Python code. There are Python counterparts to such Tidyverse stables as dply and ggplot2 in the form of Siuba and Plotnine, respectively. The syntax of these packages are not direct copies of what is executed in R but they are close enough for there to be enough familiarity for added user friendliness compared to Pandas or Matplotlib. The interoperability does not stop there for there is SQLAlchemy for connecting to MySQL and other databases (PyMySQL is needed as well) and there also is SASPy for interacting with SAS Viya.
Pyhton may not have the speed of Julia but there are plenty of packages for working with larger workloads. Of these, Dask, Modin and RAPIDS all have there uses for dealing with data volumes that make Pandas code crawl. As if to prove that there are plenty of libraries for various forms of data analytics, data science, artificial intelligence and machine learning, there also are the likes of Keras, TensorFlow and NetworkX. These are just a selection of what is available and there is no need not to check out more. It may be tempting to stick with the most popular packages all the time, especially when they do so much, but it never hurst to keep an open mind either.
The thrust of an exhortation from a computing handbook publisher comes to mind here: don’t just look things up on Google, read a book so you really understand what you are doing. Something like those words was used to sell an eBook on Github but the same sentiment applies to R or any other computing language. Using a search engine will get you going or add to existing knowledge but only a book or a training course will help to embed real competence.
In the case of R, there is a myriad of blogs out there that can be consulted as well as function and package documentation on RDocumentation or rrdr.io. For the former, R-bloggers or R Weekly can make good places to start while ones like Stats and R, Statistics Globe, STHDA, PSI’s VIS-SIG and anything from Posit (including their main blog as well as their AI one) can be worth consulting. Additionally, there is also RStudio Education and the NHS-R Community, which also have a Github repository together with a YouTube channel. Many packages have dedicated websites as well so there is no lack of documentation with all of these so here is a selection:
To come to the real subject of this post, R is unusual in that books that you can buy also have companions websites that contain the same content with the same structure. Whatever funds this approach (and some appear to be supported by RStudio itself by the looks of things), there certainly are a lot of books available freely online in HTML as you will see from the list below while a few do not have a print counterpart as far as I know:
R Programming for Data Science
R Markdown: The Definitive Guide
bookdown: Authoring Books and Technical Documents with R Markdown
blogdown: Creating Websites with R Markdown
pagedown: Create Paged HTML Documents for Printing from R Markdown
Dynamic Documents with R and knitr
Engineering Production-Grade Shiny Apps
Outstanding User Interfaces with Shiny
Happy Git and GitHub for the useR
Outstanding User Interfaces with Shiny
Engineering Production-Grade Shiny Apps
Many of the above have counterparts published by O’Reilly or Chapman & Hall, to name the two publishers that I have found so far. Aside from sharing these with you, there is also the personal motivation of having the collection of links somewhere so I can close tabs in my Firefox session. There are other web articles open in other tabs that I need to retain and share but these will need to do for now and I hope that you find them as useful as I do.
Like Fedora, Manjaro also installs a package for VirtualBox Guest Additions when you install the Linux distro in a VirtualBox virtual machine. However, it does have certain expectations when doing this. On many systems and my own is one of these, Linux guests are forced to use the VMSVGA virtual graphics controller while Windows guests are allowed to use the VBoxSVGA one. It is the latter that Manjaro expects so you get a message like the following appearing when the desktop environment has loaded:
Windows Resizing
Set your VirtualBox Graphics Controller to enable windows resizing
After ensuring that gcc, make, perl and kernel headers are installed, I usually install VirtualBox Guest Additions myself from the included ISO image and so I did the same with Manjaro. Doing that and restarting the virtual machine got me extra functionality like screen resizing and being able to copy and paste between the VM and elsewhere after choosing the Bidirectional setting in the menus under Devices > Shared Clipboard.
That still left an unwanted message popping up on startup. To get rid of that, I just needed to remove /etc/xdg/autostart/mhwd-vmsvga-alert.desktop. It can be deleted but I just moved it somewhere else and a restart proved that the message was gone as needed. Now everything is working as I wanted.
Recently, I had someone ask me how to create PNG files in SAS using ODS Graphics so I sought out the answer for them. Normally, the suggestion would have been to create RTF or PDF files instead but there was a specific need so a different approach was taken. Adding the something like following lines before an SGPLOT, SGPANEL or SGRENDER procedure does the needful:
ods listing gpath='E:\';
ods graphics / imagename="test" imagefmt=png;
Here, the ODS LISTING statement declares the destination for the desired graphics file while the ODS GRAPHICS statement defines the file name and type. In the above example, the file test.png would be created in the root of the E drive of a Windows machine. However, this also works with Linux or UNIX directory paths.
My online travel photo gallery is a self-coded set of PHP scripts that read data from tables in a MySQL database. These tables are built from input XML files using a Perl script that itself creates and executes an SQL script. The Perl script also does some image processing using GraphicsMagick commands to resize images and to add copyright information and image framing. Because this processed one image at a time sequentially, it was taking several minutes to complete and only partly used the capacity of the PC that I used.
This led me to look at adding parallel processing and that is what brought me to looking at the Parallel::ForkManager Perl module. An alternative approach might have been to add new images in such a way as not to need the full run involving hundreds of image files, but that will take more work and I fancied having a look at parallelising things anyway.
If it was not there already, the first act would have been to install build-essential to get access to the cpan command. The following command accomplishes this:
sudo apt-get install build-essential
Once that is there, the cpan command needs to be run and some questions answered to get things going. The first question to answer is whether you want setup to be as automated as possible and the default answer of yes worked for me. The next question to answer regards the approach that cpan takes when installing modules and I chose sudo here (local::lib is the default value and manual is another option). After this, cpan drops into its own command shell. Here, I issued two more commands to continue the basic setup by updating CPAN.pm to the latest version and adding Bundle::CPAN to optimise the module further:
make install
install Bundle::CPAN
Continuing the last of these may need extra intervention to confirmation the suggested default of exit at one point in its operation and that takes a little time to complete. It is after this that Parallel::ForkManager can be installed using the following command:
install Parallel::ForkManager
That completed quickly and the cpan shell was exited using its exit command. Then, the new module was available in scripting after that. The actual use of this module is something that hope to describe in another post so I am ending this one here and the same process is just as applicable to setting up cpan and adding any other Perl CPAN module.
While I have had preview builds of Windows 10 in various virtual machines for the most of twelve months, actually upgrading physical and virtual devices that you use for more critical work is a very different matter. Also, Windows 10 is set to be a rolling release with enhancements coming on an occasional basis so I would like to see what comes before it hits the actual machines that I need to use. That means that a VirtualBox instance of the preview build is being retained to see what happens to that over time.
Some might call it incautious but I have taken the plunge and completely moved from Windows 8.1 to Windows 10. The first machine that I upgraded was more expendable and success with that encouraged me to move onto others before even including a Windows 7 machine to see how that went. The 30-day restoration period allows an added degree of comfort when doing all this. The list of machines that I upgraded were a VMware VM with 32-bit Windows 8.1 Pro (itself part of a 32-bit upgrade cascade involving Windows 7 Home and Windows 8 Pro), a VirtualBox VM with 64-bit Windows 8.1, a physical PC that dual booted Linux Mint 17.2 and 64-bit Windows 8.1 and an HP Pavilion dm4 laptop (Intel Core i3 with 8 GB RAM and a 1 TB SSHD) with Windows 7.
The main issue that I uncovered with the virtual machines is that the Windows 10 update tool that is downloaded onto Windows 7 and 8.x does not accept the graphics capability on there. This is a bug because the functionality works fine on the Windows Insider builds. The solution was to download the appropriate Windows 10 ISO image for use in the ensuing upgrade. There are 32-bit and 64-bit disk images with Windows 10 and Windows 10 Pro installation files on each. My own actions used both disk images.
During the virtual machine upgrades, most of the applications that considered important were carried over from Windows 8.1 to Windows without a bother. Anyone would expect Microsoft’s own software like Word, Excel and others to make the transition, but others like Adobe’s Photoshop and Lightroom made it too, as did Mozilla’s Firefox, albeit requiring a trip to Settings to set it as the default option for opening web pages. Less well-known desktop applications like Zinio (digital magazines) or Mapyx Quo (maps for cycling, walking and the like) were the same. Classic Shell was an exception but the Windows 10 Start Menu suffices for now anyway. Also, there was a need to reinstate Bitdefender Antivirus Plus using its new Windows 10 compatible installation file. Still, the experience was a big change from the way things used to be in the days when you used to have to reinstall nearly all your software following a Windows upgrade.
The Windows 10 update tool worked well for the Windows 8.1 PC, so no installation disks were needed. Neither was the bootloader overwritten so the Windows option needed selecting from GRUB every time there was a system reboot as part of the installation process, a temporary nuisance that was tolerated since booting into Linux Mint was preserved. Again, no critical software was lost in the process apart from Kaspersky Internet Security, which needed the Windows 10 compatible version installed, much like Bitdefender, or Epson scanning software that I found was easy to reinstall anyway. Usefully, Anquet’s Outdoor Map Navigator (again used for working with walking and cycling maps) continued to function properly after the changeover.
For the Windows 7 laptop, it was much the same story, albeit with the upgrade being delivered using Windows Update. Then, the main Windows account could be connected to my Outlook account to get everything tied up with the other machines for the first time. Before the obligatory change of background picture, the browns in the one that I was using were causing interface items to appear in red, not exactly my favourite colour for application menus and the like. Now they are in blue and all the upheaval surrounding the operating system upgrade had no effect on the Dropbox or Kaspersky installations that I had in place before it all started. If there is any irritation, it is that unpinning of application tiles from the Start Menu or turning off of live tiles is not always as instantaneous as I would have liked and that is all done now anyway.
While writing the above, I could not help thinking that more observations on Windows 10 may follow, but these will do for now. Microsoft had to get this upgrade process right and it does appear that they have, so credit is due to them for that. So far, I have Windows 10 to be stable and will be seeing how things develop from here, especially when those new features arrive occasionally as is the promise that has been made to us users. Hopefully, that will be as painless as it needs to be to ensure trust is retained.
It appears that PROC SGPLOT and other statistical graphics procedures create image files even if you are creating RTF or PDF files. By default, these are PNG files but there are other possibilities. When working with PC SAS , I have seen them written to the current working directory and that could clutter up your folder structure, especially if they are unwanted.
Being unable to track down a setting that controls this behaviour, I resolved to find a way around it by sending the files to the SAS work directory so they are removed when a SAS session is ended. One option is to set the session’s working directory to be the SAS work one and that can be done in SAS code without needing to use the user interface. As a result, you get some automation.
The method is implicit though in that you need to use an X statement to tell the operating system to change folder for you. Here is the line of code that I have used:
x "cd %sysfunc(pathname(work))";
The X statement passes commands to an operating system’s command line and they are enclosed in quotes. %sysfunc then is a macro command that allows certain data step functions or call routines as well as some SCL functions to be executed. An example of the latter is pathname and this resolves library or file references and it is interrogating the location of the SAS work library here so it can be passed to the operating systems cd (change directory) command for processing. This method works on Windows and UNIX so Linux should be covered too, offering a certain amount of automation since you don’t have to specify the location of the SAS work library in every session due to the folder name changing all the while.
Of course, if someone were to tell me of another way to declare the location of the generated PNG files that works with RTF and PDF ODS destinations, then I would be all ears. Even direct output without image file creation would be even better. Until then though, the above will do nicely.
There are times when you might need to drop back to a full screen terminal session from a graphical desktop environment when running Linux. One example that I have encountered is the installation of Nvidia’s own graphics drivers on either Ubuntu or Linux Mint. Another happened to me on Arch Linux when a Cinnamon desktop environment update left me without the ability to open up a terminal window. Then, the full screen command allowed me to add in an alternative terminal emulator and Tech Drive-in’s list came in handy for this. there might have been something to sort on a FreeBSD installation that needed the same treatment. This latter pair happened to me on sessions running within VirtualBox and that has its own needs when it comes to dropping into a full screen command line session and I will come to that later.
When running Linux on a physical PC, the keyboard shortcuts that you are CTRL + ALT + F1 for entering a full screen terminal session and CTRL + ALT + F7 for returning to the graphical desktop again. When you are running a Linux guest in VirtualBox and the host operating system also is Linux, then the aforementioned shortcuts do not work within the virtual machine but instead affect the host. To get the guest operating system to drop into a full screen terminal session, the keyboard shortcut you need is [Host Key] + F1 and the default host key is the right hand CTRL key on your keyboard unless you have assigned to something else as I have. The exit the full screen terminal session and return to the graphical desktop again, the keyboard shortcut is [Host Key] + F7.
All of this works with GNOME and Cinnamon desktop environments running in X sessions and I cannot vouch for anything else unless alternatives like Wayland come my way. Hopefully, this useful functionality applies to those other set-ups too because there are times when a terminal session is needed to recover from a mishap, rare though those thankfully are and they need to be very rare if not non-existent for most users.
On upgrading from Ubuntu GNOME 13.10 to Ubuntu GNOME 14.04, a few rough edges were to be noticed. One was the display of my chosen background image: it was garbled. Later, I discovered that there is a maximum width of 2560 px for background images in GNOME Shell these days and that things get messy beyond that.
In my case, the image width was around 6000 px and I was used to its getting resized in GNOME Shell 3.8 and its predecessors. It seems that the functionality got removed after that though so the workaround of manual image resizing in the GIMP needed to be employed. Though having big images open in memory creates an additional overhead, not handling them very well at all looks like a bug caused by setting 2560 px as a maximum screen width for the GNOME Shell panel and the complete removal of Nautilus from desktop rendering duties. Let’s hope that sense is seen with ever larger screen sizes and resolutions coming our way.
It’s the sort of thing that did get me looking at adding on Cinnamon 2.2 for a while before setting background image scaling using the indispensable GNOME Tweak Tool was discovered. LinuxG.net has a useful tutorial on this for anyone with such an adventurous streak in them. For now though, I am OK with my set up but the GNOME project’s focus on minimalism could affect us in other ways so I can see why Clem Lefebvre started the Cinnamon one primarily for Linux Mint and the desktop environment is appearing elsewhere too. After all, Gedit lost its menu bar in GNOME 3.12 so it’s just as well that we have alternative choices.
Update 2014-05-06: It seems that the desktop image bug that afflicts GNOME Shell 3.10 got sorted for GNOME Shell 3.12. At least, that is the impression that an Antergos instance in a VirtualBox virtual machine gives me.
With UNIX being the basis of Linux, I have a soft spot for trying out any UNIX that can be installed on a PC. For a while, I had OpenSolaris on the go and even vaguely recall having a look at one of the BSD’s. However, any recent attempt to install one of the latter, and there are quite a few around now, got stymied by some sort of kernel panic caused by using AMD CPU’s. With the return to the Intel fold arising from the upgrade of my main home PC last year, it perhaps was time to try again.
The recent release of FreeBSD 10.0 was the cue and I downloaded a DVD image for a test installation in a VirtualBox virtual machine with 4 GB of memory and a 32 GB virtual hard drive attached (expanding storage was chosen so not all the allocated space has been taken so far). The variant of FreeBSD chosen was the 64-bit x86 one and I set to installing it in there. Though not as pretty in appearance as those in various Linux distros, the installer was not that user unfriendly to me. Mind you, I have experience of installing Arch Linux so that might have acclimatised me somewhat.
Those installation screens ask about the keyboard mapping that you want and I successfully chose one of the UK options. There was limited opportunity for adding extras though there was a short list of few from which I made some selections. User account set up also was on offer and I would have been better off knowing what groups to assign for my personal user account so as to have to avoid needing to log in as root so often following system start up later. Otherwise, all the default options were sufficient.
When the installation process was complete, it was time to boot into the new system and all that was on offer was a command line log in session. After logging in as root, it was time to press pkg into service in order to get a desktop environment in place. The first step was to install X:
pkg install xorg
Then, it was time to install a desktop environment. While using XFCE or KDE were alternatives, I chose GNOME 2 due to familiarity and more extensive instructions on the corresponding FreeBSD handbook page. Issuing the following command added GNOME and all its helper applications:
pkg install gnome2
So that GNOME starts up at the next reboot, some extra steps are needed. The first of these is to add the following line into /etc/fstab:
proc /proc procfs rw 0 0
Then, two lines were needed in /etc/rc.conf:
gdm_enable=”YES”
gnome_enable=”YES”
The first enables the GNOME display manager and the second activates other GNOME programs that are needed for a desktop session to start. With each of these in place, I got a graphical login screen at the next boot time.
With FreeDSB being a VirtualBox Guest, it was time to consult the relevant FreeBSD manual page. Here, there are sections for a number of virtual machine tools so a search was needed to find the one for VirtualBox. VirtualBox support for FreeBSD is incomplete in that there is no installation media for BSD systems though Linux and Solaris are supported along with Windows. Therefore, it is over to the FreeBSD repositories for the required software:
pkg install virtualbox-ose-additions
Aside from the virtual machine session not capturing and releasing the mouse pointer automatically, that did everything that was needed even if it was the open source edition of the drivers and their proprietary equivalents. To resolve the mouse pointer issue, I needed to temporarily disable the GNOME desktop session in /etc/rc.conf to drop down to a console only session where xorg. conf could be generated using the following commands:
Xorg -configure
cp xorg.conf.new /etc/xorg.conf
In the new xorg.conf file, the mouse section needs to be as follows:
Section “InputDevice”
Identifier “Mouse0”
Driver “vboxmouse”
EndSection
If it doesn’t look like the above and it wasn’t the case for me then it needs changing. Also, any extra lines from the default set up also need removing or the mouse will not function as it should. The ALT+F1 (for accessing GNOME menus) and ALT+F2 (for running commands) keyboard shortcuts then become crucial when your mouse is not working as it should and could avert a panic too; knowing that adjusting a single configuration file will fix a problem when doing so is less accessible is not a good feeling as I discovered to my own cost. The graphics settings were fine by default but here’s what you should have in case it isn’t for you:
Section “Device”
### Available Driver options are:-
### Values: <i>: integer, <f>: float, <bool>: “True”/”False”,
### <string>: “String”, <freq>: “<f> Hz/kHz/MHz”
### [arg]: arg optional
Identifier “Card0”
Driver “vboxvideo”
VendorName “InnoTek Systemberatung GmbH”
BoardName “VirtualBox Graphics Adapter”
BusID “PCI:0:2:0”
EndSection
The next step is to ensure that your HAL settings are as they should. I needed to create a file in /usr/local/etc/hal/fdi/policy called 90-vboxguest.fdi that contains the following:
<?xml version=”1.0″ encoding=”utf-8″?>
<!--
# Sun VirtualBox
# Hal driver description for the vboxmouse driver
# $Id: chapter.xml,v 1.33 2012-03-17 04:53:52 eadler Exp $
Copyright (C) 2008-2009 Sun Microsystems, Inc.
This file is part of VirtualBox Open Source Edition (OSE, as
available from http://www.virtualbox.org. This file is free software;
you can redistribute it and/or modify it under the terms of the GNU
General Public License (GPL) as published by the Free Software
Foundation, in version 2 as it comes in the “COPYING” file of the
VirtualBox OSE distribution. VirtualBox OSE is distributed in the
hope that it will be useful, but WITHOUT ANY WARRANTY of any kind.
Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa
Clara, CA 95054 USA or visit http://www.sun.com if you need
additional information or have any questions.
-->
<deviceinfo version=”0.2″>
<device>
<match key=”info.subsystem” string=”pci”>
<match key=”info.product” string=”VirtualBox guest Service”>
<append key=”info.capabilities” type=”strlist”>input</append>
<append key=”info.capabilities” type=”strlist”>input.mouse</append>
<merge key=”input.x11_driver” type=”string”>vboxmouse</merge>
<merge key=”input.device” type=”string”>/dev/vboxguest</merge>
</match>
</match>
</device>
</deviceinfo>
With all that set, it is time to ensure that the custom user account is added to the wheel and operator groups using this command:
pw user mod [user name] -G wheel operator
Executing the above as root means that the custom account can run the su command so that logging in as root at the start of a desktop session no longer is needed. That is what being in the wheel group allows and the anyone in the operator group can shut down or restart the system. Both are facilities readily available in Linux so I fancied having them in FreeBSD too.
Being able to switch to root in a terminal session meant that I could go on to add software like Firefox, Libreoffice, GIMP, EMACS, Geany, Netbeans, Banshee and so on. There may be a line of opinion that FreeBSD is a server operating system but all of these make it more than passable for serving as a desktop one too. There may be no package management GUI as such and the ones that come with GNOME do not work either but anyone familiar with command line working will get around that.
FreeBSD may be conservative but that has its place too and being able to build up a system one item at a time teaches far more than getting everything already sorted in one hit. So far, there is enough documentation to get me going and I hope to see where else things go too. So far, the OS hasn’t been that intimidating and that’s good to see.