Adding a Weka shortcut under Fedora Linux

…and scaling it up for High DPI screens.

I’ve recently bought the developer edition of the XPS13 with a highDPI screen and needed to install WEKA for a course I’m doing. When I first ran it from the command line it looked like a tiny little thing in the top left of the screen however a small amount of DuckDuckGo-ing resulted in the correct way of dealing with it.

I’m using Fedora (and gnome) so create a Desktop file for Weka with this pasth and filename: ~/.local/share/applications/weka.desktop

I’ve put my weka jar file in a directory called Programs in my home folder.

[Desktop Entry]
Type=Application
Encoding=UTF-8
Name=Weka
Comment=Weka Application
Exec=java -Dsun.java2d.uiScale=3.5 -jar /home/mike/Programs/weka-3-8-1/weka.jar
Icon=/home/mike/Programs/weka-3-8-1/weka.ico
Terminal=false

Save it and you should be good to go. Make sure you have the Latest Java installed (I’m using Java9 straight from Oracle). Oh! change the uiScale factor to suite your preference.

Generating a SSH key and uploading it to a server

Generate a key with:

ssh-keygen -t rsa

When you are entering the values for the key, remember to set the name and path appropriately.
Upload the public part to your server using:

ssh-copy-id -i ~/.ssh/root_key.pub root@1.2.3.4

Whatever you set the name and path to in the ssh-keygen part you will need to make sure it replaces the ~/.ssh/root_key bit in the second command.

Now you can use rsync to seamlessly copy files from one computer to another

rsync -avz --progress /home/person1/DirectoryOrFile person2@1.2.3.4:/home/person2/DirectoryOrFile

…so that’s:

rsync -avz --progress  username@server:

GitLab Docker Container on Fedora 26 Server

I’ve set up a NUC in the house to replace a Digital Ocean droplet.
My first job was to set up a headless Fedora 26 server. Currently it’s running Nextcloud which I set up using instructions from: marksei.com, and a docker container with Gitlab on it, based on instructions from this site.

To get the docker container to run I sued the following code:

docker run --detach --hostname gitlab.my.domain --publish 1443:443 --publish 180:80 --publish 122:22     --name gitlab     --restart always     --volume /srv/gitlab/config:/etc/gitlab:Z     --volume /srv/gitlab/logs:/var/log/gitlab:Z     --volume /srv/gitlab/data:/var/opt/gitlab:Z    gitlab/gitlab-ce:latest

Notice that the ports are translated to:
1443 from 443 (https)
180 from 80 (http)
122 from 22 (ssh)

…because I was already running https, http and ssh on the server.
mark the repo as the origin by going into the directory and using the following command for SSH

git remote add origin ssh://git@yourIPADDRESSorURL:122/monkeymike/MyGreatCodingProject.git

..or for http connections (using username and password – when pushing)

git remote add origin http://yourIPADDRESSorURL:180/monkeymike/MyGreatCodingProject.git

then push to it with:

git push -u origin --all

…and if you need to clear this:

git remote rm origin

If I want to connect to the docker container over SSH I use:

ssh git@myserverURL -p 122

..and when I connect to the server, front end, I use http://myserverURL:180 and it’s all good.

NuGet on Rider C# IDE from JetBrains

You can find NuGet under the Tools menu, then select NuGet and then Manage Nuget Packages.

Once you have the packages dialog open type in the name of the package, choose the package you want and use the green cross button (on the right hand side) to add it to your project.
That should add it to your project; nice and easy.

Wales Pharmacy Data. Load unto mariadb.

You can download some data from here.
Unzip the files into your data directory. The GPDataYYYYMM.csv files are the ones you want.
This assumes you have mariadb installed and all working OK.

Unfortunately the structure of the csv files changes between 2013 to 2015 has one set of columns and 2016 onward have another set.
as per Blue Peter I’ve created a database called pharmacy before hand.

create table pharmacy.GPData_OLD
(
	HB text null,
	Locality text null,
	PracticeID text null,
	BNFCode text null,
	BNFName text null,
	Items int null,
	NIC double null,
	ActCost double null,
	Quantity int null,
	Period int null
);

The new structure has a couple of extra columns: DDD and ADQ.

create table pharmacy.GPData_NEW
(
	HB text null,
	Locality text null,
	PracticeID text null,
	BNFCode text null,
	BNFName text null,
	Items int null,
	NIC double null,
	ActCost double null,
	Quantity int null,
	DDD double null,
	ADQ double null,
	Period int null
);

Now we’ve created the tables we can use a bash script to load all the csv files into the tables we’ve just knocked up.
This from something pulled from Stack Overflow:

for f in /DATA/mysql/pharmacy/GPData201[345]*.csv
do
    mysql -e "LOAD DATA INFILE '"$f"' INTO TABLE  GPData_OLD FIELDS TERMINATED BY ',' LINES TERMINATED BY '\n' IGNORE 1 LINES" -u root --password="yourrootpassword" pharmacy
echo "Done: '"$f"' at $(date)"
done

save it as something called “something” and then run it by invoking like ./something. Remember chmod u+x something

for f in /DATA/mysql/pharmacy/GPData201[67]*.csv
do
    mysql -e "LOAD DATA INFILE '"$f"' INTO TABLE  GPData201706 FIELDS TERMINATED BY ',' LINES TERMINATED BY '\n' IGNORE 1 LINES" -u root --password="TreeOfLife123$" pharmacy
echo "Done: '"$f"' at $(date)"
done

as something like “something2016” invoking with ./something2016 again make it run-able using chmod u+x something2016

Chug chug chug and you get something like 33 million rows.

PostgreSQL access over the local network.

After installing postgresql

Edit the pg_hba.conf
usually

/var/lib/pgsql/data/

under the line

# TYPE DATABASE USER ADDRESS METHOD

comment out all the lines using #
and then add

host all all 127.0.0.1/32 trust

and

host all all 192.168.0.0/24 trust

and that seems to work. I can log in to the server over my local network using DataGrip.

Remove lots of values from a large Excel without the undo memory overhead

in this case get rid of the null, this was a sparse query output!

Sub RemoveNulls_memoryCheap()
'
' RemoveNulls_memoryCheap Macro
'
For cci = 1 To 255
If Not Cells(1, cci) = "" Then
    columnCOunt = cci
End If
Next cci
    
For t = 1 To columnCOunt
Range(Columns(t), Columns(t)).EntireColumn.Select
    Selection.Replace What:="null", Replacement:="", LookAt:=xlPart, _
        SearchOrder:=xlByRows, MatchCase:=False, SearchFormat:=False, _
        ReplaceFormat:=False
Next t

some grep

Find all instances of java in the output list from the rpm installed packages list.

rpm -qa | grep "java"