3. How to...

How to...


  • How to find out which pages are visited according to user country in Google Analytics:

    how to view visited pages by country.jpg


  • How to horizontally split a file view in sublime text editor:

    Split the view by using View -> Layout -> Rows: 2. Then select Row 1 and go to File -> New View into File. It will open second tab with the same file. Then you can drag that tab to the Row 2.


  • How to push a docker image to Docker Hub:

    • You need to have an account on Docker Hub. In this example the account is named 'ning'.
    • The image must be built so that the account name is within the image name

    Eg. for docker command:

    docker tag nextflow_docker:202004201512  ning/nextflow_docker:202004201512
    • For a docker-compose file:
    version: '3'
services:
  nextflow_docker:
    build: ./nextflow_docker
    image: ning/nextflow_docker:latest
    • Then you need to push:
    docker logout
docker login
#login with your username and password (if you are on ubuntu you can ignore password keyring (agree but dont enter a password)
#NB. If you are on ubuntu and get the error message "Error calling StartServiceByName for org.freedesktop.secrets: Timeout was reached" , try the following command "sudo apt remove gnome-keyring golang-docker-credential-helpers"
docker push ning/nextflow_docker:latest

  • How to translate a singularity recipe into a Dockerfile:

    • First install spython:
    git clone https://www.github.com/singularityhub/singularity-cli.git
cd singularity-cli
python3 setup.py install
    • Then perform the conversion:
    spython recipe /input/path/to/singularity/recipe/Singularity /output/path/to/Dockerfile

  • How to resize the left and right padding in large screens with CSS. (NB. Probably better to use bootstrap with 3 columns instead):

    • In the head section:
    <head>
  <meta name="viewport" content="width=device-width, initial-scale=1.0">

  <style>   
    @media screen and (min-width: 768px) {
        body {
            padding-left: 100px;
            padding-right: 100px;
        }
    }
  </style>

</head>

  • How to horizontally write the address on an envelope in Japan:

    〒106-0044東京都港区東麻布1-8-1 ISビル4F GPlusMedia


  • How to find the meanings of symbols in R:

    ?"@"
?"[["

  • How to overcome the following error message in Rstudio:

    WARNING: Rtools is required to build R packages but is not currently installed. Please download and install the appropriate version of Rtools before proceeding:

    In the R console:

    #Find where Rtools is installed:
Sys.which("make")

#Find where .Rprofile or .Rprofile.site is (my .Rprofile.site was in C:\R\R-3.6.2\etc):
Sys.getenv("HOME")

    Add these two lines to .Rprofile.site (or .Rprofile):

    Sys.setenv(PATH = paste("C:/rtools40/usr/bin", Sys.getenv("PATH"), sep=";"))
Sys.setenv(BINPREF = "C:/rtools40/mingw_$(WIN)/bin/")

  • How to use caret library from R to create folds for k-fold validation:

    # Import caret
library(caret)
# 10-Fold cross validation
valid_sets = createFolds(1:nrow(data), k = 10)

    How to create a vector with 10 zeroes in R:

    vector = numeric(length(1:10))

  • How to find the most common element in a vector in R:

    mostCommon <- function(x){
     un <- unique(x)
     un[which.max(tabulate(match(x,un)))]
}

  • How to normalise data in R:

    
#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
####FUNCTION THAT CAN BE USED TO SCALE TO ANY RANGE 
#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
#define the function
    #x is the input vector
    #y is the scale (eg. you can use eg 1:100 or c(1,100))
normaliza <- function(x,y){
  xin<-as.numeric(x)
  yin<-as.numeric(y)
  xmin<-min(xin)
  xmax<-max(xin)
  ymin<-min(yin)
  ymax<-max(yin)
  nza<-rep(0,length(xin))
  counta<-1
  for(i in xin){
    new<-(((i - xmin)/(xmax - xmin))*(ymax-ymin))+ymin
    nza[counta]<-new
    counta=counta+1
  }
  return(nza)
}

#define the scale
y<-c(0,10)

#an example with a vector:

#define a vector
test_vector<-1:100
#test the function on a vector
nzt<-normaliza(test_vector,y)
sort(nzt)

#an example with a dataframe

#define a dataframe
test_frame<-as.data.frame(cbind(test_vector,test_vector))
#test the function on a dataframe
nzt_frame <- as.data.frame(lapply(test_frame, normaliza,y=y))
nzt_frame


#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@
#### FUNCTION THAT CAN SCALE FROM 0 to 1  
#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

#define the function:
normalize <- function(x) {
    return ((x - min(x)) / (max(x) - min(x)))
}

#test on a data frame:
nzt_frame <- lapply(test_frame, normalize)
nzt_frame
#@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@@

  • How to apply a function with multiple parameters in R:

    mylist <- list(a=1,b=2,c=3)
myfxn <- function(var1,var2){
  var1*var2
}
var2 <- 2

sapply(mylist,myfxn,var2=var2)

  • How to manually scale data in R exactly the same as the base 'scale' function:

    #define the data
vec1<-1:20
vec2<-21:40
vecz<-as.matrix(cbind(vec1,vec2))

#define the function
scalez <- function(q,sc){
  #if not scaling then give 0 for sc
  #you could also simplify below as:
  #return(q/sd(q))   
  if(sc == 0){
    return(q/sqrt(sum(q^2)/(length(q)-1)))
  }
  else{
    #you could also simplify below as:
    #return((q-mean(q))/sd(q))
    return((q-mean(q))/(sqrt((sum((q-mean(q))^2))/(length(q)-1))))
  }
}

#test our scale function on a vector without centering
uncentered_vec1<-scalez(vec1,sc=0)
uncentered_vec1

#test our scale function on a vector with centering
centered_vec1<-scalez(vec1, sc=1)
centered_vec1

#test the base scale function on a vector without centering
scale_nocenter_vec1<-scale(vec1, center = FALSE)
scale_nocenter_vec1

#test our scale function on a matrix without centering
uncentered_vecz<-apply(vecz,2, scalez, sc=0)
uncentered_vecz

#test the base scale function on a matrix without centering
scale_nocenter_vecz<-scale(vecz, center = FALSE)
scale_nocenter_vecz

#test our scale function on a matrix with centering
centered_vecz<-apply(vecz,2, scalez, sc=1)
centered_vecz

#test the base scale function on a matrix with centering
scale_center_vecz<-scale(vecz, center = TRUE)
scale_center_vecz

  • How to do reverse scaling of scaled data in R:

    #define the data
vec1<-1:20
vec2<-21:40
vecz<-as.matrix(cbind(vec1,vec2))


#######################################################################
#define the function for unscaling

unscalez <- function(z,s,c){
  if(is.vector(z)){
    #if scaling was performed with the scale function then the scaled:scale attribute will exist
    if(length(attr(z,'scaled:scale'))>0){
      #if scaling was performed with the scale function and center option then the scaled:center attribute will exist
      if(length(attr(z, 'scaled:center'))>0){
        #if centering occurred 'scaled:scale'=sd(unscaled vector)
        return((z*(attr(z, 'scaled:scale')))+attr(z, 'scaled:center'))
      }
      else{
        #if centering did not occur 'scaled:scale'=sqrt(sum(q^2)/(length(q)-1)), where q is the unscaled vector
        return(z*attr(z, 'scaled:scale'))
      } 
    }
    else{
      if(length(c)>0){
        #if centering occurred s=sd(unscaled vector),c=mean(unscaled vector)
        return((z*s)+c)
      }
      else{
        #if centering did not occur, s=sqrt(sum(q^2)/(length(q)-1)), where q is the unscaled vector
        return(z*s)
      }
    }
  }
  else{
    columnz <- list()
    for(t in 1:ncol(z)){
      #if scaling was performed with the scale function then the scaled:scale attribute will exist
      if(length(attr(z,'scaled:scale'))>0){
        #if scaling was performed with the scale function and center option then the scaled:center attribute will exist
        if(length(attr(z, 'scaled:center'))>0){
          #if centering occurred 'scaled:scale'=sd(unscaled vector)
          columnz[[t]]<-(z[,t]*(attr(z, 'scaled:scale')[t]))+attr(z, 'scaled:center')[t]
        }
        else{
          #if centering did not occur 'scaled:scale'=sqrt(sum(q^2)/(length(q)-1)), where q is the unscaled vector
          columnz[[t]]<-(z[,t]*attr(z, 'scaled:scale')[t])
        } 
      }
      else{
        if(length(c)>0){
          #if centering occurred s=sd(unscaled vector),c=mean(unscaled vector)
          columnz[[t]]<-((z[,t]*s[t])+c[t])
        }
        else{
          #if centering did not occur, s=sqrt(sum(q^2)/(length(q)-1)), where q is the unscaled vector
          columnz[[t]]<-(z[,t]*s[t])
        }
      }          
    }
    matrixz<-do.call(cbind, columnz)
    return(matrixz)
  }
}

#######################################################################

#test the unscalez function on a vector created with the base scale function without centering
scale_nocenter_vec1<-scale(vec1, center = FALSE)
scale_nocenter_vec1
unscalez(scale_nocenter_vec1)

#test the unscalez function on a vector created with the base scale function with centering
scale_center_vec1<-scale(vec1, center = TRUE)
scale_center_vec1
unscalez(scale_center_vec1)

#test the unscalezfunction on a matrix created with the base scale function without centering
scale_nocenter_vecz<-scale(vecz, center = FALSE)
scale_nocenter_vecz
unscalez(scale_nocenter_vecz)

#test the unscalez function on a matrix created with the base scale function with centering
scale_center_vecz<-scale(vecz, center = TRUE)
scale_center_vecz
unscalez(scale_center_vecz)

#######################################################################

#test our unscalez function on a vector created with the scalez function (see previous post) without centering
uncentered_vec1<-scalez(vec1,sc=0)
uncentered_vec1
unscalez(uncentered_vec1,s=12.29035,c=NULL)

#test our unscalez function on a vector created with the scalez function (see previous post) with centering
centered_vec1<-scalez(vec1, sc=1)
centered_vec1
unscalez(centered_vec1,s=5.91608,c=10.5)

#test our unscalez function on a matrix created with the scalez function (see previous post) with centering
centered_vecz<-apply(vecz,2, scalez, sc=1)
centered_vecz
unscalez(centered_vecz,s=c(5.91608,5.916080),c=c(10.5,30.5))

#test our unscalez function on a matrix created with the scalez function (see previous post) without centering
uncentered_vecz<-apply(vecz,2, scalez, sc=0)
uncentered_vecz
unscalez(uncentered_vecz,s=c(12.29035,31.84667),c=NULL)

  • How to calculate R-squared in R:

    #where x is the fitted values and y is the actual values
rsq <- function (x, y) {cor(x, y) ^ 2}

#when given the R-squared value:

adj_rsqz<- function (r,n,p) {
  1-(((1-r)*(n-1))/(n-p-1))
}
#where:
#r=r-squared value
#n=total sample size
#p=number of predictors

  • How to get all the attributes and variables associated with an object in R:

    str(your_object)

  • How to continue a loop in R when there is an error for one of the elements:

      for (j in 1:10){
    tryCatch({ 
       print(j)
    }, error=function(e){cat("ERROR :",conditionMessage(e), "\n")})
  }

  • How to do matrix multiplication:

    matrix_multiplication_explained2.png


  • How to grab a still / screenshot of an image from a video in DaVinci Resolve:

      1. Click on the 'picture' icon at the bottom.
      1. Right click on the image that appears at the top and select 'Grab still'.
        grab_still1.jpg
      1. Right click on the image that appears in the left pane and select 'Export'.
        grab_still2.jpg
      1. Choose the picture format you wish to use and save.
        grab_still3.jpg

  • How to log in to a Facebook page when it gives the error:

    "This page isn’t working right now www.facebook.com redirected you too many times. ERR_TOO_MANY_REDIRECTS"

    1. In the browsers URL type in https://www.facebook.com/home
    2. Select the "Go to News Feed" button
    3. You should now be able to browse the site
    4. To view your page go the bottom left "Your Shortcuts" menu option and click your listed page.

    NB. You may need to clear cookies and restart the browser. I got this to work on Firefox, Chrome and Opera

Log in to reply
 
Powered by ShareZomics