malomarrec

// Setup: pip3 install pandas matplotlib wordcloud 

import pandas as pd
import matplotlib.pyplot as plt
from wordcloud import WordCloud
import re


def replace_substring(test_str, s1, s2):
    # Replacing all occurrences of substring s1 with s2

malomarrec

timeline
        title Roadmap
        section Q2 FY2024
          New features:
            Autocomplete :
            Plugins :
            Guardrails / Filters MVP :
            Cody in Sourcegraph UI :
            And more!
          Context fetching:

malomarrec

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "ec2:*",
                "eks:*",
                "autoscaling:*"
            ],

malomarrec

h3 {
  text-align: center;
}

malomarrec

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "ec2:*",
                "eks:*",
                "autoscaling:*"
            ],

malomarrec

# Notes: 
# You need to have the clusterone package installed (pip install tensorport)
# Export logs and outputs to /logs, your data is in /data. 

import tensorflow as tf
from clusterone import get_data_path, get_logs_path


# Get the environment parameters for distributed TensorFlow
try:

malomarrec

#This function "slides" dataframe blocks based on NA values
#Takes a dataframe and two column names
#On each row, one of these columns have to be NA
#The function returns a dataframe with a new column called <newname> with the combined non NA values of the first columns
#Different factor levels in both columns are dealt with by casting to character, then back to factor after the merge
#If newname is not specified, the first columns is overriden by the new column
#Optionnal parameters allows to drop the old columns

slide_df <- function(df,name1,name2,newname = NULL, drop = FALSE){
  require(dplyr)

malomarrec

read.spss_wrapper <- function (path, verbose = FALSE){
  #In R 3.4.0, reading SPSS with foreign::read.spss raises errors when converting int factor levels to their labels with use.value.labels=TRUE
  #This is a workaround.
  
  list_data <- read.spss(path, to.data.frame=FALSE, use.value.labels=FALSE)
  #list_data <- read.sav(path, to.data.frame=FALSE, use.value.labels=TRUE)
  l1 = length(list_data)
  indx <- sapply(list_data, is.factor)
  list_data[indx] <- lapply(list_data[indx], function(x) {
    levels(x) <- make.unique(levels(x))

malomarrec

benjamini_hochberg <- function(df, fdr){
  #input:
    #df = dataframe where the rows are the tests and one of the column named "pval"" contains the p-values
    #fdr = false discovery rate threshold
  # output
    # "threshold_BH" is the benjamini_hochberg threshold
    # "reject" indicates whether we reject the null or not in the output
  n = nrow(df)
  bh = df %>% mutate(rank = dense_rank(pval))  %>% mutate(threshold_BH = (rank/n)*fdr, reject = pval < threshold_BH) %>% arrange(pval)
  return(bh)