make-activity

Original @JennyBC

The commit history of this repository reflects what a student might do as she works through this activity from STAT 545. This fully developed example shows:

[x] How to run an R script non-interactively
[x] How to use make

- to record which files are inputs vs. intermediates vs. outputs
- to capture how scripts and commands convert inputs to outputs
- to re-run parts of an analysis that are out-of-date  

[x] The intersection of R and make, i.e. how to

- run snippets of R code
- run an entire R script
- render an R Markdown document (or R script)

[x] The interface between RStudio and make
[x] How to use make from the shell
[x] How Git facilitates the process of building a pipeline

---

File
make
R script 1
R script 2
master rmd generator
words.txt
python script

Original output

Modified @zeeva85 for hw09

Final output

Pipeline of the automation analysis:

makefile2dot

[makefile2dot][makefile2dot] is used yo produce the proceeding image in the make pipeline The output is like this:

make output.png # output.dot is automatically removed after png is made

makefile2dot: https://github.com/vak/makefile2dot

The following calls have been added to the make pipeline:-

make

1. If the word file is avaliable in/usr/share/dict/words, it is copied from the location as words.txt, if unvaliable a .py script downloads the word list from an online source. This evaluation is done using anif else Bourne-again shell (Bash) also known as (sh) snippets script is used to make the call between download or copy.

2. An R script that contains a for loop is used to created a concatenated string vector of 26 elements with ^ followed by letters such as ^S using paste0 to be used as regex input to match the words.txt. This vectors is then matched and computed into a tibble displaying the frequency of each letter in the start position (begining) of each word in the words.txt dataset. The table is saved as freq_let.tsv.

3. Another R script is then used to generate the plots using the tsv producing the output freq_let.png (the snippets are commented out as they seem too chunky) in the makefile.

4. A different approach is taken where a master file reportgen.txt is used as a starting point using a combination of readLines and writeLines and with different lines to be read so different .rmd's could be generated, report.rmd or report2.rmd. This reduces the number of rmd in the repo making it reproducible and less clutered. report.rmd is the complete report a submission of @zeeva85 along with the original work from @jennyBC. report2.rmd is a condensed submission of purely @zeeva85 modified version. They can be accesed via analysis1/ analysis2 (see below). The rmd generates an md and html file which is kept as per the original assignment. The usage to access the individual reports:-

make analysis1 # jennyBC version 
make analysis2 # zeeva85 version
make           # complete report with @zeeva85 + @jennyBC 

clean

1. Ussage to clean

make clean_old # @jennyBC
make clean2 # @zeeva85
make clean # cleans all version, includes removal of output.png.

Non R scripts and snipperts used in make

• .py python script is used to download data
• sh snippet evaluates it to download using IF ELSE when words.txt is unavailable
• .py python script is used to make the make file workflow graph for better visualizaton