cosine similarity parallel

This software computes the cosine similarity between all columns versus all columns of tables in CSV (separator TAB). Computations are performed in parallel by groups of n columns (n=100 by default).

┌────────────────────────────────────────────┐

│ Cosine Similarity - (c)Frederic PONT 2019 │

│ Free Software GNU General Public License │

└────────────────────────────────────────────┘

precompilation...

loading CSV...done !

loading DataFrames...done !

Missing values not allowed !

Please enter the number of columns / thread (def = 100) :

Read test.csv...done

nb columns = 6

Write results/test.csv...done

elapsed time: 4.476922988891602 seconds

Caution : missing values and NA are not allowed

Manual :

1- install Julia v1.2 or later, start Julia and install packages by pressing the "]" key:

LinearAlgebra : add LinearAlgebra

CSV : add CSV

2- unzip the software

3- copy your table(s) CSV (separator TAB) in the data directory

the first line of the table must contains the columns names, and the first columns must contains row names.

4- execute the software by the command : julia main_001.jl

5- enter the number of the columns / thread

7- results are in the results directory.

8- to obtain a table use julia square_table_12.jl or non_square_table_12.jl (square table). The final table is in the table directory