This program displays a comparison of two ENDF files. It takes into account the tabular structure of the ENDF format and is able to detect insertions and deletions of numbers. It also takes into account approximate matches for the identification of matching elements.
Note: It has not been thoroughly tested yet.
To create the executable, run
gfortran -O3 -o dtwdiff dtwdiff.f90
Then, to compare ENDF files, execute
./dtwdiff <file1> <file2> | less -R
It outputs the contents of the files side-by-side.
Piping to less -R makes the output scrollable.
The left side shows the content of the first file.
Fields only present in the first file are colored red.
The right side shows the content of the second file.
Fields only present in the second file are colored green.
Exact matches (equal strings) of corresponding fields are shown in black.
Approximate matches of numbers within a relative difference
of about 3% are colored blue.
Corresponding line numbers of the original files are also printed.
The fields are printed at the same columnar position as in the original file.
In order to show only modifications, execute
./dtwdiff <file1> <file2> | grep -C5 '^!'
An html-file visualizing the differences can be created by
./dtwdiff <file1> <file2> | ansi2html --white > differences.html
Each MF/MT section present in both files is treated in isolation. For each section, dynamic time warping (DTW) is used to identify matching elements and to align the numbers of the MF/MT sections of the ENDF files. Due to computational cost, the DTW is applied on chunks in each section if the section is too large. Too large sections exceeding hundred thousand fields in total are not analyzed.