Graph Data Format Converter

Command

node app.js -f from -t to inputfile outputfile
acceptable format is as from: edge/edge2/dimacs/adj, as to:edge

example:
    shell script conv.sh usage is like following:
        conv.sh inputfile outputfile
    conv.sh is example shell script for converting from adj to edge. There are many options for node to increase
    memory availability of node.
    On intel i5 PC with 32G bytes memory equipped, 
    700Mbytes adj file can be converted to edge file in 5 minutes, using about 22G bytes memory.

Background

There are several formats to represent graph. There is no kitchensink converter supporting translation vise versa. Don't invent new one please.

if POC is proved, this program should be rewritten with more memory efficient language like C.

Text Format

1. CSR: Compressed Sparse Row

2. MATRIX: matrix of row and col. row is current, col is next: the cell has cost.

3. EDGES: list of pair of vertex id

   each lines have: from to
   

4. EDEGS2:

       firstline:  edge-count
       after: from to
       OR
       firstline:  vertex-count edge-count
       following each lines have: from to
      
       vertex id starts with 0 or 1

   As cost representation, each lines might have cost as third parameter;

5. ADJACENT: adjacent list of vertex; line number is vertex id
6. ROWCOL: colptr and rowptr
7. DIMACS: http://prolland.free.fr/works/research/dsat/dimacs.html
8. BSMR: Bitmap-based sparse matrix representation:
9. DCSR:
10. PAWAN: https://github.com/pathscale/rodinia/tree/master/cuda/bfs

       first line: node-count
       the line: from zero count upto node-count: start number-of-edges-each-node
       source-id
       length-of-edges-list
       the line: from zero count upto length-of-edges-list: edge cost

References

https://link.springer.com/article/10.1007/s41019-016-0024-y