MIGRAINE = MultIple GRaph AlIgNmEnt tool via Bron Kerbosch and VF2 algorithms
A python3-based tool that takes multiple graph objects and aligns them in a progressive way using either Bron-Kerbosch or VF2 algorithm.
- Input:
- Two or more graphs in either JSON, GRAPHML or the tool-specific GRAPH format.
- Output:
- A multiple graph matching in GRAPHML format.
git clone https://github.com/Tschichen/MIGRAINE.git
Furthermore, you can install the library as a dependency by executing:
pip3 install -e . --user
-
minimum input
python3 -m migraine -i users/path/dir_with_graphs/ -o my_name -
reads newick tree from file
python3 -m migraine -i users/path/dir_with_graphs -o my_name -n users/path/some_dir/tree.nwtree -
uses BK and predefines a clique, saves all alignment steps
python3 -m migraine -i users/path/dir_with_graphs/ -o my_name -b -c users/path/dir/clique.json -m all -
reads a list with scores and scores labels according to list, saves guide tree as newick
python3 -m migraine -i users/path/dir_with_graphs/ -o my_name -s users/path/some_dir/score_list.txt -q
Note: Clique option only available for BK. Disconnected directed graphs can only be assessed via BK algorithm. Input graphs must be all directed or undirected.
| option | explanation |
|---|---|
| --graphgen | starts built in tool, that builds (multiple) randomly created graphs. Graph properties like node count, edge probability or random labels need to be customized. |
| -h or --help | provides a list of options that can used with this tool and additional information |
| -i | Use -i and specify a path where all the graph files, you want to be aligned, are stored. All files in JSON, GRAPHML or GRAPH format in this location will be used. Mandatory Option! |
| -o | Name your current run. A directory with that name will be generated in your current location and all outputfiles will be saved there. Mandatory Option! |
| -b or --bronkerbosch | As the tool uses VF2 algorithm by default, use this parameter to switch to using Bron Kerbosch algorithm. |
| -v | random pivoting strategy for bk algorithm. Combination with pre-clique option is excluded. Default: costum pivoting. |
| -c | Clique-option: Uses the specified graph-file as anchor for Bron Kerbosch. Checks if clique is in all graphs that are to be aligned and starts multiple alignment from pre-defined clique. Input is possible in all three formats. Only works with Bron Kerbosch. |
| -s or --score <file.txt> | Allows to input a matrix of scores for node and edge labels. Check the bottom of this file for format example. Default: If no scoring list is given, the score of an alignment is generated via the size of the largest mapping. |
| -f or --forbidden | while inputting a scoring list, score via size of largest common subgraph after excluding forbidden label matches. |
| -g or --labelling | String can be "nodes", "edges" or "both". Defines what labels should be used for scoring. |
| -n or --newick | Input a guide tree for multiple alignment in newick format. If this option is not enabled the guide tree is built from scores computed via pairwise alignment. |
| -r or --randomt | saves a randomly generated guide tree for your current run |
| -t or --showt | saves the generated guide tree as .png |
| -q or --savetn | saves the generated guide tree as newick file |
| -m | saves the multiple alignment as GRAPHML file. Options for string are: "end" only end result will be saved, "all" saves every matching after one graph is added to the alignment till the end. Default: end |
| -d or --drawm | saves a visualization of the multiple graph alignment as .png |
- networkx
- sys
- os
- getopt
- glob
- math
- json
- matplotlib
- collections
- operator
- random
- pprint
- string
- Bio
- cStringIO
- time
#author; <string>
#nodes; <int>
#edges; <int>
Nodes labelled; <True/False>
Edges labelled; <True/False>
Directed Graph; <True/False>
SPACE
node count;Label(optional)
...
SPACE
node1;node2;Label(optional)
...
#author;someone
#nodes;2
#edges;1
Nodes labelled;True
Edges labelled;True
Directed Graph;False
1;Label
2;Label
1;2;Label
Please type N, if match between two labels is forbidden.
#nodelabel1 \t nodelabel2 \t nodelabel3...
score13 \t score23 \t score33
score12 \t score22
score11
#edgelabel1 \t edgelabel2 \t edgelabel3...
score13 \t score23 \t score33
score12 \t score22
score11
#C N H B
5 3 2 1
4 N 3
2 1
4
#s d i
3 3 2
7 N
2
in the example, a match between a node labeled with "N" and a node labeled with "H" would be forbidden, score between edge labels "s" and "i" would be 3.