Word Sausage for a poster
To demonstrate LM-backed normalisation
%%writefile sample.dot
digraph Sausage_21_och_10_Clean {
rankdir=LR;
fontsize=12;
labelloc="t";
label="21 och 10";
nodesep=0.35; ranksep=0.4;
node [fontname="Helvetica", shape=box, style="rounded,filled", fillcolor="#F7F9FB", fontsize=11];
edge [color=gray40, arrowsize=0.7];
// Start and End
START [shape=oval, label="start", fontsize=10, fillcolor="#EEF2F6"];
END [shape=oval, label="end", fontsize=10, fillcolor="#EEF2F6"];
// 21 variants
tjugoen [label="tjugoen"];
tjugoett [label="tjugoett"];
tjuen [label="tjuen"];
tjuett [label="tjuett"];
tjugien [label="tjugien"];
tjugiett [label="tjugiett"];
// och
och [label="och", fillcolor="#EAEFF5"];
// 10 variants
tio [label="tio"];
tie [label="tie"];
// Flow
START -> tjugoen; START -> tjugoett; START -> tjuen; START -> tjuett; START -> tjugien; START -> tjugiett;
tjugoen -> och; tjugoett -> och; tjuen -> och; tjuett -> och; tjugien -> och; tjugiett -> och;
och -> tio; och -> tie;
tio -> END; tie -> END;
// Highlight canonical path
edge [color="#2B6CB0", penwidth=2];
START -> tjugoett;
tjugoett -> och;
och -> tio;
tio -> END;
}
!dot -Tpng sample.dot -o sample.png
from IPython.display import Image
Image(filename='sample.png')