misc

kornai · Nov 25, 2015 · 81c6435807ca3a4f841c099100a10fe2d6ae3023 · 81c6435
1 parent d575a85
commit 81c6435807ca3a4f841c099100a10fe2d6ae3023
Unified Split

Showing with 43 additions and 0 deletions.

+12 −0 conf/hu_test.cfg

+2 −0 dep_to_4lang_hu.txt

+29 −0 scripts/dep_to_dot.py
diff --git a/conf/hu_test.cfg b/conf/hu_test.cfg
@@ -0,0 +1,12 @@
+  
+
+  
+[machine]
+  
+definitions_binary = test/machines/eksz_test.pickle
+  
+graph_dir = test/graphs/hu_test
+  
+
+  
+[text]
+  
+input_sens = test/input/hu_test.sens
+  
+expand = False
+  
+
+  
+[deps]
+  
+lang = hu
+  
+dep_map = dep_to_4lang_hu.txt
diff --git a/dep_to_4lang_hu.txt b/dep_to_4lang_hu.txt
@@ -15,6 +15,8 @@ mode	0,-
 subj	1,0
 obj	2,-
+  
+pred	0,-
+  
+
 obl,.*,N...2.*	-,-	IN
 obl,.*,N...x.*	-,-	IN
 obl,.*,N...i.*	-,-	INSTRUMENT

diff --git a/scripts/dep_to_dot.py b/scripts/dep_to_dot.py
@@ -0,0 +1,29 @@
+  
+import json
+  
+import sys
+  
+
+  
+HEADER = u"digraph finite_state_machine {\n\tdpi=100;\n\trankdir=LR;\n"
+  
+EXCLUDE = ("punct")
+  
+
+  
+def dep_to_dot(deps, fn):
+  
+    edges = [
+  
+        (d['dep']['lemma'], d['type'], d['gov']['lemma']) for d in deps
+  
+        if d['type'] not in EXCLUDE]
+  
+    words = set([e[0] for e in edges] + [e[2] for e in edges])
+  
+    lines = []
+  
+    for word in words:
+  
+        lines.append(u'\t{0} [shape=rectangle, label="{0}"];'.format(word))
+  
+    for dep, dtype, gov in edges:
+  
+        lines.append(u'\t{0} -> {1} [label="{2}"];'.format(dep, gov, dtype))
+  
+    with open(fn, 'w') as f:
+  
+        f.write(HEADER.encode("utf-8"))
+  
+        f.write(u"\n".join(lines).encode("utf-8"))
+  
+        f.write("}\n")
+  
+
+  
+def main():
+  
+    data = json.load(open(sys.argv[1]))
+  
+    i = 0 if len(sys.argv) == 3 else int(sys.argv[3])
+  
+    sen = data['deps'][i]
+  
+    dep_to_dot(sen, sys.argv[2])
+  
+
+  
+if __name__ == "__main__":
+  
+    main()