From 4f9c123dce1f0f7f5c3cd8fe4f14484dd09f5bb7 Mon Sep 17 00:00:00 2001
From: Gertjan van Noord <g.j.m.van.noord@rug.nl>
Date: Fri, 2 Feb 2024 15:44:55 +0100
Subject: [PATCH] list features which occur in best parse

---
 Disambiguation/features_in_best_parse.py | 65 ++++++++++++++++++++++++
 1 file changed, 65 insertions(+)
 create mode 100644 Disambiguation/features_in_best_parse.py

diff --git a/Disambiguation/features_in_best_parse.py b/Disambiguation/features_in_best_parse.py
new file mode 100644
index 000000000..71c0654f1
--- /dev/null
+++ b/Disambiguation/features_in_best_parse.py
@@ -0,0 +1,65 @@
+#!/usr/bin/env python3
+
+import sys
+import errno
+
+def score(iscore):
+    (overlap,correct,system,postaglemma,postaglemma_total) = iscore.split('|')
+    n = max(int(correct),int(system))
+    pen = n-int(overlap)
+    if n:
+        return 100*(1-pen/n)
+    else:
+        return 100
+
+
+def printit(features):
+    flist = features.split('|')
+    for pair in flist:
+        try:
+            fields = pair.split('@')
+            count=fields[0]
+            feat='@'.join(fields[1:])
+            print(feat)
+        except Exception as e:
+            if e.errno == errno.EPIPE:
+                raise
+            print(pair,file=sys.stderr)
+            print(e,file=sys.stderr)
+
+
+def main():
+    prev_key=""
+    best_score=0
+
+    for line in sys.stdin:
+        try:
+            fields = line.rstrip().split('#')   #            (key,ilen,iscore,features)
+            key = fields[0]
+            ilen = fields[1]
+            iscore = fields[2]
+            features = '#'.join(fields[3:])
+            if not key == prev_key:
+                if prev_key:
+                    printit(best_features)
+                prev_key=key
+                best_features=features
+                best_score=score(iscore)
+            else:
+                next_score=score(iscore)
+                if next_score > best_score:
+                    best_features=features
+                    best_score=next_score
+        except Exception as e: 
+            if e.errno == errno.EPIPE:
+                raise
+            print("ignoring line {}".format(line[:70]),file=sys.stderr)
+            print(e,file=sys.stderr)
+
+    if key:
+        printit(best_features)
+
+
+if __name__ == "__main__":
+    main()
+