From eb79049bdb8385b4f5f31ab2cef3858927e5b598 Mon Sep 17 00:00:00 2001
From: Vitor Santos Costa <vsc@dcc.fc.up.pt>
Date: Sun, 28 Oct 2012 18:22:09 +0000
Subject: [PATCH] more improvements to interface

---
 packages/pyswip/pl2py_examples/nltk.pl | 11 +++++++++
 packages/pyswip/python.c               | 32 ++++++++++++++++++++++----
 2 files changed, 39 insertions(+), 4 deletions(-)

diff --git a/packages/pyswip/pl2py_examples/nltk.pl b/packages/pyswip/pl2py_examples/nltk.pl
index 7650373af..789169e2c 100644
--- a/packages/pyswip/pl2py_examples/nltk.pl
+++ b/packages/pyswip/pl2py_examples/nltk.pl
@@ -1,3 +1,14 @@
+/*
+import nltk
+sentence = """At eight o'clock on Thursday morning
+... Arthur didn't feel very good."""
+tokens = nltk.word_tokenize(sentence)
+tagged = nltk.pos_tag(tokens)
+tagged[0:6]
+entities = nltk.chunk.ne_chunk(tagged)
+entities
+*/
+
 :- use_module(library(python)).
 :- use_module(library(maplist)).
 
diff --git a/packages/pyswip/python.c b/packages/pyswip/python.c
index 50f1ba8e2..204738ee1 100644
--- a/packages/pyswip/python.c
+++ b/packages/pyswip/python.c
@@ -20,6 +20,18 @@ static functor_t FUNCTOR_dollar1,
 
 static PyObject *py_Main;
 
+static inline int
+proper_ascii_string(const char *s)
+{
+  unsigned int c;
+
+  while ((c = *s++)) {
+    if (c > 127)
+      return FALSE;
+  }
+  return TRUE;
+}
+
 static PyObject *
 term_to_python(term_t t)
 {
@@ -31,9 +43,21 @@ term_to_python(term_t t)
     {
       char *s;
       
-      if (!PL_get_atom_chars(t, &s))
-	return NULL;
-      return PyString_FromStringAndSize(s, strlen(s) );
+      if (!PL_get_atom_chars(t, &s)) {
+	wchar_t *w;
+	atom_t at;
+	size_t len;
+
+	if (!PL_get_atom(t, &at))
+	  return NULL;
+	if (!(w = PL_atom_wchars(at, &len)))
+	  return NULL;
+	return PyUnicode_FromWideChar(w, wcslen(w) );
+      }
+      if (proper_ascii_string(s))
+	return PyString_FromStringAndSize(s, strlen(s) );
+      else
+	return PyUnicode_DecodeLatin1(s, strlen(s), NULL);
     }
   case PL_INTEGER:
     {
@@ -49,7 +73,7 @@ term_to_python(term_t t)
 
       if (!PL_get_string_chars(t, &s, &len))
 	return NULL;
-      return PyByteArray_FromStringAndSize(s, len );
+      return PyByteArray_FromStringAndSize(s, len);
     }
   case PL_FLOAT:
     {