From 3e3ff99ca01f12bc955f55b04fc612601c22208c Mon Sep 17 00:00:00 2001
From: Matthew Honnibal <matthew@honnibal.com>
Date: Sat, 30 Aug 2014 19:01:00 +0200
Subject: [PATCH] * Add orth features

---
 spacy/orth.py | 71 +++++++++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 71 insertions(+)
 create mode 100644 spacy/orth.py

diff --git a/spacy/orth.py b/spacy/orth.py
new file mode 100644
index 000000000..c574006c8
--- /dev/null
+++ b/spacy/orth.py
@@ -0,0 +1,71 @@
+# Binary string features
+def is_alpha(string, prob, case_stats, tag_stats):
+    return False
+
+def is_digit(string, prob, case_stats, tag_stats):
+    return False
+
+def is_punct(string, prob, case_stats, tag_stats):
+    return False
+
+def is_space(string, prob, case_stats, tag_stats):
+    return False
+
+def is_ascii(string, prob, case_stats, tag_stats):
+    return False
+
+def is_title(string, prob, case_stats, tag_stats):
+    return False
+
+def is_lower(string, prob, case_stats, tag_stats):
+    return False
+
+def is_upper(string, prob, case_stats, tag_stats):
+    return False
+
+
+# Statistics features
+def oft_case(name, thresh):
+    def wrapped(string, prob, case_stats, tag_stats):
+        return string
+    return wrapped
+
+
+def can_tag(name, thresh):
+    def wrapped(string, prob, case_stats, tag_stats):
+        return string
+    return wrapped
+
+
+# String features
+def canon_case(string, prob, cluster, case_stats, tag_stats):
+    return string
+
+def word_shape(string, *args):
+    length = len(string)
+    shape = ""
+    last = ""
+    shape_char = ""
+    seq = 0
+    for c in string:
+        if c.isalpha():
+            if c.isupper():
+                shape_char = "X"
+            else:
+                shape_char = "x"
+        elif c.isdigit():
+            shape_char = "d"
+        else:
+            shape_char = c
+        if shape_char == last:
+            seq += 1
+        else:
+            seq = 0
+            last = shape_char
+        if seq < 3:
+            shape += shape_char
+    return shape
+
+
+def non_sparse(string, prob, cluster, case_stats, tag_stats):
+    return string