lang_EN.to_year(): render correct pronunciations (#144)

* Add tests

Starting work on #141, which points out that 2017 should come out
"twenty seventeen"; it currently comes out "two thousand seventeen".

* Handle English years properly.

Resolves #141.

* Whoops -- didn't PEP-8 my tests

* Handle years MORE properly...with suffixes!

- Rendering years between 10-99 was bugged (e.g. "zero fifty").
Fixed that by thinking outside the high/low text box I had made.
- Added suffixes: default none if positive, 'BC' if negative.
User-overrideable.

num2words/lang_EN.py

@@ -82,8 +82,24 @@ class Num2Word_EN(lang_EU.Num2Word_EU):
         self.verify_ordinal(value)
         return "%s%s" % (value, self.to_ordinal(value)[-2:])
 
-    def to_year(self, val, longval=True):
-        if not (val // 100) % 10:
-            return self.to_cardinal(val)
-        return self.to_splitnum(val, hightxt="hundred", jointxt="and",
-                                longval=longval)
+    def to_year(self, val, suffix=None, longval=True):
+        if val < 0:
+            val = abs(val)
+            suffix = 'BC' if not suffix else suffix
+        high, low = (val // 100, val % 100)
+        # If year is 00XX, X00X, or beyond 9999, go cardinal.
+        if (high == 0
+                or (high % 10 == 0 and low < 10)
+                or high >= 100):
+            valtext = self.to_cardinal(val)
+        else:
+            hightext = self.to_cardinal(high)
+            if low == 0:
+                lowtext = "hundred"
+            elif low < 10:
+                lowtext = "oh-%s" % self.to_cardinal(low)
+            else:
+                lowtext = self.to_cardinal(low)
+            valtext = "%s %s" % (hightext, lowtext)
+        return (valtext if not suffix
+                else "%s %s" % (valtext, suffix))