From e2b27b619f20097df11bc133c6d6f8ceb8578e53 Mon Sep 17 00:00:00 2001
From: Tony Garnock-Jones <tonyg@leastfixedpoint.com>
Date: Mon, 16 Aug 2021 11:38:44 -0400
Subject: [PATCH] Schema encode

---
 implementations/python/preserves/schema.py | 53 ++++++++++++++++++++--
 1 file changed, 48 insertions(+), 5 deletions(-)

diff --git a/implementations/python/preserves/schema.py b/implementations/python/preserves/schema.py
index 08576ce..a53af8c 100644
--- a/implementations/python/preserves/schema.py
+++ b/implementations/python/preserves/schema.py
@@ -176,11 +176,14 @@ class Enumeration(SchemaEntity):
     def _encode(self):
         raise TypeError('Cannot encode instance of Enumeration')
 
+def safeattrname(k):
+    return k + '_' if keyword.iskeyword(k) else k
+
 def safesetattr(o, k, v):
-    if keyword.iskeyword(k):
-        setattr(o, k + '_', v)
-    else:
-        setattr(o, k, v)
+    setattr(o, safeattrname(k), v)
+
+def safegetattr(o, k):
+    return getattr(o, safeattrname(k))
 
 class Definition(SchemaEntity):
     EMPTY = False
@@ -255,7 +258,13 @@ class Definition(SchemaEntity):
         return None
 
     def _encode(self):
-        raise NotImplementedError('Not yet implemented')
+        if self.SIMPLE:
+            if self.EMPTY:
+                return encode(self.SCHEMA, ())
+            else:
+                return encode(self.SCHEMA, self.value)
+        else:
+            return encode(self.SCHEMA, self)
 
     def _as_dict(self):
         return dict((k, getattr(self, k)) for k in self.FIELD_NAMES)
@@ -266,6 +275,35 @@ class Definition(SchemaEntity):
     def __setitem__(self, name, value):
         return safesetattr(self, name, value)
 
+def encode(p, v):
+    if p == ANY:
+        return v
+    if p.key == NAMED:
+        return encode(p[1], safegetattr(v, p[0].name))
+    if p.key == ATOM:
+        return v
+    if p.key == EMBEDDED:
+        return v ## TODO: reconsider representation of embedded values?
+    if p.key == LIT:
+        return p[0]
+    if p.key == SEQOF:
+        return tuple(encode(p[0], w) for w in v)
+    if p.key == SETOF:
+        return set(encode(p[0], w) for w in v)
+    if p.key == DICTOF:
+        return dict((encode(p[0], k), encode(p[1], w)) for (k, w) in v.items())
+    if p.key == REF:
+        return v._encode()
+    if p.key == REC:
+        return Record(encode(p[0], v), encode(p[1], v))
+    if p.key == TUPLE:
+        return tuple(encode(pp, v) for pp in p[0])
+    if p.key == TUPLE_PREFIX:
+        return tuple(encode(pp, v) for pp in p[0]) + encode(p[1], v)
+    if p.key == DICT:
+        return dict((k, encode(pp, v)) for (k, pp) in p[0].items())
+    raise ValueError('Bad schema')
+
 def module_path_str(mp):
     return '.'.join([e.name for e in mp])
 
@@ -392,6 +430,8 @@ if __name__ == '__main__':
     with open(__metaschema_filename, 'rb') as f:
         x = Decoder(f.read()).next()
     print(meta.Schema.decode(x))
+    print(meta.Schema.decode(x)._encode())
+    assert meta.Schema.decode(x)._encode() == x
 
     @extend(meta.Schema)
     def f(self, x):
@@ -407,5 +447,8 @@ if __name__ == '__main__':
         x = Decoder(f.read()).next()
     print(meta.Schema.decode(x))
     print(meta.Schema.decode(x) == meta.Schema.decode(x))
+    print(meta.Schema.decode(x)._encode())
+    assert meta.Schema.decode(x)._encode() == x
+
     print()
     print(path)