Allow json schema of , resulting in unconstrained json value

outlines-dev · May 23, 2024 · c7955b4 · c7955b4
1 parent 6f655ca
commit c7955b4
Show file tree

Hide file tree

Showing 2 changed files with 31 additions and 2 deletions.
diff --git a/outlines/fsm/json_schema.py b/outlines/fsm/json_schema.py
@@ -126,7 +126,22 @@ def to_regex(
     if whitespace_pattern is None:
         whitespace_pattern = WHITESPACE
 
-    if "properties" in instance:
+    if instance == {}:
+        # JSON Schema Spec: Empty object means unconstrained, any json type is legal
+        types = [
+            {"type": "boolean"},
+            {"type": "null"},
+            {"type": "number"},
+            {"type": "integer"},
+            {"type": "string"},
+            # {"type": "array"},  # issue #913
+            {"type": "object"},
+        ]
+        regexes = [to_regex(resolver, t, whitespace_pattern) for t in types]
+        regexes = [rf"({r})" for r in regexes]
+        return rf"{'|'.join(regexes)}"
+
+    elif "properties" in instance:
         regex = ""
         regex += r"\{"
         properties = instance["properties"]
@@ -324,7 +339,7 @@ def to_regex(
                     {"type": "number"},
                     {"type": "boolean"},
                     {"type": "null"}
-                    # { "type": "array" },  # TODO: enable arrays within object-types
+                    # { "type": "array" },  # issue #913
                 ]
 
                 # We set the object depth to 2 to keep the expression finite, but the "depth"

diff --git a/tests/fsm/test_json_schema.py b/tests/fsm/test_json_schema.py
@@ -723,6 +723,20 @@ def test_format(schema, regex, examples):
                 ("1234", False),  # not an object
             ],
         ),
+        # No schema / unconstrained value
+        (
+            {},
+            [
+                ('"aaabbuecuh"', True),  # string
+                ("5.554", True),  # number
+                ("true", True),  # boolean
+                ("null", True),  # null
+                ("5999", True),  # integer
+                # ('["a", "b"]', True),  # array  # issue #913
+                ('{"key": {"k2": "value"}}', True),  # nested object
+                ("this isnt valid json", False),
+            ],
+        ),
     ],
 )
 def test_format_without_regex(schema, examples):