[Bugfix] Fallback to outlines for complex json schemas (#10899)

Signed-off-by: mgoin <michael@neuralmagic.com>

tests/entrypoints/conftest.py

@@ -69,6 +69,37 @@ def sample_json_schema():
     }
 
 
+@pytest.fixture
+def sample_complex_json_schema():
+    return {
+        "type": "object",
+        "properties": {
+            "score": {
+                "type": "integer",
+                "minimum": 0,
+                "maximum": 100  # Numeric range
+            },
+            "grade": {
+                "type": "string",
+                "pattern": "^[A-D]$"  # Regex pattern
+            },
+            "email": {
+                "type": "string",
+                "pattern": "^[a-zA-Z0-9._%+-]+@[a-zA-Z0-9.-]+\\.[a-zA-Z]{2,}$"
+            },
+            "tags": {
+                "type": "array",
+                "items": {
+                    "type": "string",
+                    "pattern":
+                    "^[a-z]{1,10}$"  # Combining length and pattern restrictions
+                }
+            }
+        },
+        "required": ["score", "grade", "email", "tags"]
+    }
+
+
 @pytest.fixture
 def sample_guided_choice():
     return [

tests/entrypoints/llm/test_guided_generate.py

@@ -76,6 +76,34 @@ def test_guided_json_completion(sample_json_schema, llm):
         jsonschema.validate(instance=output_json, schema=sample_json_schema)
 
 
+@pytest.mark.skip_global_cleanup
+def test_guided_complex_json_completion(sample_complex_json_schema, llm):
+    sampling_params = SamplingParams(
+        temperature=1.0,
+        max_tokens=1000,
+        guided_decoding=GuidedDecodingParams(json=sample_complex_json_schema))
+    outputs = llm.generate(prompts=[
+        f"Give an example JSON for an assignment grade "
+        f"that fits this schema: {sample_complex_json_schema}"
+    ] * 2,
+                           sampling_params=sampling_params,
+                           use_tqdm=True)
+
+    assert outputs is not None
+
+    for output in outputs:
+        assert output is not None
+        assert isinstance(output, RequestOutput)
+        prompt = output.prompt
+
+        generated_text = output.outputs[0].text
+        assert generated_text is not None
+        print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
+        output_json = json.loads(generated_text)
+        jsonschema.validate(instance=output_json,
+                            schema=sample_complex_json_schema)
+
+
 @pytest.mark.skip_global_cleanup
 def test_guided_choice_completion(sample_guided_choice, llm):
     sampling_params = SamplingParams(