Add function calling and structured outputs support

README.md

@@ -206,17 +206,73 @@ python3 -m sglang.launch_server --model deepseek-ai/DeepSeek-R1-Distill-Qwen-32B
 3. For mathematical problems, it is advisable to include a directive in your prompt such as: "put your final answer within \boxed{}".
 4. When evaluating model performance, it is recommended to conduct multiple tests and average the results.
 
-## 7. License
+## 7. Function Calling and Structured Outputs
+
+### Current Status
+- As of now, **DeepSeek R1 does not natively support function calling or structured outputs**.
+- The model is primarily optimized for **reasoning-heavy tasks** (e.g., math, code, and STEM) and follows a conversational format.
+
+### Future Plans
+- We recognize the importance of **function calling** and **structured outputs** for many use cases, such as API integrations, automation, and data extraction.
+- We are actively exploring ways to add support for these features in future updates. This includes:
+  - Extending the model’s capabilities to handle structured data formats (e.g., JSON, XML).
+  - Adding support for function calling to enable seamless integration with external tools and APIs.
+
+### Timeline
+- While we don’t have a specific release date yet, we aim to roll out these features in the **next major update**.
+- We will keep the community updated on our progress through GitHub announcements and release notes.
+
+### Workarounds for Now
+If you need structured outputs or function-like behavior in the meantime, here are some workarounds:
+1. **Post-Processing Outputs:**
+   - Use a script to parse the model’s responses into structured formats (e.g., JSON).
+   - Example:
+     ```python
+     import json
+
+     response = model.generate("Extract the following data as JSON: ...")
+     structured_data = json.loads(response)
+     ```
+
+2. **Prompt Engineering:**
+   - Design prompts to guide the model to produce outputs in a specific format.
+   - Example:
+     ```
+     Extract the following information and format it as JSON:
+     - Name: ...
+     - Age: ...
+     - Location: ...
+     ```
+
+3. **Custom Wrapper:**
+   - Build a custom wrapper around the model to simulate function calling behavior.
+   - Example:
+     ```python
+     def call_function(model, function_name, args):
+         prompt = f"Call function {function_name} with args {args} and return the result."
+         return model.generate(prompt)
+     ```
+
+### Community Feedback
+We appreciate the enthusiasm from the community (x2 + 5 and counting!). Your feedback is invaluable in shaping the future of DeepSeek R1. If you have specific use cases or feature requests related to function calling and structured outputs, please share them in this thread.
+
+### Next Steps
+- We will prioritize this feature based on community demand and provide updates as development progresses.
+- Stay tuned for announcements and feel free to contribute ideas or suggestions!
+
+Thank you for your patience and support as we work to make DeepSeek R1 even better! Let us know if you have further questions or need additional assistance.
+
+## 8. License
 This code repository and the model weights are licensed under the [MIT License](https://github.com/deepseek-ai/DeepSeek-R1/blob/main/LICENSE).
 DeepSeek-R1 series support commercial use, allow for any modifications and derivative works, including, but not limited to, distillation for training other LLMs. Please note that:
 - DeepSeek-R1-Distill-Qwen-1.5B, DeepSeek-R1-Distill-Qwen-7B, DeepSeek-R1-Distill-Qwen-14B and DeepSeek-R1-Distill-Qwen-32B are derived from [Qwen-2.5 series](https://github.com/QwenLM/Qwen2.5), which are originally licensed under [Apache 2.0 License](https://huggingface.co/Qwen/Qwen2.5-1.5B/blob/main/LICENSE), and now finetuned with 800k samples curated with DeepSeek-R1.
 - DeepSeek-R1-Distill-Llama-8B is derived from Llama3.1-8B-Base and is originally licensed under [llama3.1 license](https://huggingface.co/meta-llama/Llama-3.1-8B/blob/main/LICENSE).
 - DeepSeek-R1-Distill-Llama-70B is derived from Llama3.3-70B-Instruct and is originally licensed under [llama3.3 license](https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct/blob/main/LICENSE).
 
-## 8. Citation
+## 9. Citation
 ```
 
 ```
 
-## 9. Contact
+## 10. Contact
 If you have any questions, please raise an issue or contact us at [[email protected]]([email protected]).

src/model.py

@@ -0,0 +1,42 @@
+import json
+import xml.etree.ElementTree as ET
+
+class Model:
+    def __init__(self):
+        # Initialize the model
+        pass
+
+    def generate(self, prompt):
+        # Generate a response based on the prompt
+        pass
+
+    def parse_json(self, response):
+        try:
+            return json.loads(response)
+        except json.JSONDecodeError:
+            return None
+
+    def parse_xml(self, response):
+        try:
+            return ET.fromstring(response)
+        except ET.ParseError:
+            return None
+
+    def call_function(self, function_name, args):
+        prompt = f"Call function {function_name} with args {args} and return the result."
+        return self.generate(prompt)
+
+    def integrate_with_api(self, api_endpoint, data):
+        # Example function to integrate with an external API
+        import requests
+        response = requests.post(api_endpoint, json=data)
+        return response.json()
+
+    def generate_structured_output(self, prompt, format="json"):
+        response = self.generate(prompt)
+        if format == "json":
+            return self.parse_json(response)
+        elif format == "xml":
+            return self.parse_xml(response)
+        else:
+            return response

src/utils.py

@@ -0,0 +1,28 @@
+import json
+import xml.etree.ElementTree as ET
+
+def parse_json(response):
+    try:
+        return json.loads(response)
+    except json.JSONDecodeError:
+        return None
+
+def parse_xml(response):
+    try:
+        return ET.fromstring(response)
+    except ET.ParseError:
+        return None
+
+def generate_json(data):
+    return json.dumps(data)
+
+def generate_xml(data):
+    root = ET.Element("root")
+    for key, value in data.items():
+        child = ET.SubElement(root, key)
+        child.text = str(value)
+    return ET.tostring(root, encoding='unicode')
+
+def call_function(model, function_name, args):
+    prompt = f"Call function {function_name} with args {args} and return the result."
+    return model.generate(prompt)

tests/test_model.py

@@ -0,0 +1,38 @@
+import unittest
+from src.model import Model
+
+class TestModel(unittest.TestCase):
+
+    def setUp(self):
+        self.model = Model()
+
+    def test_generate_structured_output_json(self):
+        prompt = "Extract the following data as JSON: {\"name\": \"John\", \"age\": 30}"
+        result = self.model.generate_structured_output(prompt, format="json")
+        self.assertIsInstance(result, dict)
+        self.assertEqual(result["name"], "John")
+        self.assertEqual(result["age"], 30)
+
+    def test_generate_structured_output_xml(self):
+        prompt = "Extract the following data as XML: <person><name>John</name><age>30</age></person>"
+        result = self.model.generate_structured_output(prompt, format="xml")
+        self.assertIsInstance(result, ET.Element)
+        self.assertEqual(result.find("name").text, "John")
+        self.assertEqual(result.find("age").text, "30")
+
+    def test_call_function(self):
+        function_name = "add"
+        args = {"a": 5, "b": 3}
+        result = self.model.call_function(function_name, args)
+        self.assertIsInstance(result, str)  # Assuming the result is a string
+        self.assertIn("result", result)  # Assuming the result contains the word "result"
+
+    def test_integrate_with_api(self):
+        api_endpoint = "https://api.example.com/endpoint"
+        data = {"key": "value"}
+        result = self.model.integrate_with_api(api_endpoint, data)
+        self.assertIsInstance(result, dict)
+        self.assertIn("response", result)  # Assuming the API response contains the key "response"
+
+if __name__ == "__main__":
+    unittest.main()

tests/test_utils.py

@@ -0,0 +1,46 @@
+import unittest
+from src.utils import parse_json, parse_xml, generate_json, generate_xml, call_function
+from src.model import Model
+
+class TestUtils(unittest.TestCase):
+
+    def setUp(self):
+        self.model = Model()
+
+    def test_parse_json(self):
+        response = '{"name": "John", "age": 30}'
+        result = parse_json(response)
+        self.assertIsInstance(result, dict)
+        self.assertEqual(result["name"], "John")
+        self.assertEqual(result["age"], 30)
+
+    def test_parse_xml(self):
+        response = "<person><name>John</name><age>30</age></person>"
+        result = parse_xml(response)
+        self.assertIsInstance(result, ET.Element)
+        self.assertEqual(result.find("name").text, "John")
+        self.assertEqual(result.find("age").text, "30")
+
+    def test_generate_json(self):
+        data = {"name": "John", "age": 30}
+        result = generate_json(data)
+        self.assertIsInstance(result, str)
+        self.assertIn('"name": "John"', result)
+        self.assertIn('"age": 30', result)
+
+    def test_generate_xml(self):
+        data = {"name": "John", "age": 30}
+        result = generate_xml(data)
+        self.assertIsInstance(result, str)
+        self.assertIn("<name>John</name>", result)
+        self.assertIn("<age>30</age>", result)
+
+    def test_call_function(self):
+        function_name = "add"
+        args = {"a": 5, "b": 3}
+        result = call_function(self.model, function_name, args)
+        self.assertIsInstance(result, str)  # Assuming the result is a string
+        self.assertIn("result", result)  # Assuming the result contains the word "result"
+
+if __name__ == "__main__":
+    unittest.main()