opengradient 0.5.0a2__tar.gz → 0.5.1__tar.gz

{opengradient-0.5.0a2/src/opengradient.egg-info → opengradient-0.5.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opengradient
-Version: 0.5.0a2
+Version: 0.5.1
 Summary: Python SDK for OpenGradient decentralized model management & inference services
 Author-email: OpenGradient <kyle@vannalabs.ai>
 License-Expression: MIT
@@ -23,6 +23,7 @@ Requires-Dist: requests>=2.32.3
 Requires-Dist: langchain>=0.3.7
 Requires-Dist: openai>=1.58.1
 Requires-Dist: pydantic>=2.9.2
+Requires-Dist: og-test-x402==0.0.1
 Dynamic: license-file
 
 # OpenGradient Python SDK

{opengradient-0.5.0a2 → opengradient-0.5.1}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "opengradient"
-version = "0.5.0a2"
+version = "0.5.1"
 description = "Python SDK for OpenGradient decentralized model management & inference services"
 authors = [{name = "OpenGradient", email = "kyle@vannalabs.ai"}]
 readme = "README.md"
@@ -29,6 +29,7 @@ dependencies = [
     "langchain>=0.3.7",
     "openai>=1.58.1",
     "pydantic>=2.9.2",
+    "og-test-x402==0.0.1",
 ]
 
 [project.scripts]

{opengradient-0.5.0a2 → opengradient-0.5.1}/src/opengradient/client.py

@@ -14,6 +14,9 @@ from web3 import Web3
 from web3.exceptions import ContractLogicError
 from web3.logs import DISCARD
 import urllib.parse
+import asyncio
+from x402.clients.httpx import x402HttpxClient
+from x402.clients.base import decode_x_payment_response, x402Client
 
 from .exceptions import OpenGradientError
 from .proto import infer_pb2, infer_pb2_grpc
@@ -30,7 +33,12 @@ from .types import (
     ModelRepository,
     FileUploadResult,
 )
-from .defaults import DEFAULT_IMAGE_GEN_HOST, DEFAULT_IMAGE_GEN_PORT, DEFAULT_SCHEDULER_ADDRESS, DEFAULT_LLM_SERVER_URL
+from .defaults import (
+    DEFAULT_IMAGE_GEN_HOST, 
+    DEFAULT_IMAGE_GEN_PORT, 
+    DEFAULT_SCHEDULER_ADDRESS,
+    DEFAULT_LLM_SERVER_URL, 
+    DEFAULT_OPENGRADIENT_LLM_SERVER_URL)
 from .utils import convert_array_to_model_output, convert_to_model_input, convert_to_model_output
 
 _FIREBASE_CONFIG = {
@@ -53,6 +61,9 @@ DEFAULT_RETRY_DELAY_SEC = 1
 
 PRECOMPILE_CONTRACT_ADDRESS = "0x00000000000000000000000000000000000000F4"
 
+X402_PROCESSING_HASH_HEADER = "x-processing-hash"
+X402_PLACEHOLDER_API_KEY = "0x1234567890abcdef1234567890abcdef1234567890abcdef1234567890abcdef"
+
 class Client:
     _inference_hub_contract_address: str
     _blockchain: Web3
@@ -73,6 +84,7 @@ class Client:
         email: Optional[str] = None, 
         password: Optional[str] = None, 
         llm_server_url: Optional[str] = DEFAULT_LLM_SERVER_URL,
+        og_llm_server_url: Optional[str] = DEFAULT_OPENGRADIENT_LLM_SERVER_URL,
         openai_api_key: Optional[str] = None,
         anthropic_api_key: Optional[str] = None,
         google_api_key: Optional[str] = None,
@@ -106,6 +118,7 @@ class Client:
             self._hub_user = None
 
         self._llm_server_url = llm_server_url
+        self._og_llm_server_url = og_llm_server_url
         
         self._external_api_keys = {}
         if openai_api_key or os.getenv("OPENAI_API_KEY"):
@@ -404,6 +417,15 @@ class Client:
 
         return run_with_retry(execute_transaction, max_retries)
 
+    def _og_payment_selector(self, accepts, network_filter=None, scheme_filter=None, max_value=None):
+        """Custom payment selector for OpenGradient network (og-devnet)."""
+        return x402Client.default_payment_requirements_selector(
+            accepts,
+            network_filter="og-devnet",
+            scheme_filter=scheme_filter,
+            max_value=max_value,
+        )
+
     def llm_completion(
         self,
         model_cid: str,  # Changed from LLM to str to accept any model
@@ -437,7 +459,11 @@ class Client:
             OpenGradientError: If the inference fails.
         """
         # Check if this is a local model or external
-        if not local_model and not self._is_local_model(model_cid):
+        # TODO (Kyle): separate TEE and Vanilla completion requests
+        if inference_mode == LlmInferenceMode.TEE:
+            if model_cid not in TEE_LLM:
+                return OpenGradientError("That model CID is not supported yet for TEE inference")
+
             return self._external_llm_completion(
                 model=model_cid,
                 prompt=prompt,
@@ -448,11 +474,11 @@ class Client:
         
         # Original local model logic
         def execute_transaction():
-            if inference_mode != LlmInferenceMode.VANILLA and inference_mode != LlmInferenceMode.TEE:
+            if inference_mode != LlmInferenceMode.VANILLA:
                 raise OpenGradientError("Invalid inference mode %s: Inference mode must be VANILLA or TEE" % inference_mode)
 
-            if inference_mode == LlmInferenceMode.TEE and model_cid not in [llm.value for llm in TEE_LLM]:
-                raise OpenGradientError("That model CID is not supported yet supported for TEE inference")
+            if model_cid not in [llm.value for llm in LLM]:
+                raise OpenGradientError("That model CID is not yet supported for inference")
 
             contract = self._blockchain.eth.contract(address=self._inference_hub_contract_address, abi=self._inference_abi)
 
@@ -488,7 +514,7 @@ class Client:
         temperature: float = 0.0,
     ) -> TextGenerationOutput:
         """
-        Route completion request to external LLM server.
+        Route completion request to external LLM server with x402 payments.
         
         Args:
             model: Model identifier
@@ -503,35 +529,96 @@ class Client:
         Raises:
             OpenGradientError: If request fails
         """
-        url = f"{self._llm_server_url}/v1/completions"
-        
-        headers = {"Content-Type": "application/json"}
         api_key = self._get_api_key_for_model(model)
+    
         if api_key:
-            headers["Authorization"] = f"Bearer {api_key}"
-        
-        payload = {
-            "model": model,
-            "prompt": prompt,
-            "max_tokens": max_tokens,
-            "temperature": temperature,
-        }
-        
-        if stop_sequence:
-            payload["stop"] = stop_sequence
-        
-        try:
-            response = requests.post(url, json=payload, headers=headers, timeout=60)
-            response.raise_for_status()
+            logging.debug("External LLM completions using API key")
+            url = f"{self._llm_server_url}/v1/completions"
             
-            result = response.json()
+            headers = {
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {api_key}"
+            }
             
-            return TextGenerationOutput(
-                transaction_hash="external",  # No blockchain transaction for external
-                completion_output=result["completion"]
-            )
+            payload = {
+                "model": model,
+                "prompt": prompt,
+                "max_tokens": max_tokens,
+                "temperature": temperature,
+            }
             
-        except requests.RequestException as e:
+            if stop_sequence:
+                payload["stop"] = stop_sequence
+            
+            try:
+                response = requests.post(url, json=payload, headers=headers, timeout=60)
+                response.raise_for_status()
+                
+                result = response.json()
+                
+                return TextGenerationOutput(
+                    transaction_hash="external",
+                    completion_output=result.get("completion")
+                )
+                
+            except requests.RequestException as e:
+                error_msg = f"External LLM completion failed: {str(e)}"
+                if hasattr(e, 'response') and e.response is not None:
+                    try:
+                        error_detail = e.response.json()
+                        error_msg += f" - {error_detail}"
+                    except:
+                        error_msg += f" - {e.response.text}"
+                logging.error(error_msg)
+                raise OpenGradientError(error_msg)
+
+        async def make_request():
+            async with x402HttpxClient(
+                account=self._wallet_account,
+                base_url=self._og_llm_server_url,
+                payment_requirements_selector=self._og_payment_selector,
+            ) as client:
+                headers = {
+                    "Content-Type": "application/json",
+                    "Authorization": f"Bearer {X402_PLACEHOLDER_API_KEY}"
+                }
+                
+                payload = {
+                    "model": model,
+                    "prompt": prompt,
+                    "max_tokens": max_tokens,
+                    "temperature": temperature,
+                }
+                
+                if stop_sequence:
+                    payload["stop"] = stop_sequence
+                
+                try:
+                    response = await client.post("/v1/completions", json=payload, headers=headers, timeout=60)
+                    
+                    # Read the response content
+                    content = await response.aread()
+                    result = json.loads(content.decode())
+                    payment_hash = ""
+
+                    if X402_PROCESSING_HASH_HEADER in response.headers:
+                        payment_hash = response.headers[X402_PROCESSING_HASH_HEADER]
+                    
+                    return TextGenerationOutput(
+                        transaction_hash="external",
+                        completion_output=result.get("completion"),
+                        payment_hash=payment_hash
+                    )
+                    
+                except Exception as e:
+                    error_msg = f"External LLM completion request failed: {str(e)}"
+                    logging.error(error_msg)
+                    raise OpenGradientError(error_msg)
+        
+        try:
+            # Run the async function in a sync context
+            return asyncio.run(make_request())
+        except Exception as e:
             error_msg = f"External LLM completion failed: {str(e)}"
             if hasattr(e, 'response') and e.response is not None:
                 try:
@@ -544,7 +631,7 @@ class Client:
 
     def llm_chat(
         self,
-        model_cid: str,  # Changed from LLM to str
+        model_cid: str,
         messages: List[Dict],
         inference_mode: LlmInferenceMode = LlmInferenceMode.VANILLA,
         max_tokens: int = 100,
@@ -577,7 +664,11 @@ class Client:
             OpenGradientError: If the inference fails.
         """
         # Check if this is a local model or external
-        if not local_model and not self._is_local_model(model_cid):
+        # TODO (Kyle): separate TEE and Vanilla completion requests
+        if inference_mode == LlmInferenceMode.TEE:
+            if model_cid not in TEE_LLM:
+                return OpenGradientError("That model CID is not supported yet for TEE inference")
+
             return self._external_llm_chat(
                 model=model_cid,
                 messages=messages,
@@ -590,11 +681,11 @@ class Client:
         
         # Original local model logic
         def execute_transaction():
-            if inference_mode != LlmInferenceMode.VANILLA and inference_mode != LlmInferenceMode.TEE:
+            if inference_mode != LlmInferenceMode.VANILLA:
                 raise OpenGradientError("Invalid inference mode %s: Inference mode must be VANILLA or TEE" % inference_mode)
-
-            if inference_mode == LlmInferenceMode.TEE and model_cid not in TEE_LLM:
-                raise OpenGradientError("That model CID is not supported yet supported for TEE inference")
+            
+            if model_cid not in [llm.value for llm in LLM]:
+                raise OpenGradientError("That model CID is not yet supported for inference")
 
             contract = self._blockchain.eth.contract(address=self._inference_hub_contract_address, abi=self._inference_abi)
 
@@ -663,7 +754,7 @@ class Client:
         tool_choice: Optional[str] = None,
     ) -> TextGenerationOutput:
         """
-        Route chat request to external LLM server.
+        Route chat request to external LLM server with x402 payments.
         
         Args:
             model: Model identifier
@@ -680,40 +771,110 @@ class Client:
         Raises:
             OpenGradientError: If request fails
         """
-        url = f"{self._llm_server_url}/v1/chat/completions"
-        
-        headers = {"Content-Type": "application/json"}
         api_key = self._get_api_key_for_model(model)
+    
         if api_key:
-            headers["Authorization"] = f"Bearer {api_key}"
-        
-        payload = {
-            "model": model,
-            "messages": messages,
-            "max_tokens": max_tokens,
-            "temperature": temperature,
-        }
-        
-        if stop_sequence:
-            payload["stop"] = stop_sequence
-        
-        if tools:
-            payload["tools"] = tools
-            payload["tool_choice"] = tool_choice or "auto"
-        
-        try:
-            response = requests.post(url, json=payload, headers=headers, timeout=60)
-            response.raise_for_status()
+            logging.debug("External LLM completion using API key")
+            url = f"{self._llm_server_url}/v1/chat/completions"
             
-            result = response.json()
+            headers = {
+                "Content-Type": "application/json",
+                "Authorization": f"Bearer {api_key}"
+            }
             
-            return TextGenerationOutput(
-                transaction_hash="external",  # No blockchain transaction for external
-                finish_reason=result["finish_reason"],
-                chat_output=result["message"]
-            )
+            payload = {
+                "model": model,
+                "messages": messages,
+                "max_tokens": max_tokens,
+                "temperature": temperature,
+            }
             
-        except requests.RequestException as e:
+            if stop_sequence:
+                payload["stop"] = stop_sequence
+            
+            if tools:
+                payload["tools"] = tools
+                payload["tool_choice"] = tool_choice or "auto"
+            
+            try:
+                response = requests.post(url, json=payload, headers=headers, timeout=60)
+                response.raise_for_status()
+                
+                result = response.json()
+                
+                return TextGenerationOutput(
+                    transaction_hash="external",
+                    finish_reason=result.get("finish_reason"),
+                    chat_output=result.get("message")
+                )
+                
+            except requests.RequestException as e:
+                error_msg = f"External LLM chat failed: {str(e)}"
+                if hasattr(e, 'response') and e.response is not None:
+                    try:
+                        error_detail = e.response.json()
+                        error_msg += f" - {error_detail}"
+                    except:
+                        error_msg += f" - {e.response.text}"
+                logging.error(error_msg)
+                raise OpenGradientError(error_msg)
+
+        async def make_request():
+            async with x402HttpxClient(
+                account=self._wallet_account,
+                base_url=self._og_llm_server_url,
+                payment_requirements_selector=self._og_payment_selector,
+            ) as client:
+                headers = {
+                    "Content-Type": "application/json",
+                    "Authorization": f"Bearer {X402_PLACEHOLDER_API_KEY}"
+                }
+                
+                payload = {
+                    "model": model,
+                    "messages": messages,
+                    "max_tokens": max_tokens,
+                    "temperature": temperature,
+                }
+                
+                if stop_sequence:
+                    payload["stop"] = stop_sequence
+                
+                if tools:
+                    payload["tools"] = tools
+                    payload["tool_choice"] = tool_choice or "auto"
+                
+                try:
+                    response = await client.post("/v1/chat/completions", json=payload, headers=headers, timeout=60)
+                    
+                    # Read the response content
+                    content = await response.aread()
+                    result = json.loads(content.decode())
+
+                    payment_hash = ""
+                    if X402_PROCESSING_HASH_HEADER in response.headers:
+                        payment_hash = response.headers[X402_PROCESSING_HASH_HEADER]
+                    
+                    choices = result.get("choices")
+                    if not choices:
+                        raise OpenGradientError(f"Invalid response: 'choices' missing or empty in {result}")
+                    
+                    return TextGenerationOutput(
+                        transaction_hash="external",
+                        finish_reason=choices[0].get("finish_reason"),
+                        chat_output=choices[0].get("message"),
+                        payment_hash=payment_hash
+                    )
+                    
+                except Exception as e:
+                    error_msg = f"External LLM chat request failed: {str(e)}"
+                    logging.error(error_msg)
+                    raise OpenGradientError(error_msg)
+        
+        try:
+            # Run the async function in a sync context
+            return asyncio.run(make_request())
+        except Exception as e:
             error_msg = f"External LLM chat failed: {str(e)}"
             if hasattr(e, 'response') and e.response is not None:
                 try:

{opengradient-0.5.0a2 → opengradient-0.5.1}/src/opengradient/defaults.py

@@ -8,4 +8,5 @@ DEFAULT_SCHEDULER_ADDRESS = "0x7179724De4e7FF9271FA40C0337c7f90C0508eF6"
 DEFAULT_BLOCKCHAIN_EXPLORER = "https://explorer.opengradient.ai/tx/"
 DEFAULT_IMAGE_GEN_HOST = "18.217.25.69"
 DEFAULT_IMAGE_GEN_PORT = 5125
-DEFAULT_LLM_SERVER_URL = "http://35.225.197.84:8000"
+DEFAULT_LLM_SERVER_URL = "http://35.225.197.84:8000"
+DEFAULT_OPENGRADIENT_LLM_SERVER_URL = "https://llm.opengradient.ai"

{opengradient-0.5.0a2 → opengradient-0.5.1}/src/opengradient/types.py

@@ -151,6 +151,9 @@ class TextGenerationOutput:
     completion_output: Optional[str] = None
     """Raw text output from completion-style generation. Empty string if not applicable."""
 
+    payment_hash: Optional[str] = None
+    """Payment hash for x402 transaction"""
+
 
 @dataclass
 class AbiFunction:
@@ -190,18 +193,65 @@ class Abi:
 class LLM(str, Enum):
     """Enum for available LLM models"""
 
+    # Existing open-source OG hosted models
     META_LLAMA_3_8B_INSTRUCT = "meta-llama/Meta-Llama-3-8B-Instruct"
     LLAMA_3_2_3B_INSTRUCT = "meta-llama/Llama-3.2-3B-Instruct"
     QWEN_2_5_72B_INSTRUCT = "Qwen/Qwen2.5-72B-Instruct"
     META_LLAMA_3_1_70B_INSTRUCT = "meta-llama/Llama-3.1-70B-Instruct"
     DOBBY_UNHINGED_3_1_8B = "SentientAGI/Dobby-Mini-Unhinged-Llama-3.1-8B"
     DOBBY_LEASHED_3_1_8B = "SentientAGI/Dobby-Mini-Leashed-Llama-3.1-8B"
-
+    
+    # OpenAI models via TEE
+    GPT_4_1_2025_04_14 = "gpt-4.1-2025-04-14"
+    GPT_4O = "gpt-4o"
+    O4_MINI = "o4-mini"
+    
+    # Anthropic models via TEE
+    CLAUDE_3_7_SONNET = "claude-3.7-sonnet"
+    CLAUDE_3_5_HAIKU = "claude-3.5-haiku"
+    CLAUDE_4_0_SONNET = "claude-4.0-sonnet"
+    
+    # Google models via TEE
+    GEMINI_2_5_FLASH = "gemini-2.5-flash"
+    GEMINI_2_5_PRO = "gemini-2.5-pro"
+    GEMINI_2_0_FLASH = "gemini-2.0-flash"
+    
+    # xAI Grok models via TEE
+    GROK_3_MINI_BETA = "grok-3-mini-beta"
+    GROK_3_BETA = "grok-3-beta"
+    GROK_2_1212 = "grok-2-1212"
+    GROK_2_VISION_LATEST = "grok-2-vision-latest"
+    GROK_4_1_FAST = "grok-4.1-fast"
+    GROK_4_1_FAST_NON_REASONING = "grok-4-1-fast-non-reasoning"
 
 class TEE_LLM(str, Enum):
     """Enum for LLM models available for TEE execution"""
-
-    META_LLAMA_3_1_70B_INSTRUCT = "meta-llama/Llama-3.1-70B-Instruct"
+    
+    # Existing (Currently turned off)
+    # META_LLAMA_3_1_70B_INSTRUCT = "meta-llama/Llama-3.1-70B-Instruct"
+    
+    # OpenAI models via TEE
+    GPT_4_1_2025_04_14 = "gpt-4.1-2025-04-14"
+    GPT_4O = "gpt-4o"
+    O4_MINI = "o4-mini"
+    
+    # Anthropic models via TEE
+    CLAUDE_3_7_SONNET = "claude-3.7-sonnet"
+    CLAUDE_3_5_HAIKU = "claude-3.5-haiku"
+    CLAUDE_4_0_SONNET = "claude-4.0-sonnet"
+    
+    # Google models via TEE
+    GEMINI_2_5_FLASH = "gemini-2.5-flash"
+    GEMINI_2_5_PRO = "gemini-2.5-pro"
+    GEMINI_2_0_FLASH = "gemini-2.0-flash"
+    
+    # xAI Grok models via TEE
+    GROK_3_MINI_BETA = "grok-3-mini-beta"
+    GROK_3_BETA = "grok-3-beta"
+    GROK_2_1212 = "grok-2-1212"
+    GROK_2_VISION_LATEST = "grok-2-vision-latest"
+    GROK_4_1_FAST = "grok-4.1-fast"
+    GROK_4_1_FAST_NON_REASONING = "grok-4-1-fast-non-reasoning"
 
 
 @dataclass

{opengradient-0.5.0a2 → opengradient-0.5.1/src/opengradient.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: opengradient
-Version: 0.5.0a2
+Version: 0.5.1
 Summary: Python SDK for OpenGradient decentralized model management & inference services
 Author-email: OpenGradient <kyle@vannalabs.ai>
 License-Expression: MIT
@@ -23,6 +23,7 @@ Requires-Dist: requests>=2.32.3
 Requires-Dist: langchain>=0.3.7
 Requires-Dist: openai>=1.58.1
 Requires-Dist: pydantic>=2.9.2
+Requires-Dist: og-test-x402==0.0.1
 Dynamic: license-file
 
 # OpenGradient Python SDK

{opengradient-0.5.0a2 → opengradient-0.5.1}/src/opengradient.egg-info/requires.txt

@@ -8,3 +8,4 @@ requests>=2.32.3
 langchain>=0.3.7
 openai>=1.58.1
 pydantic>=2.9.2
+og-test-x402==0.0.1