hud-python 0.4.68__tar.gz → 0.4.69__tar.gz

{hud_python-0.4.68 → hud_python-0.4.69}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: hud-python
-Version: 0.4.68
+Version: 0.4.69
 Summary: SDK for the HUD platform.
 Project-URL: Homepage, https://github.com/hud-evals/hud-python
 Project-URL: Bug Tracker, https://github.com/hud-evals/hud-python/issues
@@ -520,8 +520,8 @@ Thanks to all our contributors!
 
 ```bibtex
 @software{hud2025agentevalplatform,
-  author = {HUD and Jay Ram and Lorenss Martinsons and Parth Patel and Oskars Putans and Govind Pimpale and Mayank Singamreddy and Nguyen Nhat Minh},
-  title  = {HUD: An Evaluation Platform for Agents},
+  author = {HUD and Jay Ram and Lorenss Martinsons and Parth Patel and Govind Pimpale and Dylan Bowman and Jaideep and Nguyen Nhat Minh},
+  title  = {HUD: An Evaluation and RL Envrionments Platform for Agents},
   date   = {2025-04},
   url    = {https://github.com/hud-evals/hud-python},
   langid = {en}

{hud_python-0.4.68 → hud_python-0.4.69}/README.md

@@ -403,8 +403,8 @@ Thanks to all our contributors!
 
 ```bibtex
 @software{hud2025agentevalplatform,
-  author = {HUD and Jay Ram and Lorenss Martinsons and Parth Patel and Oskars Putans and Govind Pimpale and Mayank Singamreddy and Nguyen Nhat Minh},
-  title  = {HUD: An Evaluation Platform for Agents},
+  author = {HUD and Jay Ram and Lorenss Martinsons and Parth Patel and Govind Pimpale and Dylan Bowman and Jaideep and Nguyen Nhat Minh},
+  title  = {HUD: An Evaluation and RL Envrionments Platform for Agents},
   date   = {2025-04},
   url    = {https://github.com/hud-evals/hud-python},
   langid = {en}

{hud_python-0.4.68 → hud_python-0.4.69}/environments/browser/pyproject.toml

@@ -3,7 +3,7 @@ name = "hud-browser-controller"
 version = "0.1.0"
 description = "HUD Browser Controller - MCP interface for browser environments"
 requires-python = ">=3.11,<3.14"
-dependencies = [ "pydantic>=2.6,<3", "pydantic-settings>=2.2,<3", "hud-python>=0.4.68", "playwright", "pyautogui", "httpx", "typer", "fastapi>=0.104.1", "uvicorn[standard]>=0.24.0", "python-multipart>=0.0.6",]
+dependencies = [ "pydantic>=2.6,<3", "pydantic-settings>=2.2,<3", "hud-python>=0.4.69", "playwright", "pyautogui", "httpx", "typer", "fastapi>=0.104.1", "uvicorn[standard]>=0.24.0", "python-multipart>=0.0.6",]
 
 [build-system]
 requires = [ "hatchling",]

{hud_python-0.4.68 → hud_python-0.4.69}/environments/online_mind2web/pyproject.toml

@@ -3,7 +3,7 @@ name = "hud-om2w"
 version = "0.1.0"
 description = "HUD Remote Browser Controller with MCP tools for cloud browser providers"
 requires-python = ">=3.11,<3.13"
-dependencies = [ "hud-python>=0.4.68", "anthropic>=0.74.0", "pyautogui", "playwright", "httpx", "typer", "google-api-python-client", "google-auth",]
+dependencies = [ "hud-python>=0.4.69", "anthropic>=0.74.0", "pyautogui", "playwright", "httpx", "typer", "google-api-python-client", "google-auth",]
 
 [build-system]
 requires = [ "hatchling",]

{hud_python-0.4.68 → hud_python-0.4.69}/hud/cli/eval.py

@@ -53,13 +53,19 @@ _AGENT_PRESETS: list[AgentPreset] = [
     AgentPreset(
         "Gemini 2.5 Computer Use", AgentType.GEMINI, "gemini-2.5-computer-use-preview-10-2025"
     ),
-    # HUD Gateway presets
+    # HUD Gateway presets (models via HUD Inference API)
     AgentPreset(
-        "Grok 4.1 Fast",
+        "Grok 4-1 Fast (xAI)",
         AgentType.OPENAI_COMPATIBLE,
-        "xai/grok-4-1-fast-reasoning",
+        "grok-4-1-fast",
         {"openai_compatible": {"base_url": settings.hud_gateway_url, "model_name": "Grok"}},
     ),
+    AgentPreset(
+        "GLM-4.5V (Z-AI)",
+        AgentType.OPENAI_COMPATIBLE,
+        "glm-4.5v",
+        {"openai_compatible": {"base_url": settings.hud_gateway_url, "model_name": "GLM"}},
+    ),
 ]
 
 _DEFAULT_CONFIG_TEMPLATE = """# HUD Eval Configuration
@@ -180,7 +186,7 @@ class EvalConfig(BaseModel):
                 hud_console.error("HUD_API_KEY is required for remote execution")
                 hud_console.info("Set it: hud set HUD_API_KEY=your-key-here")
                 raise typer.Exit(1)
-            if self.agent_type in (AgentType.GEMINI, AgentType.OPERATOR):
+            if self.agent_type == AgentType.GEMINI:
                 hud_console.error(
                     f"Remote execution is not supported for {self.agent_type.value} agent"
                 )
@@ -229,6 +235,12 @@ class EvalConfig(BaseModel):
 
         if self.agent_type == AgentType.OPENAI_COMPATIBLE:
             base_url = kwargs.get("base_url", "")
+            model_name = kwargs.get("model_name", "")
+            if model_name:
+                kwargs["model_name"] = model_name
+            else:
+                kwargs["model_name"] = "OpenAI Compatible"
+
             if "api_key" not in kwargs:
                 # Use HUD API key for gateway, otherwise fall back to OpenAI API key
                 if settings.hud_gateway_url in base_url:
@@ -730,7 +742,11 @@ def eval_command(
 
     # Run
     start_time = time.time()
-    results, tasks = asyncio.run(_run_evaluation(cfg))
+    try:
+        results, tasks = asyncio.run(_run_evaluation(cfg))
+    except ValueError as e:
+        hud_console.error(str(e))
+        raise typer.Exit(1) from None
     elapsed = time.time() - start_time
 
     if cfg.remote:

{hud_python-0.4.68 → hud_python-0.4.69}/hud/datasets/utils.py

@@ -12,8 +12,10 @@ from pydantic import BaseModel, Field, field_validator, model_validator
 
 from hud.settings import settings
 from hud.types import AgentType, Task, Trace
+from hud.utils.hud_console import HUDConsole
 
 logger = logging.getLogger(__name__)
+hud_console = HUDConsole()
 
 
 class SingleTaskRequest(BaseModel):
@@ -92,6 +94,31 @@ async def submit_rollouts(
     if not settings.api_key:
         raise ValueError("HUD_API_KEY is required for remote execution")
 
+    # Validate tasks have remote-compatible mcp_config (URL-based, not command-based)
+    local_task_servers: list[tuple[int, str, str]] = []  # (task_idx, task_id, server_name)
+    affected_task_indices: set[int] = set()
+    for i, task in enumerate(tasks):
+        if task.mcp_config:
+            for server_name, server_cfg in task.mcp_config.items():
+                if (
+                    isinstance(server_cfg, dict)
+                    and "command" in server_cfg
+                    and not server_cfg.get("url")
+                ):
+                    local_task_servers.append((i, task.id or f"task_{i}", server_name))
+                    affected_task_indices.add(i)
+
+    if local_task_servers:
+        task_details = ", ".join(f"{tid} ({srv})" for _, tid, srv in local_task_servers[:3])
+        if len(local_task_servers) > 3:
+            task_details += f", ... and {len(local_task_servers) - 3} more"
+        raise ValueError(
+            f"Remote execution requires URL-based mcp_config, but "
+            f"{len(affected_task_indices)} task(s) use local Docker configs "
+            f"(command-based): {task_details}. "
+            "Convert to remote with: hud convert <tasks_file>"
+        )
+
     # Build single task requests
     requests: list[SingleTaskRequest] = []
     for task_idx, task in enumerate(tasks):
@@ -136,31 +163,34 @@ async def submit_rollouts(
                 total_accepted += result.get("accepted", 0)
                 total_rejected += result.get("rejected", 0)
 
-                logger.info(
-                    "Batch %d/%d: %d/%d accepted",
-                    (i // batch_size) + 1,
-                    (len(requests) + batch_size - 1) // batch_size,
-                    result.get("accepted", 0),
-                    len(batch),
+                for item in result.get("results", []):
+                    if isinstance(item, dict) and item.get("status") == "rejected":
+                        hud_console.warning(f"Task rejected: {item.get('error', 'Unknown reason')}")
+
+                batch_num = (i // batch_size) + 1
+                total_batches = (len(requests) + batch_size - 1) // batch_size
+                hud_console.info(
+                    f"Batch {batch_num}/{total_batches}: "
+                    f"{result.get('accepted', 0)}/{len(batch)} accepted"
                 )
 
             except httpx.HTTPStatusError as exc:
-                logger.error(
-                    "Batch submission failed: %s - %s", exc.response.status_code, exc.response.text
-                )
+                if 400 <= exc.response.status_code < 500:
+                    raise ValueError(f"Submission failed: {exc.response.text}") from exc
+                hud_console.error(f"Batch submission failed: {exc.response.status_code}")
                 total_rejected += len(batch)
 
             except Exception as exc:
-                logger.exception("Batch submission failed: %s", exc)
+                hud_console.error(f"Batch submission failed: {exc}")
                 total_rejected += len(batch)
 
     # Log final summary
-    logger.info(
-        "Submitted %d/%d requests (%d rejected)",
-        total_accepted,
-        len(requests),
-        total_rejected,
-    )
+    if total_rejected > 0:
+        hud_console.warning(
+            f"Submitted {total_accepted}/{len(requests)} requests ({total_rejected} rejected)"
+        )
+    else:
+        hud_console.info(f"Submitted {total_accepted}/{len(requests)} requests")
 
 
 async def cancel_job(job_id: str) -> dict[str, Any]:

{hud_python-0.4.68 → hud_python-0.4.69}/hud/utils/tests/test_version.py

@@ -5,4 +5,4 @@ def test_import():
     """Test that the package can be imported."""
     import hud
 
-    assert hud.__version__ == "0.4.68"
+    assert hud.__version__ == "0.4.69"

{hud_python-0.4.68 → hud_python-0.4.69}/hud/version.py

@@ -4,4 +4,4 @@ Version information for the HUD SDK.
 
 from __future__ import annotations
 
-__version__ = "0.4.68"
+__version__ = "0.4.69"

{hud_python-0.4.68 → hud_python-0.4.69}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "hud-python"
-version = "0.4.68"
+version = "0.4.69"
 description = "SDK for the HUD platform."
 readme = "README.md"
 requires-python = ">=3.11, <3.13"