The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
fix(llm): forward max_tokens to Ollama num_predict
Aniruddha Adak6502cf24bf89
diff --git a/src/llm/providers/ollama.py b/src/llm/providers/ollama.pyindex 2f83338d0..257d17900 100644--- a/src/llm/providers/ollama.py+++ b/src/llm/providers/ollama.py@@ -70,8 +70,13 @@ class OllamaProvider(LLMProvider):"messages": [msg.to_dict() for msg in input.messages],"stream": False,}+ options: dict[str, Any] = {}if input.temperature != 1.0:- payload["options"] = {"temperature": input.temperature}+ options["temperature"] = input.temperature+ if input.max_tokens is not None:+ options["num_predict"] = input.max_tokens+ if options:+ payload["options"] = optionsdata = json.dumps(payload).encode("utf-8")req = urllib.request.Request(url, data=data, headers={"Content-Type": "application/json"})diff --git a/tests/test_provider_tools.py b/tests/test_provider_tools.pyindex 4c9c76f92..ba4c09415 100644--- a/tests/test_provider_tools.py+++ b/tests/test_provider_tools.py@@ -1,3 +1,6 @@+import json+import urllib.request+from io import BytesIOfrom types import SimpleNamespaceimport pytest@@ -5,6 +8,7 @@ import pytestfrom llm.core.types import LLMInput, Message, Role, ToolDefinitionfrom llm.providers.claude import ClaudeProviderfrom llm.providers.constants import EMPTY_FILTERED_RESPONSE_ERROR+from llm.providers.ollama import OllamaProviderfrom llm.providers.openai import OpenAIProvider@@ -114,6 +118,56 @@ def test_openai_provider_allows_missing_usage():assert output.usage is None+ ("max_tokens", "temperature", "expected_options"),+ [+ (128, 1.0, {"num_predict": 128}),+ (128, 0.2, {"temperature": 0.2, "num_predict": 128}),+ (128, 0.0, {"temperature": 0.0, "num_predict": 128}),+ (0, 1.0, {"num_predict": 0}),+ (None, 1.0, {}),+ (None, 0.2, {"temperature": 0.2}),+ ],+)+def test_ollama_provider_serializes_generation_options(+ monkeypatch, max_tokens, temperature, expected_options+):+ requests = []++ def fake_urlopen(request, timeout):+ requests.append((request, timeout))+ return BytesIO(b'{"message": {"content": "ok"}, "done_reason": "stop"}')++ monkeypatch.setattr(urllib.request, "urlopen", fake_urlopen)+ provider = OllamaProvider(base_url="http://localhost:11434", default_model="llama3.2")++ output = provider.generate(+ LLMInput(+ messages=[Message(role=Role.USER, content="hi")],+ max_tokens=max_tokens,+ temperature=temperature,+ )+ )++ assert len(requests) == 1+ request, timeout = requests[0]+ expected_payload = {+ "model": "llama3.2",+ "messages": [{"role": "user", "content": "hi"}],+ "stream": False,+ }+ if expected_options:+ expected_payload["options"] = expected_options+ assert json.loads(request.data) == expected_payload+ assert request.full_url == "http://localhost:11434/api/chat"+ assert request.get_method() == "POST"+ assert request.get_header("Content-type") == "application/json"+ assert timeout == 60+ assert output.content == "ok"+ assert output.model == "llama3.2"+ assert output.stop_reason == "stop"++def test_claude_provider_serializes_tools_for_messages_api():provider = ClaudeProvider(api_key="test")client = _AnthropicClient()