wehub-resource-sync
52 lines
1.8 KiB
Python
52 lines
1.8 KiB
Python
import llm
|
|
from llm.tools import llm_version
|
|
import os
|
|
import pytest
|
|
|
|
API_KEY = os.environ.get("PYTEST_OPENAI_API_KEY", None) or "badkey"
|
|
|
|
|
|
# This response contains streaming variant "a" where arguments="" is followed by arguments="{}"
|
|
@pytest.mark.vcr(record_mode="none")
|
|
def test_tools_streaming_variant_a():
|
|
model = llm.get_model("gpt-4.1-mini")
|
|
chain = model.chain(
|
|
"What is the current llm version?", tools=[llm_version], key=API_KEY
|
|
)
|
|
assert "".join(chain) == "The current version of *llm* is **0.fixed-version**."
|
|
|
|
|
|
# This response contains streaming variant "b" where arguments="{}" is the first partial stream received.
|
|
@pytest.mark.vcr(record_mode="none")
|
|
def test_tools_streaming_variant_b():
|
|
model = llm.get_model("gpt-4.1-mini")
|
|
chain = model.chain(
|
|
"What is the current llm version?", tools=[llm_version], key=API_KEY
|
|
)
|
|
assert "".join(chain) == "The current version of *llm* is **0.fixed-version**."
|
|
|
|
|
|
# This response contains streaming variant "c".
|
|
@pytest.mark.vcr(record_mode="none")
|
|
def test_tools_streaming_variant_c():
|
|
model = llm.get_model("gpt-4.1-mini")
|
|
chain = model.chain(
|
|
"What is the current llm version?", tools=[llm_version], key=API_KEY
|
|
)
|
|
assert (
|
|
"".join(chain)
|
|
== "The installed version of LLM on this system is 0.fixed-version."
|
|
)
|
|
|
|
|
|
# This response contains streaming variant "d" where a no-argument tool call
|
|
# streams arguments=null and never sends a "{}" chunk, so the accumulated
|
|
# arguments string stays empty - json.loads("") used to raise here.
|
|
@pytest.mark.vcr(record_mode="none")
|
|
def test_tools_streaming_variant_d():
|
|
model = llm.get_model("gpt-4.1-mini")
|
|
chain = model.chain(
|
|
"What is the current llm version?", tools=[llm_version], key=API_KEY
|
|
)
|
|
assert "".join(chain) == "The current version of *llm* is **0.fixed-version**."
|