This commit is contained in:
Robert Keus 2026-10-28 06:32:34 +00:00 • committed by GitHub
commit fa78f430c8
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
17 changed files with 642 additions and 5 deletions

View file

@ -276,6 +276,7 @@ jobs:
tests/unit/llms/google_pse
tests/unit/llms/github_copilot
tests/unit/llms/gradient_ai
tests/unit/llms/greenpt
tests/unit/llms/groq
tests/unit/llms/heroku
tests/unit/llms/hosted_vllm

View file

@ -0,0 +1 @@

View file

@ -0,0 +1 @@

View file

@ -0,0 +1,64 @@
from collections.abc import Mapping, Sized
from typing import Final
from litellm.llms.hosted_vllm.rerank.transformation import HostedVLLMRerankConfig
from litellm.secret_managers.main import get_secret_str
from litellm.types.rerank import RerankBilledUnits
from litellm.types.utils import ModelInfo
class GreenPTRerankConfig(HostedVLLMRerankConfig):
def get_supported_cohere_rerank_params(self, model: str) -> list[str]: # mutable-ok: mirrors override contract
return ["query", "documents", "top_n", "return_documents"]
def validate_environment(
self,
headers: dict[str, str], # mutable-ok: mirrors override contract
model: str,
api_key: str | None = None,
optional_params: dict[str, object] | None = None, # mutable-ok: mirrors override contract
litellm_params: Mapping[str, object] | None = None,
) -> dict[str, str]: # mutable-ok: mirrors override contract
resolved_api_key: Final = api_key or get_secret_str("GREENPT_API_KEY")
if resolved_api_key is None:
raise ValueError("GreenPT API key is required. Set GREENPT_API_KEY.")
return {
"Authorization": f"Bearer {resolved_api_key}",
"accept": "application/json",
"content-type": "application/json",
**headers,
}
def transform_rerank_request(
self,
model: str,
optional_rerank_params: dict[str, object], # mutable-ok: mirrors override contract
headers: dict[str, str], # mutable-ok: mirrors override contract
litellm_params: dict[str, object] | None = None, # mutable-ok: mirrors override contract
) -> dict[str, object]: # mutable-ok: mirrors override contract
request: Final[Mapping[str, object]] = super().transform_rerank_request( # pyright: ignore[reportUnknownMemberType, reportUnknownVariableType] # HostedVLLMRerankConfig returns an untyped dict
model=model,
optional_rerank_params=optional_rerank_params,
headers=headers,
litellm_params=litellm_params,
)
documents: Final = optional_rerank_params["documents"]
top_n_default: Final = len(documents) if isinstance(documents, Sized) else None
return {"top_n": top_n_default, **request} if top_n_default is not None else {**request}
def calculate_rerank_cost(
self,
model: str,
custom_llm_provider: str | None = None,
billed_units: RerankBilledUnits | None = None,
model_info: ModelInfo | None = None,
) -> tuple[float, float]:
if model_info is None or billed_units is None:
return 0.0, 0.0
input_cost_per_token: Final = model_info.get("input_cost_per_token")
if input_cost_per_token is None:
return 0.0, 0.0
total_tokens: Final = billed_units.get("total_tokens")
if total_tokens is None:
return 0.0, 0.0
return input_cost_per_token * total_tokens, 0.0

View file

@ -79,12 +79,20 @@ class JSONProviderRegistry:
return next((provider for provider in cls._providers.values() if provider.base_url == base_url), None)
@classmethod
def supports_responses_api(cls, slug: str) -> bool:
"""Check if a JSON provider supports the Responses API"""
def _supports_endpoint(cls, slug: str, endpoint: str) -> bool:
provider: Final = cls._providers.get(slug)
if provider is None:
return False
return "/v1/responses" in provider.supported_endpoints
return endpoint in provider.supported_endpoints
@classmethod
def supports_responses_api(cls, slug: str) -> bool:
"""Check if a JSON provider supports the Responses API"""
return cls._supports_endpoint(slug, "/v1/responses")
@classmethod
def supports_embeddings_api(cls, slug: str) -> bool:
return cls._supports_endpoint(slug, "/v1/embeddings")
@classmethod
def list_providers(cls) -> list:

View file

@ -168,6 +168,15 @@
},
"supported_endpoints": ["/v1/chat/completions", "/v1/responses"]
},
"greenpt": {
"base_url": "https://api.greenpt.ai/v1",
"api_key_env": "GREENPT_API_KEY",
"api_base_env": "GREENPT_API_BASE",
"param_mappings": {
"max_completion_tokens": "max_tokens"
},
"supported_endpoints": ["/v1/chat/completions", "/v1/embeddings"]
},
"meta": {
"base_url": "https://api.meta.ai/v1",
"api_key_env": "META_API_KEY",

View file

@ -6961,6 +6961,7 @@ def embedding(
custom_llm_provider == "openai_like"
or custom_llm_provider == "llamafile"
or custom_llm_provider == "lm_studio"
or JSONProviderRegistry.supports_embeddings_api(custom_llm_provider)
):
api_base = api_base or litellm.api_base or get_secret_str("OPENAI_LIKE_API_BASE")

View file

@ -61437,6 +61437,148 @@
],
"supports_audio_output": true
},
"greenpt/glm-5.2": {
"input_cost_per_token": 1.501764e-06,
"litellm_provider": "greenpt",
"max_input_tokens": 1000000,
"max_output_tokens": 131072,
"max_tokens": 131072,
"mode": "chat",
"output_cost_per_token": 5.256174e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/green-embedding": {
"input_cost_per_token": 2.2754e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_output_tokens": 2560,
"max_tokens": 2560,
"mode": "embedding",
"output_cost_per_token": 0,
"source": "https://docs.greenpt.ai/pricing"
},
"greenpt/green-rerank": {
"input_cost_per_token": 1.36524e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_tokens": 32768,
"mode": "rerank",
"source": "https://docs.greenpt.ai/pricing"
},
"greenpt/green-l": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "chat",
"output_cost_per_token": 9.1016e-07,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-l-raw": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "chat",
"output_cost_per_token": 9.1016e-07,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-r": {
"input_cost_per_token": 3.98195e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 131072,
"max_output_tokens": 32768,
"max_tokens": 32768,
"mode": "chat",
"output_cost_per_token": 1.080815e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-r-raw": {
"input_cost_per_token": 3.98195e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 131072,
"max_output_tokens": 32768,
"max_tokens": 32768,
"mode": "chat",
"output_cost_per_token": 1.080815e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/kimi-k2.6": {
"input_cost_per_token": 8.30521e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 262144,
"max_output_tokens": 262144,
"max_tokens": 262144,
"mode": "chat",
"output_cost_per_token": 4.380145e-06,
"source": "https://docs.greenpt.ai/pricing",
"supported_modalities": [
"text",
"image"
],
"supported_output_modalities": [
"text"
],
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/kimi-k2.7-code": {
"input_cost_per_token": 9.44291e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 262144,
"max_output_tokens": 262144,
"max_tokens": 262144,
"mode": "chat",
"output_cost_per_token": 4.380145e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/minimax-m2.5": {
"input_cost_per_token": 1.93409e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 200000,
"max_output_tokens": 128000,
"max_tokens": 128000,
"mode": "chat",
"output_cost_per_token": 1.126323e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/qwen3-embedding-8b": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "embedding",
"output_cost_per_token": 0,
"source": "https://docs.greenpt.ai/pricing"
},
"fallback_generalizations": {
"rules": [
{

View file

@ -33,7 +33,7 @@ async def arerank(
query: str,
documents: list[str | dict[str, object]],
custom_llm_provider: (
Literal["cohere", "together_ai", "deepinfra", "fireworks_ai", "voyage", "watsonx"] | None
Literal["cohere", "together_ai", "deepinfra", "fireworks_ai", "voyage", "watsonx", "greenpt"] | None
) = None,
top_n: int | None = None,
rank_fields: list[str] | None = None,
@ -106,6 +106,7 @@ def rerank(
"fireworks_ai",
"voyage",
"watsonx",
"greenpt",
]
| None
) = None,

View file

@ -4532,6 +4532,7 @@ class LlmProviders(str, Enum):
TOPAZ = "topaz"
SAP_GENERATIVE_AI_HUB = "sap"
ASSEMBLYAI = "assemblyai"
GREENPT = "greenpt"
AZURE_SPEECH = "azure_speech"
CHARITY_ENGINE = "charity_engine"
GITHUB_COPILOT = "github_copilot"

View file

@ -9583,7 +9583,12 @@ class ProviderConfigManager:
)
return get_dashscope_family_rerank_config(provider.value)
return litellm.CohereRerankConfig()
from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig
config_class: Final = {
litellm.LlmProviders.GREENPT: GreenPTRerankConfig,
}.get(provider, litellm.CohereRerankConfig)
return config_class()
@staticmethod
@lru_cache(maxsize=1)

View file

@ -61437,6 +61437,148 @@
],
"supports_audio_output": true
},
"greenpt/glm-5.2": {
"input_cost_per_token": 1.501764e-06,
"litellm_provider": "greenpt",
"max_input_tokens": 1000000,
"max_output_tokens": 131072,
"max_tokens": 131072,
"mode": "chat",
"output_cost_per_token": 5.256174e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/green-embedding": {
"input_cost_per_token": 2.2754e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_output_tokens": 2560,
"max_tokens": 2560,
"mode": "embedding",
"output_cost_per_token": 0,
"source": "https://docs.greenpt.ai/pricing"
},
"greenpt/green-rerank": {
"input_cost_per_token": 1.36524e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_tokens": 32768,
"mode": "rerank",
"source": "https://docs.greenpt.ai/pricing"
},
"greenpt/green-l": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "chat",
"output_cost_per_token": 9.1016e-07,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-l-raw": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 128000,
"max_output_tokens": 32000,
"max_tokens": 32000,
"mode": "chat",
"output_cost_per_token": 9.1016e-07,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-r": {
"input_cost_per_token": 3.98195e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 131072,
"max_output_tokens": 32768,
"max_tokens": 32768,
"mode": "chat",
"output_cost_per_token": 1.080815e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/green-r-raw": {
"input_cost_per_token": 3.98195e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 131072,
"max_output_tokens": 32768,
"max_tokens": 32768,
"mode": "chat",
"output_cost_per_token": 1.080815e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/kimi-k2.6": {
"input_cost_per_token": 8.30521e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 262144,
"max_output_tokens": 262144,
"max_tokens": 262144,
"mode": "chat",
"output_cost_per_token": 4.380145e-06,
"source": "https://docs.greenpt.ai/pricing",
"supported_modalities": [
"text",
"image"
],
"supported_output_modalities": [
"text"
],
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true,
"supports_vision": true
},
"greenpt/kimi-k2.7-code": {
"input_cost_per_token": 9.44291e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 262144,
"max_output_tokens": 262144,
"max_tokens": 262144,
"mode": "chat",
"output_cost_per_token": 4.380145e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/minimax-m2.5": {
"input_cost_per_token": 1.93409e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 200000,
"max_output_tokens": 128000,
"max_tokens": 128000,
"mode": "chat",
"output_cost_per_token": 1.126323e-06,
"source": "https://docs.greenpt.ai/pricing",
"supports_function_calling": true,
"supports_reasoning": true,
"supports_tool_choice": true
},
"greenpt/qwen3-embedding-8b": {
"input_cost_per_token": 2.84425e-07,
"litellm_provider": "greenpt",
"max_input_tokens": 32768,
"max_output_tokens": 4096,
"max_tokens": 4096,
"mode": "embedding",
"output_cost_per_token": 0,
"source": "https://docs.greenpt.ai/pricing"
},
"fallback_generalizations": {
"rules": [
{

View file

@ -5248,6 +5248,24 @@
"a2a": false,
"generateContent": false
}
},
"greenpt": {
"display_name": "GreenPT (`greenpt`)",
"url": "https://docs.greenpt.ai",
"endpoints": {
"chat_completions": true,
"messages": false,
"responses": false,
"embeddings": true,
"image_generations": false,
"audio_transcriptions": false,
"audio_speech": false,
"moderations": false,
"batches": false,
"rerank": true,
"a2a": false,
"interactions": false
}
}
},
"endpoints": {

View file

View file

@ -0,0 +1,133 @@
import json
from unittest.mock import Mock, patch
import pytest
import litellm
from litellm.cost_calculator import rerank_cost
from litellm.llms.custom_httpx.http_handler import HTTPHandler
from litellm.llms.greenpt.rerank.transformation import GreenPTRerankConfig
def test_greenpt_supported_params():
assert GreenPTRerankConfig().get_supported_cohere_rerank_params("green-rerank") == [
"query",
"documents",
"top_n",
"return_documents",
]
def test_validate_environment_with_api_key():
headers = GreenPTRerankConfig().validate_environment(
headers={},
model="green-rerank",
api_key="test-key",
)
assert headers == {
"Authorization": "Bearer test-key",
"accept": "application/json",
"content-type": "application/json",
}
def test_validate_environment_requires_api_key(monkeypatch):
monkeypatch.delenv("GREENPT_API_KEY", raising=False)
with pytest.raises(ValueError, match="GreenPT API key is required"):
GreenPTRerankConfig().validate_environment(
headers={},
model="green-rerank",
)
def test_transform_rerank_request_defaults_top_n():
request = GreenPTRerankConfig().transform_rerank_request(
model="green-rerank",
optional_rerank_params={
"query": "low-carbon inference",
"documents": ["GreenPT", "Other provider"],
"return_documents": False,
},
headers={},
)
assert request == {
"model": "green-rerank",
"query": "low-carbon inference",
"documents": ["GreenPT", "Other provider"],
"top_n": 2,
"return_documents": False,
}
def test_transform_rerank_request_preserves_top_n():
request = GreenPTRerankConfig().transform_rerank_request(
model="green-rerank",
optional_rerank_params={
"query": "low-carbon inference",
"documents": ["GreenPT", "Other provider"],
"top_n": 1,
},
headers={},
)
assert request["top_n"] == 1
def test_calculate_rerank_cost():
prompt_cost, completion_cost = GreenPTRerankConfig().calculate_rerank_cost(
model="green-rerank",
billed_units={"total_tokens": 1000},
model_info={"input_cost_per_token": 1.36524e-07},
)
assert prompt_cost == pytest.approx(0.000136524)
assert completion_cost == 0.0
def test_public_rerank_cost_uses_token_pricing():
prompt_cost, completion_cost = rerank_cost(
model="greenpt/green-rerank",
custom_llm_provider=None,
billed_units={"total_tokens": 1000},
)
assert prompt_cost == pytest.approx(0.000136524)
assert completion_cost == 0.0
@pytest.mark.parametrize(
("billed_units", "model_info"),
[
(None, None),
({}, {"input_cost_per_token": 1.36524e-07}),
({"total_tokens": 1000}, {}),
],
)
def test_calculate_rerank_cost_missing_usage_or_pricing(billed_units, model_info):
assert GreenPTRerankConfig().calculate_rerank_cost(
model="green-rerank",
billed_units=billed_units,
model_info=model_info,
) == (0.0, 0.0)
def test_public_rerank_call_sends_request_and_costs_tokens():
response = Mock()
response.raise_for_status = Mock()
response.json.return_value = {
"id": "rerank-1",
"results": [{"index": 0, "relevance_score": 0.9}],
"usage": {"total_tokens": 1000},
}
with patch.object(HTTPHandler, "post", return_value=response) as mock_post:
result = litellm.rerank(
model="greenpt/green-rerank",
query="low-carbon inference",
documents=["GreenPT", "Other provider"],
api_key="test-key",
client=HTTPHandler(),
)
assert mock_post.call_args.kwargs["url"] == "https://api.greenpt.ai/v1/rerank"
assert mock_post.call_args.kwargs["headers"]["Authorization"] == "Bearer test-key"
assert json.loads(mock_post.call_args.kwargs["data"])["top_n"] == 2
assert result.results[0]["index"] == 0
assert result._hidden_params["response_cost"] == pytest.approx(0.000136524)

View file

@ -0,0 +1,110 @@
import json
from unittest.mock import Mock, patch
import litellm
from litellm.llms.custom_httpx.http_handler import HTTPHandler
from litellm.llms.openai_like.dynamic_config import create_config_class
from litellm.llms.openai_like.json_loader import JSONProviderRegistry
from litellm.types.utils import LlmProviders
from litellm.utils import ProviderConfigManager
GREENPT_BASE_URL = "https://api.greenpt.ai/v1"
def _get_config():
provider = JSONProviderRegistry.get("greenpt")
assert provider is not None
config_class = create_config_class(provider)
return config_class()
def test_greenpt_provider_registered():
provider = JSONProviderRegistry.get("greenpt")
assert provider is not None
assert provider.base_url == GREENPT_BASE_URL
assert provider.api_key_env == "GREENPT_API_KEY"
assert provider.api_base_env == "GREENPT_API_BASE"
def test_greenpt_resolves_env_api_key(monkeypatch):
config = _get_config()
monkeypatch.setenv("GREENPT_API_KEY", "test-key")
api_base, api_key = config._get_openai_compatible_provider_info(None, None)
assert api_base == GREENPT_BASE_URL
assert api_key == "test-key"
def test_greenpt_maps_max_completion_tokens():
config = _get_config()
params = config.map_openai_params(
non_default_params={"max_completion_tokens": 256},
optional_params={},
model="greenpt/glm-5.2",
drop_params=False,
)
assert params == {"max_tokens": 256}
def test_greenpt_complete_url_appends_endpoint():
config = _get_config()
url = config.get_complete_url(
api_base=GREENPT_BASE_URL,
api_key="test-key",
model="greenpt/glm-5.2",
optional_params={},
litellm_params={},
stream=False,
)
assert url == f"{GREENPT_BASE_URL}/chat/completions"
def test_greenpt_provider_resolution():
from litellm.litellm_core_utils.get_llm_provider_logic import get_llm_provider
model, provider, _, api_base = get_llm_provider(
model="greenpt/glm-5.2",
custom_llm_provider=None,
api_base=None,
api_key=None,
)
assert model == "glm-5.2"
assert provider == "greenpt"
assert api_base == GREENPT_BASE_URL
def test_greenpt_embedding_dispatches():
response = Mock()
response.raise_for_status = Mock()
response.json.return_value = {
"object": "list",
"data": [{"object": "embedding", "embedding": [0.1, 0.2], "index": 0}],
"model": "green-embedding",
"usage": {"prompt_tokens": 2, "total_tokens": 2},
}
client = HTTPHandler()
with patch.object(HTTPHandler, "post", return_value=response) as mock_post:
result = litellm.embedding(
model="greenpt/green-embedding",
input=["renewable inference"],
api_key="test-key",
client=client,
)
assert result.data[0]["embedding"] == [0.1, 0.2]
assert mock_post.call_args.args[0] == f"{GREENPT_BASE_URL}/embeddings"
assert json.loads(mock_post.call_args.kwargs["data"]) == {
"model": "green-embedding",
"input": ["renewable inference"],
}
def test_greenpt_rerank_config():
config = ProviderConfigManager.get_provider_rerank_config(
model="green-rerank",
provider=LlmProviders.GREENPT,
api_base=GREENPT_BASE_URL,
present_version_params=[],
)
assert config.get_complete_url(GREENPT_BASE_URL, "green-rerank") == (f"{GREENPT_BASE_URL}/rerank")