refactor(opencode): move completion dispatch logic out of main.py into the provider package

This commit is contained in:
Sean Murphy 2026-09-14 10:33:25 -07:00
parent 0192faa6bd
commit 0ddf4ff30f
2 changed files with 159 additions and 101 deletions

View file

@ -0,0 +1,139 @@
"""
Completion dispatch for the OpenCode providers.
Resolves the key, base URL and session header, then sends the request over the
wire format the model speaks: Anthropic Messages through the Anthropic chat
handler, which translates OpenAI input and output, or OpenAI Chat Completions.
"""
from collections.abc import Callable, Coroutine, Mapping
from types import MappingProxyType
from typing import TYPE_CHECKING, Final
import httpx
import tiktoken
import litellm
from litellm.litellm_core_utils.streaming_handler import CustomStreamWrapper
from litellm.llms.anthropic.chat.handler import AnthropicChatCompletion
from litellm.llms.custom_httpx.http_handler import AsyncHTTPHandler, HTTPHandler
from litellm.llms.custom_httpx.llm_http_handler import BaseLLMHTTPHandler
from litellm.llms.opencode.chat.messages_transformation import (
OpenCodeMessagesConfig,
is_messages_model,
)
from litellm.llms.opencode.chat.transformation import OpenCodeConfig
from litellm.llms.opencode.common_utils import (
resolve_opencode_api_base,
resolve_opencode_api_key,
with_opencode_session_header,
)
from litellm.types.llms.openai import AllMessageValues
from litellm.types.utils import ModelResponse
if TYPE_CHECKING:
from aiohttp import ClientSession
from litellm.litellm_core_utils.litellm_logging import Logging as LiteLLMLoggingObj
def complete_opencode(
*,
custom_llm_provider: str,
model: str,
messages: list[AllMessageValues], # mutable-ok: both handlers take the message list as-is
api_key: str | None,
api_base: str | None,
headers: Mapping[str, str] | None,
litellm_params: dict[str, object], # mutable-ok: both handlers take litellm_params as a dict
optional_params: dict[str, object], # mutable-ok: both handlers take optional_params as a dict
model_response: ModelResponse,
logging_obj: "LiteLLMLoggingObj",
acompletion: bool,
stream: bool | None,
timeout: float | str | httpx.Timeout | None,
client: HTTPHandler | AsyncHTTPHandler | None,
shared_session: "ClientSession | None",
anthropic_chat_handler: AnthropicChatCompletion,
chat_handler: BaseLLMHTTPHandler,
encoding: tiktoken.Encoding,
print_verbose: Callable[[str], object],
) -> ModelResponse | CustomStreamWrapper | Coroutine[object, object, ModelResponse | CustomStreamWrapper]:
surface: Final = "go" if custom_llm_provider == "opencode_go" else "zen"
provider: Final = f"opencode_{surface}"
resolved_key: Final = resolve_opencode_api_key(surface, api_key)
resolved_base: Final = resolve_opencode_api_base(surface, api_base)
session_headers: Final = with_opencode_session_header(
surface,
headers or litellm.headers or {}, # mutable-ok: empty dict fallback for headers
litellm_params,
)
no_params: Final = MappingProxyType({})
if is_messages_model(surface, model):
# AnthropicChatCompletion builds its headers from AnthropicConfig
# directly, and that resolves a missing key from ANTHROPIC_API_KEY, so
# a missing OpenCode key has to be rejected here rather than sending a
# first-party Anthropic credential to opencode.ai.
if resolved_key is None:
raise ValueError(
f"OpenCode API key is required. Set OPENCODE_{surface.upper()}_API_KEY "
f"or OPENCODE_API_KEY, or pass api_key."
)
return anthropic_chat_handler.completion(
model=model,
messages=messages,
api_base=OpenCodeMessagesConfig(surface=surface).get_complete_url(
api_base=resolved_base,
api_key=resolved_key,
model=model,
optional_params=no_params,
litellm_params=no_params,
),
acompletion=acompletion,
custom_prompt_dict=litellm.custom_prompt_dict,
model_response=model_response,
print_verbose=print_verbose,
optional_params=optional_params,
litellm_params=litellm_params,
logger_fn=None,
encoding=encoding,
api_key=resolved_key,
logging_obj=logging_obj,
headers=session_headers, # Bearer is chat-arm only; the config injects x-api-key
timeout=timeout, # pyright: ignore[reportArgumentType] # the dispatch context widens timeout to str|None for every provider; narrowing it is a repo-wide gap, not opencode-specific
client=client,
custom_llm_provider=provider,
)
response: Final = chat_handler.completion(
model=model,
stream=stream,
messages=messages,
acompletion=acompletion,
api_base=OpenCodeConfig(surface=surface).get_complete_url(
api_base=resolved_base,
api_key=resolved_key,
model=model,
optional_params=no_params,
litellm_params=no_params,
),
model_response=model_response,
optional_params=optional_params,
litellm_params=litellm_params,
shared_session=shared_session,
custom_llm_provider=provider,
timeout=timeout, # pyright: ignore[reportArgumentType] # the dispatch context widens timeout to str|None for every provider; narrowing it is a repo-wide gap, not opencode-specific
headers=(
{**session_headers, "Authorization": f"Bearer {resolved_key}"} # mutable-ok: handler mutates headers
if resolved_key is not None
else session_headers
),
encoding=encoding,
api_key=resolved_key,
logging_obj=logging_obj,
client=client,
)
logging_obj.post_call(input=messages, api_key=resolved_key, original_response=response)
return response

View file

@ -3484,110 +3484,29 @@ def _complete_openrouter(ctx: _CompletionDispatchContext) -> _CompletionDispatch
def _complete_opencode(
ctx: _CompletionDispatchContext,
) -> _CompletionDispatchResult:
acompletion: Final = ctx.acompletion
api_base = ctx.api_base # rebind-ok: resolved below from module/env fallbacks
api_key = ctx.api_key # rebind-ok: resolved below from module/env fallbacks
client: Final = ctx.client
custom_llm_provider: Final = ctx.custom_llm_provider
headers: Final = ctx.headers
litellm_params: Final = ctx.litellm_params
logging: Final = ctx.logging
messages: Final = ctx.messages
model: Final = ctx.model
model_response: Final = ctx.model_response
optional_params: Final = ctx.optional_params
shared_session: Final = ctx.shared_session
stream: Final = ctx.stream
timeout: Final = ctx.timeout
from litellm.llms.opencode.chat.handler import complete_opencode
from litellm.llms.opencode.common_utils import (
resolve_opencode_api_base,
resolve_opencode_api_key,
with_opencode_session_header,
)
surface: Final = "go" if custom_llm_provider == "opencode_go" else "zen"
_base_url: Final = "https://opencode.ai/zen/go/v1" if surface == "go" else "https://opencode.ai/zen/v1"
api_base = (
resolve_opencode_api_base(surface, api_base) or _base_url
) # rebind-ok: resolved from module/env fallbacks
api_key = resolve_opencode_api_key(surface, api_key) # rebind-ok: resolved from module/env fallbacks
base_headers: Final = with_opencode_session_header(
surface,
headers or litellm.headers or {}, # mutable-ok: empty dict fallback for headers
litellm_params,
)
_headers: Final = (
{**base_headers, "Authorization": f"Bearer {api_key}"} # mutable-ok: dict literal for request headers
if api_key is not None
else base_headers
)
from litellm.llms.opencode.chat.messages_transformation import is_messages_model
if is_messages_model(surface, model):
# AnthropicChatCompletion builds its headers from AnthropicConfig
# directly, and that resolves a missing key from ANTHROPIC_API_KEY, so
# a missing OpenCode key has to be rejected here rather than sending a
# first-party Anthropic credential to opencode.ai.
if api_key is None:
raise ValueError(
f"OpenCode API key is required. Set OPENCODE_{surface.upper()}_API_KEY "
f"or OPENCODE_API_KEY, or pass api_key."
)
messages_base: Final = api_base.rstrip("/")
messages_url: Final = (
messages_base
if messages_base.endswith("/v1/messages")
else f"{messages_base}/messages"
if messages_base.endswith("/v1")
else f"{messages_base}/v1/messages"
)
return anthropic_chat_completions.completion(
model=model,
messages=messages,
api_base=messages_url,
acompletion=acompletion,
custom_prompt_dict=litellm.custom_prompt_dict,
model_response=model_response,
print_verbose=print_verbose,
optional_params=optional_params,
litellm_params=litellm_params,
logger_fn=None,
encoding=_get_encoding(),
api_key=api_key,
logging_obj=logging,
headers=base_headers, # Bearer is chat-arm only; the config injects x-api-key
timeout=timeout, # pyright: ignore[reportArgumentType] # ctx.timeout is str|None-widened same as every other _complete_* dispatch handler; narrowing is a pre-existing, repo-wide gap in _CompletionDispatchContext, not opencode-specific
client=client,
custom_llm_provider=f"opencode_{surface}",
)
## COMPLETION CALL (chat arm)
response = base_llm_http_handler.completion( # rebind-ok: chat arm result
model=model,
stream=stream,
messages=messages,
acompletion=acompletion,
api_base=api_base,
model_response=model_response,
optional_params=optional_params,
litellm_params=litellm_params,
shared_session=shared_session,
custom_llm_provider=f"opencode_{surface}",
timeout=timeout, # pyright: ignore[reportArgumentType] # ctx.timeout is str|None-widened same as every other _complete_* dispatch handler; narrowing is a pre-existing, repo-wide gap in _CompletionDispatchContext, not opencode-specific
headers=_headers,
return complete_opencode(
custom_llm_provider=ctx.custom_llm_provider,
model=ctx.model,
messages=ctx.messages,
api_key=ctx.api_key,
api_base=ctx.api_base,
headers=ctx.headers,
litellm_params=ctx.litellm_params,
optional_params=ctx.optional_params,
model_response=ctx.model_response,
logging_obj=ctx.logging,
acompletion=ctx.acompletion,
stream=ctx.stream,
timeout=ctx.timeout,
client=ctx.client,
shared_session=ctx.shared_session,
anthropic_chat_handler=anthropic_chat_completions,
chat_handler=base_llm_http_handler,
encoding=_get_encoding(),
api_key=api_key,
logging_obj=logging,
client=client,
print_verbose=print_verbose,
)
## LOGGING
logging.post_call(input=messages, api_key=api_key, original_response=response)
return response
def _complete_vercel_ai_gateway(