free-claude-code/config/settings.py
Pavel Yurchenko e719e4aed2
feat: deepseek api support (#118)
## Summary

* add native DeepSeek provider support via the shared OpenAI-compatible
provider base
* allow `deepseek/...` model prefixes in config validation
* add `DEEPSEEK_API_KEY` and `DEEPSEEK_BASE_URL` settings
* add DeepSeek entries to `.env.example` and `config/env.example`
* implement `DeepSeekProvider` and register it in provider dependencies
* add a DeepSeek request builder with DeepSeek-specific thinking payload
handling
* preserve Anthropic thinking blocks as `reasoning_content` for
DeepSeek-compatible continuation flows
* update `claude-pick` to discover DeepSeek models from the DeepSeek API
* document DeepSeek usage in `README.md`
* add tests for config validation, provider dependency wiring, request
building, and streaming behavior

## Motivation

DeepSeek exposes an OpenAI-compatible API and can be used directly
without routing through OpenRouter. This lets users spend their existing
DeepSeek balance through the proxy while keeping the same Claude Code
workflow and per-model provider mapping.

## Example

```dotenv
DEEPSEEK_API_KEY="sk-..."
DEEPSEEK_BASE_URL="https://api.deepseek.com"

MODEL_OPUS="deepseek/deepseek-reasoner"
MODEL_SONNET="deepseek/deepseek-chat"
MODEL_HAIKU="deepseek/deepseek-chat"
MODEL="deepseek/deepseek-chat"

---------

Co-authored-by: Alishahryar1 <alishahryar2@gmail.com>
2026-04-22 17:06:01 -07:00

304 lines
11 KiB
Python

"""Centralized configuration using Pydantic Settings."""
import os
from collections.abc import Mapping
from functools import lru_cache
from pathlib import Path
from typing import Any
from pydantic import Field, field_validator, model_validator
from pydantic_settings import BaseSettings, SettingsConfigDict
from .nim import NimSettings
def _env_files() -> tuple[Path, ...]:
"""Return env file paths in priority order (later overrides earlier)."""
files: list[Path] = [
Path.home() / ".config" / "free-claude-code" / ".env",
Path(".env"),
]
if explicit := os.environ.get("FCC_ENV_FILE"):
files.append(Path(explicit))
return tuple(files)
def _configured_env_files(model_config: Mapping[str, Any]) -> tuple[Path, ...]:
"""Return the currently configured env files for Settings."""
configured = model_config.get("env_file")
if configured is None:
return ()
if isinstance(configured, (str, Path)):
return (Path(configured),)
return tuple(Path(item) for item in configured)
def _env_file_contains_key(path: Path, key: str) -> bool:
"""Check whether a dotenv-style file defines the given key."""
if not path.is_file():
return False
try:
for line in path.read_text(encoding="utf-8").splitlines():
stripped = line.strip()
if not stripped or stripped.startswith("#"):
continue
if stripped.startswith("export "):
stripped = stripped[7:].lstrip()
name, sep, _value = stripped.partition("=")
if sep and name.strip() == key:
return True
except OSError:
return False
return False
def _removed_env_var_message(model_config: Mapping[str, Any]) -> str | None:
"""Return a migration error for removed env vars, if present."""
removed_key = "NIM_ENABLE_THINKING"
replacement = "ENABLE_THINKING"
if removed_key in os.environ:
return (
f"{removed_key} has been removed in this release. "
f"Rename it to {replacement}."
)
for env_file in _configured_env_files(model_config):
if _env_file_contains_key(env_file, removed_key):
return (
f"{removed_key} has been removed in this release. "
f"Rename it to {replacement}. Found in {env_file}."
)
return None
class Settings(BaseSettings):
"""Application settings loaded from environment variables."""
# ==================== OpenRouter Config ====================
open_router_api_key: str = Field(default="", validation_alias="OPENROUTER_API_KEY")
# ==================== DeepSeek Config ====================
deepseek_api_key: str = Field(default="", validation_alias="DEEPSEEK_API_KEY")
# ==================== Messaging Platform Selection ====================
# Valid: "telegram" | "discord"
messaging_platform: str = Field(
default="discord", validation_alias="MESSAGING_PLATFORM"
)
# ==================== NVIDIA NIM Config ====================
nvidia_nim_api_key: str = ""
# ==================== LM Studio Config ====================
lm_studio_base_url: str = Field(
default="http://localhost:1234/v1",
validation_alias="LM_STUDIO_BASE_URL",
)
# ==================== Llama.cpp Config ====================
llamacpp_base_url: str = Field(
default="http://localhost:8080/v1",
validation_alias="LLAMACPP_BASE_URL",
)
# ==================== Model ====================
# All Claude model requests are mapped to this single model (fallback)
# Format: provider_type/model/name
model: str = "nvidia_nim/meta/llama3-70b-instruct"
# Per-model overrides (optional, falls back to MODEL)
# Each can use a different provider
model_opus: str | None = Field(default=None, validation_alias="MODEL_OPUS")
model_sonnet: str | None = Field(default=None, validation_alias="MODEL_SONNET")
model_haiku: str | None = Field(default=None, validation_alias="MODEL_HAIKU")
# ==================== Provider Rate Limiting ====================
provider_rate_limit: int = Field(default=40, validation_alias="PROVIDER_RATE_LIMIT")
provider_rate_window: int = Field(
default=60, validation_alias="PROVIDER_RATE_WINDOW"
)
provider_max_concurrency: int = Field(
default=5, validation_alias="PROVIDER_MAX_CONCURRENCY"
)
enable_thinking: bool = Field(default=True, validation_alias="ENABLE_THINKING")
# ==================== HTTP Client Timeouts ====================
http_read_timeout: float = Field(
default=300.0, validation_alias="HTTP_READ_TIMEOUT"
)
http_write_timeout: float = Field(
default=10.0, validation_alias="HTTP_WRITE_TIMEOUT"
)
http_connect_timeout: float = Field(
default=2.0, validation_alias="HTTP_CONNECT_TIMEOUT"
)
# ==================== Fast Prefix Detection ====================
fast_prefix_detection: bool = True
# ==================== Optimizations ====================
enable_network_probe_mock: bool = True
enable_title_generation_skip: bool = True
enable_suggestion_mode_skip: bool = True
enable_filepath_extraction_mock: bool = True
# ==================== NIM Settings ====================
nim: NimSettings = Field(default_factory=NimSettings)
# ==================== Voice Note Transcription ====================
voice_note_enabled: bool = Field(
default=True, validation_alias="VOICE_NOTE_ENABLED"
)
# Device: "cpu" | "cuda" | "nvidia_nim"
# - "cpu"/"cuda": local Whisper (requires voice_local extra: uv sync --extra voice_local)
# - "nvidia_nim": NVIDIA NIM Whisper API (requires voice extra: uv sync --extra voice)
whisper_device: str = Field(default="cpu", validation_alias="WHISPER_DEVICE")
# Whisper model ID or short name (for local Whisper) or NVIDIA NIM model (for nvidia_nim)
# Local Whisper: "tiny", "base", "small", "medium", "large-v2", "large-v3", "large-v3-turbo"
# NVIDIA NIM: "nvidia/parakeet-ctc-1.1b-asr", "openai/whisper-large-v3", etc.
whisper_model: str = Field(default="base", validation_alias="WHISPER_MODEL")
# Hugging Face token for faster model downloads (optional, for local Whisper)
hf_token: str = Field(default="", validation_alias="HF_TOKEN")
# ==================== Bot Wrapper Config ====================
telegram_bot_token: str | None = None
allowed_telegram_user_id: str | None = None
discord_bot_token: str | None = Field(
default=None, validation_alias="DISCORD_BOT_TOKEN"
)
allowed_discord_channels: str | None = Field(
default=None, validation_alias="ALLOWED_DISCORD_CHANNELS"
)
claude_workspace: str = "./agent_workspace"
allowed_dir: str = ""
# ==================== Server ====================
host: str = "0.0.0.0"
port: int = 8082
log_file: str = "server.log"
# Optional server API key to protect endpoints (Anthropic-style)
# Set via env `ANTHROPIC_AUTH_TOKEN`. When empty, no auth is required.
anthropic_auth_token: str = Field(
default="", validation_alias="ANTHROPIC_AUTH_TOKEN"
)
@model_validator(mode="before")
@classmethod
def reject_removed_env_vars(cls, data: Any) -> Any:
"""Fail fast when removed environment variables are still configured."""
if message := _removed_env_var_message(cls.model_config):
raise ValueError(message)
return data
# Handle empty strings for optional string fields
@field_validator(
"telegram_bot_token",
"allowed_telegram_user_id",
"discord_bot_token",
"allowed_discord_channels",
mode="before",
)
@classmethod
def parse_optional_str(cls, v: Any) -> Any:
if v == "":
return None
return v
@field_validator("whisper_device")
@classmethod
def validate_whisper_device(cls, v: str) -> str:
if v not in ("cpu", "cuda", "nvidia_nim"):
raise ValueError(
f"whisper_device must be 'cpu', 'cuda', or 'nvidia_nim', got {v!r}"
)
return v
@field_validator("model", "model_opus", "model_sonnet", "model_haiku")
@classmethod
def validate_model_format(cls, v: str | None) -> str | None:
if v is None:
return None
valid_providers = (
"nvidia_nim",
"open_router",
"deepseek",
"lmstudio",
"llamacpp",
)
if "/" not in v:
raise ValueError(
f"Model must be prefixed with provider type. "
f"Valid providers: {', '.join(valid_providers)}. "
f"Format: provider_type/model/name"
)
provider = v.split("/", 1)[0]
if provider not in valid_providers:
raise ValueError(
f"Invalid provider: '{provider}'. "
f"Supported: 'nvidia_nim', 'open_router', 'deepseek', 'lmstudio', 'llamacpp'"
)
return v
@model_validator(mode="after")
def check_nvidia_nim_api_key(self) -> Settings:
if (
self.voice_note_enabled
and self.whisper_device == "nvidia_nim"
and not self.nvidia_nim_api_key.strip()
):
raise ValueError(
"NVIDIA_NIM_API_KEY is required when WHISPER_DEVICE is 'nvidia_nim'. "
"Set it in your .env file."
)
return self
@property
def provider_type(self) -> str:
"""Extract provider type from the default model string."""
return self.model.split("/", 1)[0]
@property
def model_name(self) -> str:
"""Extract the actual model name from the default model string."""
return self.model.split("/", 1)[1]
def resolve_model(self, claude_model_name: str) -> str:
"""Resolve a Claude model name to the configured provider/model string.
Classifies the incoming Claude model (opus/sonnet/haiku) and
returns the model-specific override if configured, otherwise the fallback MODEL.
"""
name_lower = claude_model_name.lower()
if "opus" in name_lower and self.model_opus is not None:
return self.model_opus
if "haiku" in name_lower and self.model_haiku is not None:
return self.model_haiku
if "sonnet" in name_lower and self.model_sonnet is not None:
return self.model_sonnet
return self.model
@staticmethod
def parse_provider_type(model_string: str) -> str:
"""Extract provider type from any 'provider/model' string."""
return model_string.split("/", 1)[0]
@staticmethod
def parse_model_name(model_string: str) -> str:
"""Extract model name from any 'provider/model' string."""
return model_string.split("/", 1)[1]
model_config = SettingsConfigDict(
env_file=_env_files(),
env_file_encoding="utf-8",
extra="ignore",
)
@lru_cache
def get_settings() -> Settings:
"""Get cached settings instance."""
return Settings()