Buckets:

download
raw
9.88 kB
"""
Configuration for ContextFS.
"""
from enum import Enum
from pathlib import Path
from pydantic import Field
from pydantic_settings import BaseSettings
class BackendType(str, Enum):
"""Storage backend types."""
SQLITE = "sqlite" # SQLite + ChromaDB (default)
POSTGRES = "postgres" # PostgreSQL with pgvector + AGE
SQLITE_FALKORDB = "sqlite+falkordb" # SQLite + ChromaDB + FalkorDB
POSTGRES_FALKORDB = "postgres+falkordb" # PostgreSQL + FalkorDB
class MergeStrategyType(str, Enum):
"""Default merge strategy for lineage operations."""
UNION = "union"
INTERSECTION = "intersection"
LATEST = "latest"
OLDEST = "oldest"
class Config(BaseSettings):
"""
ContextFS configuration.
Loaded from environment variables with CONTEXTFS_ prefix.
"""
# =================================================================
# Backend Selection
# =================================================================
backend: BackendType = BackendType.SQLITE
# Data directory (for sqlite/chromadb backends)
data_dir: Path = Field(default=Path.home() / ".contextfs")
# =================================================================
# SQLite Configuration
# =================================================================
sqlite_filename: str = "context.db"
# =================================================================
# PostgreSQL Configuration
# =================================================================
postgres_url: str = "postgresql://contextfs:contextfs@localhost:5432/contextfs"
postgres_pgvector: bool = True # Enable pgvector extension
postgres_age: bool = True # Enable Apache AGE extension
# =================================================================
# ChromaDB Configuration
# =================================================================
embedding_model: str = "all-MiniLM-L6-v2"
chroma_collection: str = "contextfs_memories"
# None = embedded mode (for tests/dev), "localhost" = HTTP server mode (recommended for production)
# Set CONTEXTFS_CHROMA_HOST=localhost in production to prevent corruption from concurrent access
chroma_host: str | None = None
chroma_port: int = 8000
# Auto-start ChromaDB server if not running (only applies when chroma_host is set)
chroma_auto_server: bool = True
# =================================================================
# Embedding Backend Configuration
# =================================================================
# Backend: "auto" (fastembed if installed, else sentence_transformers),
# "fastembed" (ONNX, faster), or "sentence_transformers"
embedding_backend: str = "auto"
# GPU acceleration: None = auto-detect, True = force GPU, False = force CPU
use_gpu: bool | None = None
# Parallel workers for embedding (None = auto, 0 = all cores)
embedding_parallel_workers: int | None = None
# Batch size for embedding generation (larger = fewer model calls but more memory)
embedding_batch_size: int = 256
# =================================================================
# FalkorDB Configuration
# =================================================================
falkordb_enabled: bool = False
falkordb_host: str = "localhost"
falkordb_port: int = 6379
falkordb_password: str | None = None
falkordb_graph_name: str = "contextfs_memory"
# =================================================================
# Memory Lineage Settings (CORE FEATURE)
# =================================================================
lineage_auto_track: bool = True # Auto-track evolution on updates
lineage_merge_strategy: MergeStrategyType = MergeStrategyType.UNION
lineage_preserve_tags: bool = True # Preserve tags when evolving
auto_link_enabled: bool = True # Auto-link memories based on semantic similarity
auto_link_threshold: float = 0.55 # Minimum similarity score to auto-link (0.0-1.0)
auto_link_max: int = 3 # Maximum auto-links per memory save
# =================================================================
# Search Settings
# =================================================================
default_search_limit: int = 10
min_similarity_score: float = 0.3
# Chunking
chunk_size: int = 1000
chunk_overlap: int = 200
# =================================================================
# Indexing Settings
# =================================================================
max_commits: int = 500 # Maximum git commits to index (0 = unlimited)
# =================================================================
# Session Settings
# =================================================================
auto_save_sessions: bool = True
auto_load_on_startup: bool = True
session_timeout_minutes: int = 60
# =================================================================
# API Keys (optional, for LLM features)
# =================================================================
anthropic_api_key: str | None = None
openai_api_key: str | None = None
# Default AI model for summaries
default_ai_model: str = "claude" # or "openai"
claude_model: str = "claude-3-sonnet-20240229"
openai_model: str = "gpt-3.5-turbo"
# =================================================================
# MCP Server Settings
# =================================================================
mcp_enabled: bool = True
mcp_host: str = "127.0.0.1"
mcp_port: int = 8003
mcp_sse_path: str = "/sse" # SSE endpoint path (FastMCP standard)
mcp_message_path: str = "/messages/" # Message endpoint path
# =================================================================
# Logging
# =================================================================
log_level: str = "INFO"
log_file: str | None = None
# =================================================================
# Development
# =================================================================
debug: bool = False
test_mode: bool = False
# Top 20 programming languages + extras
supported_extensions: list[str] = Field(
default=[
# Top 20 programming languages
".py", # 1. Python
".js", # 2. JavaScript
".ts", # 3. TypeScript
".java", # 4. Java
".cpp",
".cc",
".cxx",
".hpp", # 5. C++
".c",
".h", # 6. C
".cs", # 7. C#
".go", # 8. Go
".rs", # 9. Rust
".php", # 10. PHP
".rb", # 11. Ruby
".swift", # 12. Swift
".kt",
".kts", # 13. Kotlin
".scala", # 14. Scala
".r",
".R", # 15. R
".m",
".mm", # 16. Objective-C / MATLAB
".pl",
".pm", # 17. Perl
".lua", # 18. Lua
".hs",
".lhs", # 19. Haskell
".ex",
".exs", # 20. Elixir
# Additional languages
".dart", # Dart
".jl", # Julia
".clj",
".cljs", # Clojure
".erl",
".hrl", # Erlang
".fs",
".fsx", # F#
".v", # V / Verilog
".zig", # Zig
".nim", # Nim
".cr", # Crystal
".groovy", # Groovy
# Web
".html",
".htm",
".css",
".scss",
".sass",
".less",
".jsx",
".tsx",
".vue",
".svelte",
# Config/Data
".json",
".yaml",
".yml",
".toml",
".xml",
".ini",
".conf",
".env",
# Documentation
".md",
".rst",
".txt",
".tex",
".org",
".adoc",
# Shell
".sh",
".bash",
".zsh",
".fish",
".ps1",
".bat",
".cmd",
# Database
".sql",
".prisma",
".graphql",
# DevOps
".dockerfile",
".tf",
".hcl",
# Other
".proto",
".thrift",
]
)
ignored_directories: list[str] = Field(
default=[
".git",
".svn",
".hg",
"__pycache__",
".pytest_cache",
".mypy_cache",
".ruff_cache",
"node_modules",
"vendor",
"venv",
".venv",
"env",
"build",
"dist",
"target",
"out",
"bin",
"obj",
".idea",
".vscode",
".vs",
"coverage",
".coverage",
"htmlcov",
".next",
".nuxt",
".output",
".contextfs", # Don't index our own data
]
)
class Config:
env_prefix = "CONTEXTFS_"
env_nested_delimiter = "__"
# Global config instance
_config: Config | None = None
def get_config() -> Config:
"""Get global config instance."""
global _config
if _config is None:
_config = Config()
return _config
def set_config(config: Config) -> None:
"""Set global config instance."""
global _config
_config = config
def reset_config() -> None:
"""Reset global config instance (useful for tests)."""
global _config
_config = None

Xet Storage Details

Size:
9.88 kB
·
Xet hash:
cb2b425916232de6ccd7014ba74eba938780d815c410846bf01df9f3bce5e9c3

Xet efficiently stores files, intelligently splitting them into unique chunks and accelerating uploads and downloads. More info.