mullama 0.3.0

Comprehensive Rust bindings for llama.cpp with memory-safe API and advanced features
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152

//! # Mullama Daemon
//!
//! A multi-model daemon for serving LLMs with IPC and OpenAI-compatible HTTP API.
//!
//! ## Architecture
//!
//! ```text
//!                                    ┌──────────────────────────────────┐
//!                                    │           Daemon                 │
//! ┌─────────────┐                    │  ┌────────────────────────────┐  │
//! │  TUI Client │◄── nng (IPC) ─────►│  │     Model Manager          │  │
//! └─────────────┘                    │  │  ┌───────┐  ┌───────┐      │  │
//!                                    │  │  │Model 1│  │Model 2│ ...  │  │
//! ┌─────────────┐                    │  │  └───────┘  └───────┘      │  │
//! │   curl/app  │◄── HTTP/REST ─────►│  └────────────────────────────┘  │
//! └─────────────┘   (OpenAI API)     │                                  │
//!                                    │  Endpoints:                      │
//! ┌─────────────┐                    │  • /v1/chat/completions          │
//! │ Other Client│◄── nng (IPC) ─────►│  • /v1/completions               │
//! └─────────────┘                    │  • /v1/models                    │
//!                                    │  • /v1/embeddings                │
//!                                    └──────────────────────────────────┘
//! ```
//!
//! ## Usage
//!
//! ```bash
//! # Start the daemon
//! mullama serve --model llama:./models/llama.gguf --model mistral:./models/mistral.gguf
//!
//! # Interactive TUI chat
//! mullama chat
//!
//! # One-shot generation
//! mullama run "What is the meaning of life?"
//!
//! # Use with curl (OpenAI compatible)
//! curl http://localhost:8080/v1/chat/completions \
//!   -H "Content-Type: application/json" \
//!   -d '{"model": "llama", "messages": [{"role": "user", "content": "Hello!"}]}'
//! ```

mod anthropic;
mod client;
pub mod defaults;
mod hf;
mod models;
pub mod ollama;
pub mod ollama_api;
pub mod ollama_template;
mod openai;
mod protocol;
pub mod provider;
pub mod registry;
mod server;
pub mod spawn;
pub mod store;
mod tui;
pub mod ui;

// Protocol types (IPC)
pub use protocol::{
    format_size, generate_completion_id, ChatChoice as IpcChatChoice, ChatCompletionParams,
    ChatCompletionResponse as IpcChatCompletionResponse, ChatMessage,
    CompletionChoice as IpcCompletionChoice, CompletionParams,
    CompletionResponse as IpcCompletionResponse, DaemonStats, DaemonStatus, EmbeddingInput,
    ErrorCode, MessageContent, ModelDetailedStats, ModelInfo, ModelLoadParams, ModelStatus, Request, Response,
    Usage,
};

// Model management
pub use models::{
    estimate_model_memory, LoadedModel, MemoryEstimate, ModelConfig, ModelLoadConfig, ModelManager,
    ModelStats, RequestGuard, DEFAULT_CONTEXT_POOL_SIZE,
};

// Server
pub use server::{
    Daemon, DaemonBuilder, DaemonConfig, EvictionPolicy, HttpConfig, ModelDefaultsConfig,
    ResourceConfig,
};

// Persistent store
pub use store::{DaemonStore, StorageBackend};

// Model provider trait
pub use provider::{ModelProvider, ResolvedModelPath};

// OpenAI-compatible HTTP API types
pub use openai::{
    create_openai_router, ApiError, AppState, ChatChoice, ChatCompletionRequest,
    ChatCompletionResponse, CompletionChoice, CompletionRequest, CompletionResponse,
    EmbeddingObject, EmbeddingsRequest, EmbeddingsResponse, ErrorDetail, ErrorResponse,
    ModelObject, ModelsResponse,
};

// Client
pub use client::{ChatResult, CompletionResult, DaemonClient};

// TUI
pub use tui::TuiApp;

// HuggingFace downloader
pub use hf::{
    resolve_model_path, CachedModel, GgufFileInfo, HfDownloader, HfModelSpec, HfSearchResult,
};

// Ollama registry client
pub use ollama::{
    OllamaClient, OllamaManifest, OllamaModel, OllamaModelIndex, OllamaModelRef, OllamaParameters,
};

// Ollama template conversion
pub use ollama_template::ChatTemplate;

// Model registry (aliases)
pub use registry::{
    registry, resolve_model_name, ModelAlias, ModelRegistry, ParsedModelSpec, RegistryError,
    ResolvedModel,
};

// Anthropic API types
pub use anthropic::{
    AnthropicMessage, AnthropicUsage, ContentBlock, MessageContent as AnthropicMessageContent,
    MessagesRequest, MessagesResponse, ResponseContentBlock,
};

// Daemon spawn utilities
pub use spawn::{
    daemon_status, ensure_daemon_running, is_daemon_running, spawn_daemon, stop_daemon, DaemonInfo,
    SpawnConfig, SpawnResult,
};

// Embedded Web UI
pub use ui::{serve_ui, ui_available};

// Default models
pub use defaults::{
    get_default, list_default_infos, list_defaults, DefaultModel, DefaultModelInfo,
};

/// Default IPC socket path
#[cfg(unix)]
pub const DEFAULT_SOCKET: &str = "ipc:///tmp/mullama.sock";
#[cfg(windows)]
pub const DEFAULT_SOCKET: &str = "ipc://mullama";

/// Default HTTP port for OpenAI API
pub const DEFAULT_HTTP_PORT: u16 = 8080;

/// Default model alias when none specified
pub const DEFAULT_MODEL: &str = "default";