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
// #![cfg_attr(not(feature = "std"), no_std)]
//! # Tracel SDK
//!
//! High-level Tracel SDK.
//!
//! This crate re-exports the main crates used to build training, experiment tracking, inference,
//! and fleet workflows on top of Tracel.
//!
//! Features:
//! - Artifact creation and management
//! - Experiment tracking
//! - Logging metrics
//! - Model versioning
//!
//! ## Crate Layout
//!
//! The most commonly used re-exports are:
//! - [`experiment`]: experiment runs, logging, artifacts, and Burn learner integrations.
//! - [`app`]: job registration, plus CLI and HTTP server front-ends to run those jobs.
//! - [`console`]: organizations, projects, and users on the Tracel console.
//! - [`datasets`]: the dataset domain and the capability over it.
//! - [`models`]: the model domain and the capability over it.
//! - [`artifact`]: bundle and artifact utilities.
//!
//! ## Registering Routines
//!
//! Wrap a routine into a job with [`experiment::ExperimentModule::create`], then register it
//! with a [`app::cli::Cli`] (to dispatch from the command line) or a `app::server::Server`
//! (to dispatch over HTTP, requires the `server` feature):
//!
//! ```ignore
//! use tracel::app::cli::Cli;
//! use tracel::app::cli::mapper::JsonMapper;
//! use tracel::experiment::ExperimentRun;
//! use tracel::{Connection, Context};
//!
//! fn main() -> anyhow::Result<()> {
//! let module = Context::new(Connection::Cloud)?.experiment();
//!
//! let job = module.create("my_training_procedure", |session: &ExperimentRun, config| {
//! // Your training code here
//! my_training_function(session, config)
//! });
//!
//! Cli::new()
//! .register(job, JsonMapper::with_default(MyConfig::default()))
//! .run()?;
//!
//! Ok(())
//! }
//! ```
//!
//! The job's name (`"my_training_procedure"` above) is what callers use to select it, whether
//! from the CLI or from an HTTP request path.
/// Experiment tracking and management.
/// The Tracel console: its organizations, projects, and users.
pub use tracel_console as console;
/// A Tracel Station: experiments, models, and datasets served by one Station URL (requires the
/// `station` feature).
pub use tracel_station as station;
/// The dataset domain the console or a Station serves, and the operations over it.
pub use tracel_datasets as datasets;
/// The model domain the console or a Station serves, and the operations over it.
pub use tracel_models as models;
/// Inference contracts and adapters.
pub use tracel_inference as inference;
/// Artifact bundle utilities and adapters.
pub use tracel_artifact as artifact;
/// App module for job registration, CLI, and config mappers
pub use tracel_app as app;
/// Station runner front-end: serve registered jobs to a Tracel Station job queue (requires the
/// `runner` feature).
pub use tracel_runner as runner;
pub use Connection;
pub use Context;
pub use ContextError;