kalosm_language_model/chat/
task.rs

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
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
use std::mem::MaybeUninit;
use std::ops::Deref;

use crate::ModelConstraints;
use crate::NoConstraints;

use super::Chat;
use super::ChatMessage;
use super::ChatResponseBuilder;
use super::CreateChatSession;
use super::CreateDefaultChatConstraintsForType;
use super::MessageType;

/// A task session lets you efficiently run a task with a model. The task session will reuse the model's cache to avoid re-feeding the task prompt repeatedly.
///
/// # Example
/// ```rust, no_run
/// use kalosm::language::*;
///
/// #[tokio::main]
/// async fn main() {
///     let mut llm = Llama::new_chat().await.unwrap();
///     let mut task = llm.task("You are a math assistant who helps students with their homework. You solve equations and answer questions. When solving problems, you will always solve problems step by step.");
///
///     println!("question 1");
///     // The first time we use the task, it will load the model and prompt.
///     task.run("What is 2 + 2?")
///         .to_std_out()
///         .await
///         .unwrap();
///     
///     println!("question 2");
///     // After the first time, the model and prompt are cached.
///     task.run("What is 4 + 4?")
///         .to_std_out()
///         .await
///         .unwrap();
/// }
/// ```
#[derive(Debug)]
pub struct Task<M: CreateChatSession, Constraints = NoConstraints> {
    chat: Chat<M>,
    constraints: Constraints,
}

impl<M: CreateChatSession, Constraints: Clone> Clone for Task<M, Constraints> {
    fn clone(&self) -> Self {
        Self {
            chat: self.chat.clone(),
            constraints: self.constraints.clone(),
        }
    }
}

impl<M: CreateChatSession> Task<M> {
    /// Create a new task with no constraints and the default sampler.
    pub fn new(model: M, description: impl ToString) -> Self {
        let chat = Chat::new(model).with_system_prompt(description);
        Self {
            chat,
            constraints: NoConstraints,
        }
    }
}

impl<M: CreateChatSession, Constraints> Task<M, Constraints> {
    /// Add an example to the task. Examples help the model perform better by allowing it to mimic the format of the examples.
    ///
    /// # Example
    /// ```rust, no_run
    /// use kalosm::language::*;
    ///
    /// #[tokio::main]
    /// async fn main() {
    ///     let model = Llama::new_chat().await.unwrap();
    ///     let task = model.task("You are a math assistant who helps students with their homework. You solve equations and answer questions. When solving problems, you will always solve problems step by step.")
    ///         .with_example("What is 1 + 2?", "Step 1: 1 + 2 = 3\nOutput: 3");
    ///     let mut stream = task("What is 2 + 2?");
    ///     stream.to_std_out().await.unwrap();
    /// }
    /// ```
    pub fn with_example(mut self, input: impl ToString, output: impl ToString) -> Self {
        self.chat
            .add_message(ChatMessage::new(MessageType::UserMessage, input));
        self.chat
            .add_message(ChatMessage::new(MessageType::ModelAnswer, output));
        self
    }

    /// Add multiple examples to the task. Examples help the model perform better by allowing it to mimic the format of the examples.
    ///
    /// # Example
    /// ```rust, no_run
    /// use kalosm::language::*;
    ///
    /// #[tokio::main]
    /// async fn main() {
    ///     let model = Llama::new_chat().await.unwrap();
    ///     let task = model.task("You are a math assistant who helps students with their homework. You solve equations and answer questions. When solving problems, you will always solve problems step by step.")
    ///         .with_examples([
    ///             ("What is 1 + 2?", "Step 1: 1 + 2 = 3\nOutput: 3"),
    ///             ("What is 3 + 4?", "Step 1: 3 + 4 = 7\nOutput: 7"),
    ///             ("What is (4 + 8) / 3?", "Step 1: 4 + 8 = 12\nStep 2: 12 / 3 = 4\nOutput: 4"),
    ///         ]);
    ///     let mut stream = task("What is 3 + 4?");
    ///     stream.to_std_out().await.unwrap();
    /// }
    /// ```
    pub fn with_examples(
        mut self,
        examples: impl IntoIterator<Item = (impl ToString, impl ToString)>,
    ) -> Self {
        for (input, output) in examples {
            self = self.with_example(input, output);
        }
        self
    }

    /// Set the constraints for the task. The constraints force the format of all outputs of the task to fit
    /// the constraints. This can be used to make the model return a specific type. This method does the same thing
    /// as [`ChatResponseBuilder::with_constraints`] except it is called once on the task instead of any time you
    /// run the task.
    ///
    /// # Example
    /// ```rust, no_run
    /// use kalosm::language::*;
    /// use std::sync::Arc;
    ///
    /// #[tokio::main]
    /// async fn main() {
    ///     let model = Llama::new_chat().await.unwrap();
    ///     let task = model
    ///         .task("You are a math assistant. Respond with just the number answer and nothing else.")
    ///         .with_constraints(Arc::new(i32::new_parser()));
    ///     let mut stream = task("What is 2 + 2?");
    ///     stream.to_std_out().await.unwrap();
    ///     let result: i32 = stream.await.unwrap();
    ///     println!("{result}");
    /// }
    /// ```
    pub fn with_constraints<NewConstraints>(
        self,
        constraints: NewConstraints,
    ) -> Task<M, NewConstraints> {
        Task {
            chat: self.chat,
            constraints,
        }
    }

    /// Create a task with the default constraints for the given type. This is the same as calling [`Task::with_constraints`] with the default constraints for the given type.
    ///
    /// # Example
    /// ```rust, no_run
    /// use kalosm::language::*;
    ///
    /// #[tokio::main]
    /// async fn main() {
    ///     let model = Llama::new_chat().await.unwrap();
    ///     let task = model
    ///         .task("You are a math assistant. Respond with just the number answer and nothing else.")
    ///         .typed();
    ///     let mut stream = task("What is 2 + 2?");
    ///     stream.to_std_out().await.unwrap();
    ///     let result: i32 = stream.await.unwrap();
    ///     println!("{result}");
    /// }
    /// ```
    pub fn typed<T>(
        self,
    ) -> Task<M, <M as CreateDefaultChatConstraintsForType<T>>::DefaultConstraints>
    where
        M: CreateDefaultChatConstraintsForType<T>,
    {
        self.with_constraints(M::create_default_constraints())
    }
}

impl<M: CreateChatSession, Constraints: Clone> Task<M, Constraints> {
    /// Run the task with a message.
    ///
    /// # Example
    /// ```rust, no_run
    /// use kalosm::language::*;
    ///
    /// #[tokio::main]
    /// async fn main() {
    ///     let mut llm = Llama::new_chat().await.unwrap();
    ///     let task = llm.task("You are a math assistant who helps students with their homework. You solve equations and answer questions. When solving problems, you will always solve problems step by step.");
    ///
    ///     let result = task("What is 2 + 2?").await.unwrap();
    ///     println!("{result}");
    /// }
    /// ```
    pub fn run(&self, message: impl ToString) -> ChatResponseBuilder<'static, M, Constraints> {
        self.chat
            .clone()
            .into_add_message(message)
            .with_constraints(self.constraints.clone())
    }
}

impl<M: CreateChatSession + 'static, Constraints: ModelConstraints + Clone + 'static> Deref
    for Task<M, Constraints>
{
    type Target = dyn Fn(&str) -> ChatResponseBuilder<'static, M, Constraints>;

    fn deref(&self) -> &Self::Target {
        // https://github.com/dtolnay/case-studies/tree/master/callable-types

        // Create an empty allocation for Self.
        let uninit_callable = MaybeUninit::<Self>::uninit();
        // Move a closure that captures just self into the uninitialized memory. Closures create an anonymous type that implement
        // FnOnce. In this case, the layout of the type should just be Self because self is the only field in the closure type.
        let uninit_closure =
            move |input: &str| Self::run(unsafe { &*uninit_callable.as_ptr() }, input);

        // Make sure the layout of the closure and Self is the same.
        let size_of_closure = std::alloc::Layout::for_value(&uninit_closure);
        assert_eq!(size_of_closure, std::alloc::Layout::new::<Self>());

        // Then cast the lifetime of the closure to the lifetime of &self.
        fn cast_lifetime<'a, T>(_a: &T, b: &'a T) -> &'a T {
            b
        }
        let reference_to_closure = cast_lifetime(
            {
                // The real closure that we will never use.
                &uninit_closure
            },
            #[allow(clippy::missing_transmute_annotations)]
            // We transmute self into a reference to the closure. This is safe because we know that the closure has the same memory layout as Self so &Closure == &Self.
            unsafe {
                std::mem::transmute(self)
            },
        );

        // Cast the closure to a trait object.
        reference_to_closure as &_
    }
}

impl<M: CreateChatSession + 'static> Deref for Task<M> {
    type Target = dyn Fn(&str) -> ChatResponseBuilder<'static, M>;

    fn deref(&self) -> &Self::Target {
        // https://github.com/dtolnay/case-studies/tree/master/callable-types

        // Create an empty allocation for Self.
        let uninit_callable = MaybeUninit::<Self>::uninit();
        // Move a closure that captures just self into the uninitialized memory. Closures create an anonymous type that implement
        // FnOnce. In this case, the layout of the type should just be Self because self is the only field in the closure type.
        let uninit_closure =
            move |input: &str| Self::run(unsafe { &*uninit_callable.as_ptr() }, input);

        // Make sure the layout of the closure and Self is the same.
        let size_of_closure = std::alloc::Layout::for_value(&uninit_closure);
        assert_eq!(size_of_closure, std::alloc::Layout::new::<Self>());

        // Then cast the lifetime of the closure to the lifetime of &self.
        fn cast_lifetime<'a, T>(_a: &T, b: &'a T) -> &'a T {
            b
        }
        let reference_to_closure = cast_lifetime(
            {
                // The real closure that we will never use.
                &uninit_closure
            },
            #[allow(clippy::missing_transmute_annotations)]
            // We transmute self into a reference to the closure. This is safe because we know that the closure has the same memory layout as Self so &Closure == &Self.
            unsafe {
                std::mem::transmute(self)
            },
        );

        // Cast the closure to a trait object.
        reference_to_closure as &_
    }
}