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
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
// Copyright 2018-2022 argmin developers
//
// Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
// http://apache.org/licenses/LICENSE-2.0> or the MIT license <LICENSE-MIT or
// http://opensource.org/licenses/MIT>, at your option. This file may not be
// copied, modified, or distributed except according to those terms.

//! # Observers
//!
//! Argmin offers an interface to observe the state of the solver at initialization as well as
//! after every iteration. This includes the parameter vector, gradient, Jacobian, Hessian,
//! iteration number, cost values and many more as well as solver-specific metrics. This interface
//! can be used to implement loggers, send the information to a storage or to plot metrics.
//!
//! The observer [`WriteToFile`](`crate::core::observers::WriteToFile`) saves the parameter vector
//! to disk and as such requires the parameter vector to be serializable. Hence this feature is
//! only available with the `serde1` feature.
//!
//! The observer [`SlogLogger`](`crate::core::observers::SlogLogger`) logs the progress of the
//! optimization to screen or to disk. This requires the `slog-logger` feature. Writing to disk
//! in addition requires the `serde1` feature.
//!
//! For each observer it can be defined how often it will observe the progress of the solver. This
//! is indicated via the enum `ObserverMode` which can be either `Always`, `Never`, `NewBest`
//! (whenever a new best solution is found) or `Every(i)` which means every `i`th iteration.
//!
//! Custom observers can be used as well by implementing the [`crate::core::observers::Observe`]
//! trait.
//!
//! ## Example
//!
//! ```rust
//! # #![allow(unused_imports)]
//! # extern crate argmin;
//! # extern crate argmin_testfunctions;
//! # use argmin::core::{Error, Executor, CostFunction, Gradient, observers::ObserverMode};
//! # #[cfg(feature = "slog-logger")]
//! # use argmin::core::observers::SlogLogger;
//! # #[cfg(feature = "serde1")]
//! # use argmin::core::observers::{WriteToFile, WriteToFileSerializer};
//! # use argmin::solver::gradientdescent::SteepestDescent;
//! # use argmin::solver::linesearch::MoreThuenteLineSearch;
//! # use argmin_testfunctions::{rosenbrock_2d, rosenbrock_2d_derivative};
//! #
//! # struct Rosenbrock {
//! #     a: f64,
//! #     b: f64,
//! # }
//! #
//! # /// Implement `CostFunction` for `Rosenbrock`
//! # impl CostFunction for Rosenbrock {
//! #     /// Type of the parameter vector
//! #     type Param = Vec<f64>;
//! #     /// Type of the return value computed by the cost function
//! #     type Output = f64;
//! #
//! #     /// Apply the cost function to a parameter `p`
//! #     fn cost(&self, p: &Self::Param) -> Result<Self::Output, Error> {
//! #         Ok(rosenbrock_2d(p, 1.0, 100.0))
//! #     }
//! # }
//! #
//! # /// Implement `Gradient` for `Rosenbrock`
//! # impl Gradient for Rosenbrock {
//! #     /// Type of the parameter vector
//! #     type Param = Vec<f64>;
//! #     /// Type of the return value computed by the cost function
//! #     type Gradient = Vec<f64>;
//! #
//! #     /// Compute the gradient at parameter `p`.
//! #     fn gradient(&self, p: &Self::Param) -> Result<Self::Gradient, Error> {
//! #         Ok(rosenbrock_2d_derivative(p, 1.0, 100.0))
//! #     }
//! # }
//! #
//! # fn run() -> Result<(), Error> {
//! #
//! # // Define cost function (must implement `CostFunction` and `Gradient`)
//! # let problem = Rosenbrock { a: 1.0, b: 100.0 };
//! #
//! # // Define initial parameter vector
//! # let init_param: Vec<f64> = vec![-1.2, 1.0];
//! #
//! # // Set up line search
//! # let linesearch = MoreThuenteLineSearch::new();
//! #
//! # // Set up solver
//! # let solver = SteepestDescent::new(linesearch);
//! #
//! // [...]
//!
//! let res = Executor::new(problem, solver)
//!     .configure(|config| config.param(init_param).max_iters(2))
//! # ;
//! # #[cfg(feature = "slog-logger")]
//! # let res = res
//!     // Add an observer which will log all iterations to the terminal (without blocking)
//!     .add_observer(SlogLogger::term_noblock(), ObserverMode::Always)
//! # ;
//! # #[cfg(feature = "serde1")]
//! # #[cfg(feature = "slog-logger")]
//! # let res = res
//!     // Log to file whenever a new best solution is found
//!     .add_observer(SlogLogger::file("solver.log", false)?, ObserverMode::NewBest)
//!     // Write parameter vector to `params/param.arg` every 20th iteration
//!     .add_observer(
//!         WriteToFile::new("params", "param", WriteToFileSerializer::JSON),
//!         ObserverMode::Every(20)
//!     )
//! # ;
//! # let res = res
//!     // run the solver on the defined problem
//!     .run()?;
//!
//! // [...]
//! #     Ok(())
//! # }
//! #
//! # fn main() {
//! #     if let Err(ref e) = run() {
//! #         println!("{}", e);
//! #         std::process::exit(1);
//! #     }
//! # }
//! ```

#[cfg(feature = "serde1")]
pub mod file;
#[cfg(feature = "slog-logger")]
pub mod slog_logger;

#[cfg(feature = "serde1")]
pub use file::*;
#[cfg(feature = "slog-logger")]
pub use slog_logger::*;

use crate::core::{Error, State, KV};
use std::default::Default;
use std::sync::{Arc, Mutex};

/// An interface which every observer is required to implement
///
/// # Example
///
/// ```
/// use argmin::core::{Error, KV, State};
/// use argmin::core::observers::Observe;
///
/// struct MyObserver {}
///
/// impl<I> Observe<I> for MyObserver
/// where
///     // Optional constraint on `I`. The `State` trait, which every state used in argmin needs to
///     // implement, offers a range of methods which can be useful.
///     I: State,
/// {
///     fn observe_init(&mut self, name: &str, kv: &KV) -> Result<(), Error> {
///         // Do something with `name` and/or `kv`
///         // Is executed after initialization of a solver
///         Ok(())
///     }
///
///     fn observe_iter(&mut self, state: &I, kv: &KV) -> Result<(), Error> {
///         // Do something with `state` and/or `kv`
///         // Is executed after each iteration of a solver
///         Ok(())
///     }
/// }
/// ```
pub trait Observe<I> {
    /// Called once after initialization of the solver.
    ///
    /// Has access to the name of the solver via `name` and to a key-value store `kv` with entries
    /// specific for each solver.
    fn observe_init(&mut self, _name: &str, _kv: &KV) -> Result<(), Error> {
        Ok(())
    }

    /// Called at every iteration of the solver
    ///
    /// Has access to the current `state` of the solver (which always implements
    /// [`State`](`crate::core::State`)) and to a key-value store `kv` with entries specific for
    /// each solver.
    fn observe_iter(&mut self, _state: &I, _kv: &KV) -> Result<(), Error> {
        Ok(())
    }
}

type ObserversVec<I> = Vec<(Arc<Mutex<dyn Observe<I>>>, ObserverMode)>;

/// Container for observers.
///
/// This type also implements [`Observe`] and therefore can be used like a single observer.
/// Each observer has an [`ObserverMode`] attached which indicates when the observer will be
/// called.
#[derive(Clone, Default)]
pub struct Observers<I> {
    /// Vector of `Observe`rs with the corresponding `ObserverMode`
    observers: ObserversVec<I>,
}

impl<I> Observers<I> {
    /// Construct a new empty `Observers` instance.
    ///
    /// # Example
    ///
    /// ```
    /// use argmin::core::observers::Observers;
    /// use argmin::core::IterState;
    ///
    /// let observers: Observers<IterState<Vec<f64>, (), (), (), f64>> = Observers::new();
    /// # assert!(observers.is_empty());
    /// ```
    pub fn new() -> Self {
        Observers { observers: vec![] }
    }

    /// Add another observer with a corresponding [`ObserverMode`].
    ///
    /// # Example
    ///
    /// ```
    /// use argmin::core::observers::{Observers, ObserverMode};
    /// # #[cfg(feature = "slog-logger")]
    /// use argmin::core::observers::SlogLogger;
    /// use argmin::core::IterState;
    ///
    /// let mut observers: Observers<IterState<Vec<f64>, (), (), (), f64>> = Observers::new();
    ///
    /// # #[cfg(feature = "slog-logger")]
    /// let logger = SlogLogger::term();
    /// # #[cfg(feature = "slog-logger")]
    /// observers.push(logger, ObserverMode::Always);
    /// # #[cfg(feature = "slog-logger")]
    /// # assert!(!observers.is_empty());
    /// ```
    pub fn push<OBS: Observe<I> + 'static>(
        &mut self,
        observer: OBS,
        mode: ObserverMode,
    ) -> &mut Self {
        self.observers.push((Arc::new(Mutex::new(observer)), mode));
        self
    }

    /// Returns true if there are no observers stored.
    ///
    /// # Example
    ///
    /// ```
    /// use argmin::core::observers::Observers;
    /// use argmin::core::IterState;
    ///
    /// let observers: Observers<IterState<Vec<f64>, (), (), (), f64>> = Observers::new();
    /// assert!(observers.is_empty());
    /// ```
    pub fn is_empty(&self) -> bool {
        self.observers.is_empty()
    }
}

/// Implementing [`Observe`] for [`Observers`] allows to use it like a single observer. In its
/// implementation it will loop over all stored observers, checks if the conditions for observing
/// are met and calls the actual observers if required.
impl<I: State> Observe<I> for Observers<I> {
    /// After initialization of the solver, this loops over all stored observers and calls them.
    fn observe_init(&mut self, name: &str, kv: &KV) -> Result<(), Error> {
        for l in self.observers.iter() {
            l.0.lock().unwrap().observe_init(name, kv)?
        }
        Ok(())
    }

    /// Called after each iteration.
    ///
    /// Loops over all observers, and based on whether the condition for calling the observers are
    /// met, calls them.
    fn observe_iter(&mut self, state: &I, kv: &KV) -> Result<(), Error> {
        for l in self.observers.iter_mut() {
            let iter = state.get_iter();
            let observer = &mut l.0.lock().unwrap();
            match l.1 {
                ObserverMode::Always => observer.observe_iter(state, kv),
                ObserverMode::Every(i) if iter % i == 0 => observer.observe_iter(state, kv),
                ObserverMode::NewBest if state.is_best() => observer.observe_iter(state, kv),
                ObserverMode::Never | ObserverMode::Every(_) | ObserverMode::NewBest => Ok(()),
            }?
        }
        Ok(())
    }
}

/// Indicates when to call an observer.
///
/// `Always` calls the observer in every iteration, `Every(X)` calls the observer every X
/// iterations, `NewBest` calls the observer only when a new best parameter vector is found and
/// `Never` deactivates the observer.
#[derive(Copy, Clone, Debug, Eq, PartialEq)]
pub enum ObserverMode {
    /// Never call the observer
    Never,
    /// Call observer in every iteration
    Always,
    /// Call observer every N iterations
    Every(u64),
    /// Call observer when new best is found
    NewBest,
}

impl Default for ObserverMode {
    /// The default for `ObserverMode` is `Always`
    fn default() -> ObserverMode {
        ObserverMode::Always
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    send_sync_test!(observermode, ObserverMode);

    #[test]
    fn test_observers() {
        use crate::core::observers::Observe;
        use crate::core::{Error, IterState, KV};

        struct TestStor {
            pub solver_name: String,
            pub init_called: usize,
            pub iter_called: usize,
        }

        impl TestStor {
            fn new() -> Arc<Mutex<TestStor>> {
                Arc::new(Mutex::new(TestStor {
                    solver_name: String::new(),
                    init_called: 0,
                    iter_called: 0,
                }))
            }
        }

        struct TestObs {
            data: Arc<Mutex<TestStor>>,
        }

        impl<I> Observe<I> for TestObs {
            fn observe_init(&mut self, name: &str, _kv: &KV) -> Result<(), Error> {
                self.data.lock().unwrap().solver_name = name.into();
                self.data.lock().unwrap().init_called += 1;
                Ok(())
            }

            fn observe_iter(&mut self, _state: &I, _kv: &KV) -> Result<(), Error> {
                self.data.lock().unwrap().iter_called += 1;
                Ok(())
            }
        }

        let test_stor_1 = TestStor::new();
        let test_obs_1 = TestObs {
            data: test_stor_1.clone(),
        };

        let test_stor_2 = TestStor::new();
        let test_obs_2 = TestObs {
            data: test_stor_2.clone(),
        };

        let test_stor_3 = TestStor::new();
        let test_obs_3 = TestObs {
            data: test_stor_3.clone(),
        };

        let test_stor_4 = TestStor::new();
        let test_obs_4 = TestObs {
            data: test_stor_4.clone(),
        };

        let storages = [test_stor_1, test_stor_2, test_stor_3, test_stor_4];

        type TState = IterState<Vec<f64>, (), (), (), f64>;

        let mut obs: Observers<TState> = Observers::new();
        obs.push(test_obs_1, ObserverMode::Never)
            .push(test_obs_2, ObserverMode::Always)
            .push(test_obs_3, ObserverMode::Every(3))
            .push(test_obs_4, ObserverMode::NewBest);

        obs.observe_init("test_solver", &make_kv!()).unwrap();

        // all `init_called` should be 1, all `iter_called` 0
        for s in storages.iter() {
            let observer = s.lock().unwrap();
            assert_eq!(observer.solver_name, "test_solver");
            assert_eq!(observer.init_called, 1);
            assert_eq!(observer.iter_called, 0);
        }

        let mut state: TState = IterState::new();
        obs.observe_iter(&state, &make_kv!()).unwrap();

        assert_eq!(storages[0].lock().unwrap().init_called, 1);
        assert_eq!(storages[0].lock().unwrap().iter_called, 0);
        assert_eq!(storages[1].lock().unwrap().init_called, 1);
        assert_eq!(storages[1].lock().unwrap().iter_called, 1);
        assert_eq!(storages[2].lock().unwrap().init_called, 1);
        assert_eq!(storages[2].lock().unwrap().iter_called, 1);
        assert_eq!(storages[3].lock().unwrap().init_called, 1);
        assert_eq!(storages[3].lock().unwrap().iter_called, 1);

        state.increment_iter();
        obs.observe_iter(&state, &make_kv!()).unwrap();

        assert_eq!(storages[0].lock().unwrap().init_called, 1);
        assert_eq!(storages[0].lock().unwrap().iter_called, 0);
        assert_eq!(storages[1].lock().unwrap().init_called, 1);
        assert_eq!(storages[1].lock().unwrap().iter_called, 2);
        assert_eq!(storages[2].lock().unwrap().init_called, 1);
        assert_eq!(storages[2].lock().unwrap().iter_called, 1);
        assert_eq!(storages[3].lock().unwrap().init_called, 1);
        assert_eq!(storages[3].lock().unwrap().iter_called, 1);

        state.increment_iter();
        state.increment_iter();
        obs.observe_iter(&state, &make_kv!()).unwrap();

        assert_eq!(storages[0].lock().unwrap().init_called, 1);
        assert_eq!(storages[0].lock().unwrap().iter_called, 0);
        assert_eq!(storages[1].lock().unwrap().init_called, 1);
        assert_eq!(storages[1].lock().unwrap().iter_called, 3);
        assert_eq!(storages[2].lock().unwrap().init_called, 1);
        assert_eq!(storages[2].lock().unwrap().iter_called, 2);
        assert_eq!(storages[3].lock().unwrap().init_called, 1);
        assert_eq!(storages[3].lock().unwrap().iter_called, 1);

        state.increment_iter();
        // "new best found"
        state.last_best_iter = state.iter;
        obs.observe_iter(&state, &make_kv!()).unwrap();

        assert_eq!(storages[0].lock().unwrap().init_called, 1);
        assert_eq!(storages[0].lock().unwrap().iter_called, 0);
        assert_eq!(storages[1].lock().unwrap().init_called, 1);
        assert_eq!(storages[1].lock().unwrap().iter_called, 4);
        assert_eq!(storages[2].lock().unwrap().init_called, 1);
        assert_eq!(storages[2].lock().unwrap().iter_called, 2);
        assert_eq!(storages[3].lock().unwrap().init_called, 1);
        assert_eq!(storages[3].lock().unwrap().iter_called, 2);
    }
}