Renamed internal structures.

README.md

@@ -42,7 +42,7 @@ features[3] = -0.221184;
 
 svm.predict_value(&mut problem)?;
 
-assert_eq!(problem.result(), SVMResult::Label(42));
+assert_eq!(problem.result(), Outcome::Label(42));
 ```
 
 From C / FFI:

examples/basic.rs

@@ -3,7 +3,7 @@
 use ffsvm::*;
 use std::convert::TryFrom;
 
-fn main() -> Result<(), SVMError> {
+fn main() -> Result<(), Error> {
     let svm = DenseSVM::try_from(SAMPLE_MODEL)?;
 
     let mut problem = Problem::from(&svm);
@@ -16,7 +16,7 @@ fn main() -> Result<(), SVMError> {
 
     svm.predict_value(&mut problem)?;
 
-    assert_eq!(problem.result(), SVMResult::Label(12));
+    assert_eq!(problem.result(), Outcome::Label(12));
 
     Ok(())
 }

src/errors.rs

@@ -5,7 +5,7 @@ use std::{
 
 /// Possible error types when classifying with a [SVMCore].
 #[derive(Debug)]
-pub enum SVMError {
+pub enum Error {
     /// This can be emitted when creating a [SVM] from a [ModelFile]. For models generated by
     /// libSVM's `svm-train`, the most common reason this occurs is skipping attributes.
     /// All attributes must be in sequential order 0, 1, 2, ..., n. If they are not, this
@@ -44,24 +44,24 @@ pub enum SVMError {
     ParsingError(String),
 }
 
-// impl<'a, T> From<Error<'a, T>> for SVMError {
+// impl<'a, T> From<Error<'a, T>> for Error {
 //     fn from(_: Error<'a, T>) -> Self {
-//         SVMError::ParsingError
+//         Error::ParsingError
 //     }
 // }
 
-impl<'a> From<pest::Error<'a, crate::parser::Rule>> for SVMError {
-    fn from(e: pest::Error<'a, crate::parser::Rule>) -> Self { SVMError::ParsingError(format!("{}", e)) }
+impl<'a> From<pest::Error<'a, crate::parser::Rule>> for Error {
+    fn from(e: pest::Error<'a, crate::parser::Rule>) -> Self { Error::ParsingError(format!("{}", e)) }
 }
 
-impl From<NoneError> for SVMError {
-    fn from(_: NoneError) -> Self { SVMError::ParsingError("NoneError".to_owned()) }
+impl From<NoneError> for Error {
+    fn from(_: NoneError) -> Self { Error::ParsingError("NoneError".to_owned()) }
 }
 
-impl From<ParseFloatError> for SVMError {
-    fn from(_e: ParseFloatError) -> Self { SVMError::ParsingError("ParseFloatError".to_owned()) }
+impl From<ParseFloatError> for Error {
+    fn from(_e: ParseFloatError) -> Self { Error::ParsingError("ParseFloatError".to_owned()) }
 }
 
-impl From<ParseIntError> for SVMError {
-    fn from(_: ParseIntError) -> Self { SVMError::ParsingError("ParseIntError".to_owned()) }
+impl From<ParseIntError> for Error {
+    fn from(_: ParseIntError) -> Self { Error::ParsingError("ParseIntError".to_owned()) }
 }

src/lib.rs

@@ -46,7 +46,7 @@
 //! use ffsvm::*;
 //! use std::convert::TryFrom;
 //!
-//! fn main() -> Result<(), SVMError> {
+//! fn main() -> Result<(), Error> {
 //!     // Replace `SAMPLE_MODEL` with a `&str` to your model.
 //!     let svm = DenseSVM::try_from(SAMPLE_MODEL)?;
 //!
@@ -60,7 +60,7 @@
 //!
 //!     svm.predict_value(&mut problem)?;
 //!
-//!     assert_eq!(problem.result(), SVMResult::Label(42));
+//!     assert_eq!(problem.result(), Outcome::Label(42));
 //!
 //!     Ok(())
 //! }
@@ -82,13 +82,13 @@ mod vectors;
 pub static SAMPLE_MODEL: &str = include_str!("sample.model");
 
 pub use crate::{
-    errors::SVMError,
+    errors::Error,
     parser::ModelFile,
     svm::{
         core::SVMCore,
         kernel::{KernelDense, Linear, Poly, Rbf, Sigmoid},
         predict::Predict,
-        problem::{Problem, SVMResult},
+        problem::{Outcome, Problem},
         DenseSVM, SVMType, SparseSVM,
     },
 };

src/parser/mod.rs

@@ -7,7 +7,7 @@ pub use self::raw::*;
 use pest::Parser;
 use pest_derive::Parser;
 
-use crate::errors::SVMError;
+use crate::errors::Error;
 
 // Hack to make `pest` re-generate parser every time file changes.
 #[cfg(debug_assertions)]
@@ -43,10 +43,10 @@ macro_rules! convert {
 }
 
 impl<'a> TryFrom<&'a str> for ModelFile<'a> {
-    type Error = SVMError;
+    type Error = Error;
 
     /// Parses a string into a SVM model
-    fn try_from(input: &str) -> Result<ModelFile<'_>, SVMError> {
+    fn try_from(input: &str) -> Result<ModelFile<'_>, Error> {
         let parsed = LibSVMModel::parse(Rule::file, input)?.next()?;
 
         let mut svm_type = Option::None;

src/svm/core/dense.rs

@@ -2,14 +2,14 @@ use simd_aligned::{f32s, f64s, RowOptimized, SimdMatrix, SimdVector};
 use std::{convert::TryFrom, marker::PhantomData};
 
 use crate::{
-    errors::SVMError,
+    errors::Error,
     parser::ModelFile,
     svm::{
         class::Class,
         core::SVMCore,
         kernel::{KernelDense, Linear, Poly, Rbf, Sigmoid},
         predict::Predict,
-        problem::{Problem, SVMResult},
+        problem::{Problem, Outcome},
         Probabilities, SVMType,
     },
     util::{find_max_index, set_all, sigmoid_predict},
@@ -36,7 +36,7 @@ impl SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, Simd
     // based on Method 2 from the paper "Probability Estimates for Multi-class
     // Classification by Pairwise Coupling", Journal of Machine Learning Research 5 (2004) 975-1005,
     // by Ting-Fan Wu, Chih-Jen Lin and Ruby C. Weng.
-    crate fn compute_multiclass_probabilities(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), SVMError> {
+    crate fn compute_multiclass_probabilities(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), Error> {
         let num_classes = self.classes.len();
         let max_iter = 100.max(num_classes);
         let mut q = problem.q.flat_mut();
@@ -96,7 +96,7 @@ impl SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, Simd
             // In case we are on the last iteration round past the threshold
             // we know something went wrong. Signal we exceeded the threshold.
             if i == max_iter {
-                return Err(SVMError::IterationsExceeded);
+                return Err(Error::IterationsExceeded);
             }
 
             // This seems to be the main function performing (23) and (24).
@@ -170,19 +170,19 @@ impl SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, Simd
 
         sum -= self.rho[0];
 
-        problem.result = SVMResult::Value(sum as f32);
+        problem.result = Outcome::Value(sum as f32);
     }
 }
 
 impl Predict<SimdVector<f32s>, SimdVector<f64s>> for SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>> {
-    fn predict_probability(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), SVMError> {
+    fn predict_probability(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), Error> {
         match self.svm_type {
             SVMType::CSvc | SVMType::NuSvc => {
                 const MIN_PROB: f64 = 1e-7;
 
                 // Ensure we have probabilities set. If not, somebody used us the wrong way
                 if self.probabilities.is_none() {
-                    return Err(SVMError::NoProbabilities);
+                    return Err(Error::NoProbabilities);
                 }
 
                 let num_classes = self.classes.len();
@@ -217,7 +217,7 @@ impl Predict<SimdVector<f32s>, SimdVector<f64s>> for SVMCore<KernelDense, SimdMa
                 }
 
                 let max_index = find_max_index(problem.probabilities.flat());
-                problem.result = SVMResult::Label(self.classes[max_index].label);
+                problem.result = Outcome::Label(self.classes[max_index].label);
 
                 Ok(())
             }
@@ -227,7 +227,7 @@ impl Predict<SimdVector<f32s>, SimdVector<f64s>> for SVMCore<KernelDense, SimdMa
     }
 
     // Predict the value for one problem.
-    fn predict_value(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), SVMError> {
+    fn predict_value(&self, problem: &mut Problem<SimdVector<f32s>>) -> Result<(), Error> {
         match self.svm_type {
             SVMType::CSvc | SVMType::NuSvc => {
                 // Compute kernel, decision values and eventually the label
@@ -236,7 +236,7 @@ impl Predict<SimdVector<f32s>, SimdVector<f64s>> for SVMCore<KernelDense, SimdMa
 
                 // Compute highest vote
                 let highest_vote = find_max_index(&problem.vote);
-                problem.result = SVMResult::Label(self.classes[highest_vote].label);
+                problem.result = Outcome::Label(self.classes[highest_vote].label);
 
                 Ok(())
             }
@@ -250,18 +250,18 @@ impl Predict<SimdVector<f32s>, SimdVector<f64s>> for SVMCore<KernelDense, SimdMa
 }
 
 impl<'a, 'b> TryFrom<&'a str> for SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>> {
-    type Error = SVMError;
+    type Error = Error;
 
-    fn try_from(input: &'a str) -> Result<SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>>, SVMError> {
+    fn try_from(input: &'a str) -> Result<SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>>, Error> {
         let raw_model = ModelFile::try_from(input)?;
         Self::try_from(&raw_model)
     }
 }
 
 impl<'a, 'b> TryFrom<&'a ModelFile<'b>> for SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>> {
-    type Error = SVMError;
+    type Error = Error;
 
-    fn try_from(raw_model: &'a ModelFile) -> Result<SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>>, SVMError> {
+    fn try_from(raw_model: &'a ModelFile) -> Result<SVMCore<KernelDense, SimdMatrix<f32s, RowOptimized>, SimdVector<f32s>, SimdVector<f64s>>, Error> {
         // To quickly check what broke again during parsing ...
         // println!("{:?}", raw_model);
 
@@ -361,7 +361,7 @@ impl<'a, 'b> TryFrom<&'a ModelFile<'b>> for SVMCore<KernelDense, SimdMatrix<f32s
                         // In case we have seen an attribute already, this one must be strictly
                         // the successor attribute
                         if attribute.index != last + 1 {
-                            return Result::Err(SVMError::AttributesUnordered {
+                            return Result::Err(Error::AttributesUnordered {
                                 index: attribute.index,
                                 value: attribute.value,
                                 last_index: last,

src/svm/core/mod.rs

@@ -5,13 +5,13 @@ use simd_aligned::{f32s, f64s, RowOptimized, SimdMatrix, SimdVector};
 use std::{convert::TryFrom, marker::PhantomData};
 
 use crate::{
-    errors::SVMError,
+    errors::Error,
     parser::ModelFile,
     svm::{
         class::Class,
         kernel::{KernelDense, Linear, Poly, Rbf, Sigmoid},
         predict::Predict,
-        problem::{Problem, SVMResult},
+        problem::{Outcome, Problem},
         Probabilities, SVMType,
     },
     util::{find_max_index, set_all, sigmoid_predict},

src/svm/core/sparse.rs

@@ -3,14 +3,14 @@ use crate::sparse::{SparseMatrix, SparseVector};
 use std::{convert::TryFrom, marker::PhantomData};
 
 use crate::{
-    errors::SVMError,
+    errors::Error,
     parser::ModelFile,
     svm::{
         class::Class,
         core::SVMCore,
         kernel::{KernelSparse, Linear, Poly, Rbf, Sigmoid},
         predict::Predict,
-        problem::{Problem, SVMResult},
+        problem::{Outcome, Problem},
         Probabilities, SVMType,
     },
     util::{find_max_index, set_all, sigmoid_predict},
@@ -37,7 +37,7 @@ impl SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f6
     // based on Method 2 from the paper "Probability Estimates for Multi-class
     // Classification by Pairwise Coupling", Journal of Machine Learning Research 5 (2004) 975-1005,
     // by Ting-Fan Wu, Chih-Jen Lin and Ruby C. Weng.
-    crate fn compute_multiclass_probabilities(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), SVMError> {
+    crate fn compute_multiclass_probabilities(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), Error> {
         let num_classes = self.classes.len();
         let max_iter = 100.max(num_classes);
         let mut q = problem.q.flat_mut();
@@ -97,7 +97,7 @@ impl SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f6
             // In case we are on the last iteration round past the threshold
             // we know something went wrong. Signal we exceeded the threshold.
             if i == max_iter {
-                return Err(SVMError::IterationsExceeded);
+                return Err(Error::IterationsExceeded);
             }
 
             // This seems to be the main function performing (23) and (24).
@@ -171,19 +171,19 @@ impl SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f6
 
         sum -= self.rho[0];
 
-        problem.result = SVMResult::Value(sum as f32);
+        problem.result = Outcome::Value(sum as f32);
     }
 }
 
 impl Predict<SparseVector<f32>, SparseVector<f64>> for SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>> {
-    fn predict_probability(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), SVMError> {
+    fn predict_probability(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), Error> {
         match self.svm_type {
             SVMType::CSvc | SVMType::NuSvc => {
                 const MIN_PROB: f64 = 1e-7;
 
                 // Ensure we have probabilities set. If not, somebody used us the wrong way
                 if self.probabilities.is_none() {
-                    return Err(SVMError::NoProbabilities);
+                    return Err(Error::NoProbabilities);
                 }
 
                 let num_classes = self.classes.len();
@@ -218,7 +218,7 @@ impl Predict<SparseVector<f32>, SparseVector<f64>> for SVMCore<KernelSparse, Spa
                 }
 
                 let max_index = find_max_index(problem.probabilities.flat());
-                problem.result = SVMResult::Label(self.classes[max_index].label);
+                problem.result = Outcome::Label(self.classes[max_index].label);
 
                 Ok(())
             }
@@ -228,7 +228,7 @@ impl Predict<SparseVector<f32>, SparseVector<f64>> for SVMCore<KernelSparse, Spa
     }
 
     // Predict the value for one problem.
-    fn predict_value(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), SVMError> {
+    fn predict_value(&self, problem: &mut Problem<SparseVector<f32>>) -> Result<(), Error> {
         match self.svm_type {
             SVMType::CSvc | SVMType::NuSvc => {
                 // Compute kernel, decision values and eventually the label
@@ -237,7 +237,7 @@ impl Predict<SparseVector<f32>, SparseVector<f64>> for SVMCore<KernelSparse, Spa
 
                 // Compute highest vote
                 let highest_vote = find_max_index(&problem.vote);
-                problem.result = SVMResult::Label(self.classes[highest_vote].label);
+                problem.result = Outcome::Label(self.classes[highest_vote].label);
 
                 Ok(())
             }
@@ -251,18 +251,18 @@ impl Predict<SparseVector<f32>, SparseVector<f64>> for SVMCore<KernelSparse, Spa
 }
 
 impl<'a, 'b> TryFrom<&'a str> for SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>> {
-    type Error = SVMError;
+    type Error = Error;
 
-    fn try_from(input: &'a str) -> Result<SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>>, SVMError> {
+    fn try_from(input: &'a str) -> Result<SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>>, Error> {
         let raw_model = ModelFile::try_from(input)?;
         Self::try_from(&raw_model)
     }
 }
 
 impl<'a, 'b> TryFrom<&'a ModelFile<'b>> for SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>> {
-    type Error = SVMError;
+    type Error = Error;
 
-    fn try_from(raw_model: &'a ModelFile) -> Result<SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>>, SVMError> {
+    fn try_from(raw_model: &'a ModelFile) -> Result<SVMCore<KernelSparse, SparseMatrix<f32>, SparseVector<f32>, SparseVector<f64>>, Error> {
         // To quickly check what broke again during parsing ...
         // println!("{:?}", raw_model);
 

src/svm/kernel/poly.rs

@@ -2,7 +2,7 @@ use std::convert::{From, TryFrom};
 
 use super::{KernelDense, KernelSparse};
 use crate::{
-    errors::SVMError,
+    errors::Error,
     parser::ModelFile,
     sparse::{SparseMatrix, SparseVector},
 };
@@ -60,12 +60,12 @@ impl KernelSparse for Poly {
 }
 
 impl<'a, 'b> TryFrom<&'a ModelFile<'b>> for Poly {
-    type Error = SVMError;
+    type Error = Error;
 
-    fn try_from(raw_model: &'a ModelFile<'b>) -> Result<Poly, SVMError> {
-        let gamma = raw_model.header.gamma.ok_or(SVMError::NoGamma)?;
-        let coef0 = raw_model.header.coef0.ok_or(SVMError::NoCoef0)?;
-        let degree = raw_model.header.degree.ok_or(SVMError::NoDegree)?;
+    fn try_from(raw_model: &'a ModelFile<'b>) -> Result<Poly, Error> {
+        let gamma = raw_model.header.gamma.ok_or(Error::NoGamma)?;
+        let coef0 = raw_model.header.coef0.ok_or(Error::NoCoef0)?;
+        let degree = raw_model.header.degree.ok_or(Error::NoDegree)?;
 
         Ok(Poly { gamma, coef0, degree })
     }