Set up SGF reading and start on the game database #47

Merged
savanni merged 20 commits from kifu/sgf into main 2023-07-26 13:54:28 +00:00
3 changed files with 751 additions and 711 deletions
Showing only changes of commit 9137909a64 - Show all commits

269
go-sgf/src/go.rs Normal file
View File

@ -0,0 +1,269 @@
// https://red-bean.com/sgf/user_guide/index.html
// https://red-bean.com/sgf/sgf4.html
// todo: support collections in a file
// Properties to support. Remove each one as it gets support.
// B
// KO
// MN
// W
// AB
// AE
// AW
// PL
// C
// DM
// GB
// GW
// HO
// N
// UC
// V
// BM
// DO
// IT
// TE
// AR
// CR
// DD
// LB
// LN
// MA
// SL
// SQ
// TR
// AP
// CA
// FF
// GM
// ST
// SZ
// AN
// BR
// BT
// CP
// DT
// EV
// GN
// GC
// ON
// OT
// PB
// PC
// PW
// RE
// RO
// RU
// SO
// TM
// US
// WR
// WT
// BL
// OB
// OW
// WL
// FG
// PM
// VW
use crate::tree::{parse_collection, parse_size, ParseSizeError, Size};
use nom::IResult;
#[derive(Debug)]
pub enum Error<'a> {
InvalidField,
InvalidBoardSize,
Incomplete,
InvalidSgf(nom::error::VerboseError<&'a str>),
}
impl<'a> From<nom::Err<nom::error::VerboseError<&'a str>>> for Error<'a> {
fn from(err: nom::Err<nom::error::VerboseError<&'a str>>) -> Self {
match err {
nom::Err::Incomplete(_) => Error::Incomplete,
nom::Err::Error(e) => Error::InvalidSgf(e.clone()),
nom::Err::Failure(e) => Error::InvalidSgf(e.clone()),
}
}
}
impl<'a> From<ParseSizeError> for Error<'a> {
fn from(_: ParseSizeError) -> Self {
Self::InvalidBoardSize
}
}
#[derive(Debug)]
pub struct GameTree {
pub file_format: i8,
pub app: Option<String>,
pub game_type: GameType,
pub board_size: Size,
pub text: String,
}
pub struct GameInfo {
pub annotator: Option<String>,
pub copyright: Option<String>,
pub event: Option<String>,
// Games can be played across multiple days, even multiple years. The format specifies
// shortcuts.
pub date_time: Vec<chrono::NaiveDate>,
pub location: Option<String>,
// special rules for the round-number and type
pub round: Option<String>,
pub ruleset: Option<String>,
pub source: Option<String>,
pub time_limits: Option<std::time::Duration>,
pub game_keeper: Option<String>,
pub game_name: Option<String>,
pub game_comments: Option<String>,
pub black_player: Option<String>,
pub black_rank: Option<String>,
pub black_team: Option<String>,
pub white_player: Option<String>,
pub white_rank: Option<String>,
pub white_team: Option<String>,
pub opening: Option<String>,
pub overtime: Option<String>,
pub result: Option<GameResult>,
}
pub enum GameResult {
Annulled,
Draw,
Black(Win),
White(Win),
}
pub enum Win {
Score(i32),
Resignation,
Forfeit,
Time,
}
#[derive(Debug, PartialEq)]
pub enum GameType {
Go,
Unsupported,
}
enum PropType {
Move,
Setup,
Root,
GameInfo,
}
enum PropValue {
Empty,
Number,
Real,
Double,
Color,
SimpleText,
Text,
Point,
Move,
Stone,
}
pub fn parse_sgf<'a>(input: &'a str) -> Result<Vec<GameTree>, Error<'a>> {
let (input, trees) = parse_collection::<nom::error::VerboseError<&'a str>>(input)?;
let games = trees
.into_iter()
.map(|tree| {
let file_format = match tree.sequence[0].find_prop("FF") {
Some(prop) => prop.values[0].parse::<i8>().unwrap(),
None => 4,
};
let app = tree.sequence[0]
.find_prop("AP")
.map(|prop| prop.values[0].clone());
let board_size = match tree.sequence[0].find_prop("SZ") {
Some(prop) => Size::try_from(prop.values[0].as_str())?,
None => Size {
width: 19,
height: 19,
},
};
Ok(GameTree {
file_format,
app,
game_type: GameType::Go,
board_size,
text: input.to_owned(),
})
})
.collect::<Result<Vec<GameTree>, Error>>()?;
Ok(games)
}
#[cfg(test)]
mod tests {
use super::*;
use crate::tree::Size;
use std::fs::File;
use std::io::Read;
const EXAMPLE: &'static str = "(;FF[4]C[root](;C[a];C[b](;C[c])
(;C[d];C[e]))
(;C[f](;C[g];C[h];C[i])
(;C[j])))";
fn with_text(text: &str, f: impl FnOnce(Vec<GameTree>)) {
let games = parse_sgf(text).unwrap();
f(games);
}
fn with_file(path: &std::path::Path, f: impl FnOnce(Vec<GameTree>)) {
let mut file = File::open(path).unwrap();
let mut text = String::new();
let _ = file.read_to_string(&mut text);
with_text(&text, f);
}
#[test]
fn it_parses_game_root() {
with_text(EXAMPLE, |trees| {
assert_eq!(trees.len(), 1);
let tree = &trees[0];
assert_eq!(tree.file_format, 4);
assert_eq!(tree.app, None);
assert_eq!(tree.game_type, GameType::Go);
assert_eq!(
tree.board_size,
Size {
width: 19,
height: 19
}
);
// assert_eq!(tree.text, EXAMPLE.to_owned());
});
with_file(std::path::Path::new("test_data/print1.sgf"), |trees| {
assert_eq!(trees.len(), 1);
let tree = &trees[0];
assert_eq!(tree.file_format, 4);
assert_eq!(tree.app, None);
assert_eq!(tree.game_type, GameType::Go);
assert_eq!(
tree.board_size,
Size {
width: 19,
height: 19
}
);
});
}
}

View File

@ -1,81 +1,6 @@
// https://red-bean.com/sgf/user_guide/index.html pub mod go;
// https://red-bean.com/sgf/sgf4.html pub mod tree;
// todo: support collections in a file
// Properties to support. Remove each one as it gets support.
// B
// KO
// MN
// W
// AB
// AE
// AW
// PL
// C
// DM
// GB
// GW
// HO
// N
// UC
// V
// BM
// DO
// IT
// TE
// AR
// CR
// DD
// LB
// LN
// MA
// SL
// SQ
// TR
// AP
// CA
// FF
// GM
// ST
// SZ
// AN
// BR
// BT
// CP
// DT
// EV
// GN
// GC
// ON
// OT
// PB
// PC
// PW
// RE
// RO
// RU
// SO
// TM
// US
// WR
// WT
// BL
// OB
// OW
// WL
// FG
// PM
// VW
use nom::{
branch::alt,
bytes::complete::{escaped, escaped_transform, is_not, tag},
character::complete::{alpha1, digit1, multispace0, multispace1, none_of, one_of},
combinator::{opt, value},
multi::{many0, many1, separated_list1},
sequence::delimited,
Finish, IResult,
};
use thiserror::Error; use thiserror::Error;
pub enum Warning {} pub enum Warning {}
@ -115,637 +40,3 @@ impl From<nom::error::VerboseError<&str>> for ParseError {
} }
} }
*/ */
// todo: support ST root node
#[derive(Debug)]
pub struct GameTree {
pub file_format: i8,
pub app: Option<String>,
pub game_type: GameType,
pub board_size: Size,
pub text: String,
}
pub struct GameInfo {
pub annotator: Option<String>,
pub copyright: Option<String>,
pub event: Option<String>,
// Games can be played across multiple days, even multiple years. The format specifies
// shortcuts.
pub date_time: Vec<chrono::NaiveDate>,
pub location: Option<String>,
// special rules for the round-number and type
pub round: Option<String>,
pub ruleset: Option<String>,
pub source: Option<String>,
pub time_limits: Option<std::time::Duration>,
pub game_keeper: Option<String>,
pub game_name: Option<String>,
pub game_comments: Option<String>,
pub black_player: Option<String>,
pub black_rank: Option<String>,
pub black_team: Option<String>,
pub white_player: Option<String>,
pub white_rank: Option<String>,
pub white_team: Option<String>,
pub opening: Option<String>,
pub overtime: Option<String>,
pub result: Option<GameResult>,
}
pub enum GameResult {
Annulled,
Draw,
Black(Win),
White(Win),
}
pub enum Win {
Score(i32),
Resignation,
Forfeit,
Time,
}
#[derive(Debug, PartialEq)]
pub struct Size {
width: i32,
height: i32,
}
#[derive(Debug, PartialEq)]
pub enum GameType {
Go,
Unsupported,
}
// struct Sequence(Node);
/*
struct Node {
// properties
}
*/
enum PropType {
Move,
Setup,
Root,
GameInfo,
}
enum PropValue {
Empty,
Number,
Real,
Double,
Color,
SimpleText,
Text,
Point,
Move,
Stone,
}
pub fn parse_sgf(input: &str) -> Result<Vec<GameTree>, ParseError> {
let (_, trees) = parse_collection::<nom::error::Error<&str>>(input).finish()?;
trees
.into_iter()
.map(|tree| {
let file_format = match tree.sequence[0].find_prop("FF") {
Some(prop) => prop.values[0].parse::<i8>().unwrap(),
None => 4,
};
let app = tree.sequence[0]
.find_prop("AP")
.map(|prop| prop.values[0].clone());
let board_size = match tree.sequence[0].find_prop("SZ") {
Some(prop) => {
let (_, size) =
parse_size::<nom::error::Error<&str>>(prop.values[0].as_str()).finish()?;
size
}
None => Size {
width: 19,
height: 19,
},
};
Ok(GameTree {
file_format,
app,
game_type: GameType::Go,
board_size,
text: input.to_owned(),
})
})
.collect()
}
#[derive(Debug, PartialEq)]
struct Tree {
sequence: Vec<Node>,
sub_sequences: Vec<Tree>,
}
impl ToString for Tree {
fn to_string(&self) -> String {
let sequence = self
.sequence
.iter()
.map(|node| node.to_string())
.collect::<String>();
let subsequences = self
.sub_sequences
.iter()
.map(|seq| seq.to_string())
.collect::<String>();
format!("({}{})", sequence, subsequences)
}
}
#[derive(Debug, PartialEq)]
struct Node {
properties: Vec<Property>,
}
impl ToString for Node {
fn to_string(&self) -> String {
let props = self
.properties
.iter()
.map(|prop| prop.to_string())
.collect::<String>();
format!(";{}", props)
}
}
impl Node {
fn find_prop(&self, ident: &str) -> Option<Property> {
self.properties
.iter()
.find(|prop| prop.ident == ident)
.cloned()
}
}
#[derive(Clone, Debug, PartialEq)]
struct Property {
ident: String,
values: Vec<String>,
}
impl ToString for Property {
fn to_string(&self) -> String {
let values = self
.values
.iter()
.map(|val| format!("[{}]", val))
.collect::<String>();
format!("{}{}", self.ident, values)
}
}
fn parse_collection<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Vec<Tree>, E> {
separated_list1(multispace1, parse_tree)(input)
}
// note: must preserve unknown properties
// note: must fix or preserve illegally formatted game-info properties
// note: must correct or delete illegally foramtted properties, but display a warning
fn parse_tree<'a, E: nom::error::ParseError<&'a str>>(input: &'a str) -> IResult<&'a str, Tree, E> {
println!("::: parse_tree: {}", input);
let (input, _) = multispace0(input)?;
delimited(tag("("), parse_sequence, tag(")"))(input)
}
fn parse_sequence<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Tree, E> {
println!("::: parse_sequence: {}", input);
let (input, _) = multispace0(input)?;
let (input, nodes) = many1(parse_node)(input)?;
let (input, sub_sequences) = many0(parse_tree)(input)?;
Ok((
input,
Tree {
sequence: nodes,
sub_sequences,
},
))
}
fn parse_node<'a, E: nom::error::ParseError<&'a str>>(input: &'a str) -> IResult<&'a str, Node, E> {
println!("::: parse_node: {}", input);
let (input, _) = multispace0(input)?;
let (input, _) = tag(";")(input)?;
let (input, properties) = many1(parse_property)(input)?;
Ok((input, Node { properties }))
}
fn parse_property<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Property, E> {
println!(":: parse_property: {}", input);
let (input, _) = multispace0(input)?;
let (input, ident) = alpha1(input)?;
let (input, values) = many1(parse_propval)(input)?;
let values = values
.into_iter()
.map(|v| v.to_owned())
.collect::<Vec<String>>();
Ok((
input,
Property {
ident: ident.to_owned(),
values,
},
))
}
fn parse_propval<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, String, E> {
let (input, _) = multispace0(input)?;
println!("- {}", input);
let (input, _) = tag("[")(input)?;
println!("-- {}", input);
let (input, value) = parse_propval_text(input)?;
println!("--- {}", input);
let (input, _) = tag("]")(input)?;
Ok((input, value.unwrap_or(String::new())))
}
fn parse_propval_text<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Option<String>, E> {
let (input, value) = opt(escaped_transform(
none_of("\\]"),
'\\',
alt((
value("]", tag("]")),
value("\\", tag("\\")),
value("", tag("\n")),
)),
))(input)?;
Ok((input, value.map(|v| v.to_owned())))
}
fn parse_size<'a, E: nom::error::ParseError<&'a str>>(input: &'a str) -> IResult<&'a str, Size, E> {
let (input, dimensions) = separated_list1(tag(":"), digit1)(input)?;
let (width, height) = match dimensions.as_slice() {
[width] => (width.parse::<i32>().unwrap(), width.parse::<i32>().unwrap()),
[width, height] => (
width.parse::<i32>().unwrap(),
height.parse::<i32>().unwrap(),
),
_ => (19, 19),
};
Ok((input, Size { width, height }))
}
#[cfg(test)]
mod tests {
use std::{fs::File, io::Read};
use super::*;
const EXAMPLE: &'static str = "(;FF[4]C[root](;C[a];C[b](;C[c])
(;C[d];C[e]))
(;C[f](;C[g];C[h];C[i])
(;C[j])))";
#[test]
fn it_can_parse_properties() {
let (_, prop) = parse_property::<nom::error::VerboseError<&str>>("C[a]").unwrap();
assert_eq!(
prop,
Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()]
}
);
let (_, prop) = parse_property::<nom::error::VerboseError<&str>>("C[a][b][c]").unwrap();
assert_eq!(
prop,
Property {
ident: "C".to_owned(),
values: vec!["a".to_owned(), "b".to_owned(), "c".to_owned()]
}
);
}
#[test]
fn it_can_parse_a_standalone_node() {
let (_, node) = parse_node::<nom::error::VerboseError<&str>>(";B[ab]").unwrap();
assert_eq!(
node,
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
}
);
let (_, node) =
parse_node::<nom::error::VerboseError<&str>>(";B[ab];W[dp];B[pq]C[some comments]")
.unwrap();
assert_eq!(
node,
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
}
);
}
#[test]
fn it_can_parse_a_simple_sequence() {
let (_, sequence) =
parse_tree::<nom::error::VerboseError<&str>>("(;B[ab];W[dp];B[pq]C[some comments])")
.unwrap();
assert_eq!(
sequence,
Tree {
sequence: vec![
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
},
Node {
properties: vec![Property {
ident: "W".to_owned(),
values: vec!["dp".to_owned()]
}]
},
Node {
properties: vec![
Property {
ident: "B".to_owned(),
values: vec!["pq".to_owned()]
},
Property {
ident: "C".to_owned(),
values: vec!["some comments".to_owned()]
}
]
}
],
sub_sequences: vec![],
}
);
}
#[test]
fn it_can_parse_a_sequence_with_subsequences() {
let text = "(;C[a];C[b](;C[c])(;C[d];C[e]))";
let (_, sequence) = parse_tree::<nom::error::VerboseError<&str>>(text).unwrap();
let main_sequence = vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()],
}],
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["b".to_owned()],
}],
},
];
let subsequence_1 = Tree {
sequence: vec![Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["c".to_owned()],
}],
}],
sub_sequences: vec![],
};
let subsequence_2 = Tree {
sequence: vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["d".to_owned()],
}],
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["e".to_owned()],
}],
},
],
sub_sequences: vec![],
};
assert_eq!(
sequence,
Tree {
sequence: main_sequence,
sub_sequences: vec![subsequence_1, subsequence_2],
}
);
}
#[test]
fn it_can_parse_example_1() {
let (_, ex_tree) = parse_tree::<nom::error::VerboseError<&str>>(EXAMPLE).unwrap();
assert_eq!(ex_tree.sequence.len(), 1);
assert_eq!(ex_tree.sequence[0].properties.len(), 2);
assert_eq!(
ex_tree.sequence[0].properties[0],
Property {
ident: "FF".to_owned(),
values: vec!["4".to_owned()]
}
);
assert_eq!(ex_tree.sub_sequences.len(), 2);
assert_eq!(ex_tree.sub_sequences[0].sequence.len(), 2);
assert_eq!(
ex_tree.sub_sequences[0].sequence,
vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()]
}]
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["b".to_owned()]
}]
},
]
);
assert_eq!(ex_tree.sub_sequences[0].sub_sequences.len(), 2);
}
#[test]
fn it_can_regenerate_the_tree() {
let (_, tree1) = parse_tree::<nom::error::VerboseError<&str>>(EXAMPLE).unwrap();
assert_eq!(
tree1.to_string(),
"(;FF[4]C[root](;C[a];C[b](;C[c])(;C[d];C[e]))(;C[f](;C[g];C[h];C[i])(;C[j])))"
);
let (_, tree2) = parse_tree::<nom::error::VerboseError<&str>>(&tree1.to_string()).unwrap();
assert_eq!(tree1, tree2);
}
#[test]
fn it_parses_propvals() {
let (_, propval) = parse_propval::<nom::error::VerboseError<&str>>("[]").unwrap();
assert_eq!(propval, "".to_owned());
let (_, propval) =
parse_propval::<nom::error::VerboseError<&str>>("[normal propval]").unwrap();
assert_eq!(propval, "normal propval".to_owned());
let (_, propval) =
parse_propval::<nom::error::VerboseError<&str>>(r"[need an [escape\] in the propval]")
.unwrap();
assert_eq!(propval, "need an [escape] in the propval".to_owned());
}
#[test]
fn it_parses_propvals_with_hard_linebreaks() {
let (_, propval) = parse_propval_text::<nom::error::VerboseError<&str>>(
"There are hard linebreaks & soft linebreaks.
Soft linebreaks...",
)
.unwrap();
assert_eq!(
propval,
Some(
"There are hard linebreaks & soft linebreaks.
Soft linebreaks..."
.to_owned()
)
);
}
#[test]
fn it_parses_propvals_with_escaped_closing_brackets() {
let (_, propval) =
parse_propval_text::<nom::error::VerboseError<&str>>(r"escaped closing \] bracket")
.unwrap();
assert_eq!(
propval,
Some(r"escaped closing ] bracket".to_owned()).to_owned()
);
}
#[test]
fn it_parses_propvals_with_soft_linebreaks() {
let (_, propval) = parse_propval_text::<nom::error::VerboseError<&str>>(
r"Soft linebreaks are linebreaks preceeded by '\\' like this one >o\
k<. Hard line breaks are all other linebreaks.",
)
.unwrap();
assert_eq!(
propval,
Some("Soft linebreaks are linebreaks preceeded by '\\' like this one >ok<. Hard line breaks are all other linebreaks.".to_owned())
.to_owned()
);
}
fn with_text(text: &str, f: impl FnOnce(Vec<GameTree>)) {
f(parse_sgf(text).unwrap());
}
fn with_file(path: &std::path::Path, f: impl FnOnce(Vec<GameTree>)) {
let mut file = File::open(path).unwrap();
let mut text = String::new();
let _ = file.read_to_string(&mut text);
with_text(&text, f);
}
#[test]
fn it_parses_game_root() {
with_text(EXAMPLE, |trees| {
assert_eq!(trees.len(), 1);
let tree = &trees[0];
assert_eq!(tree.file_format, 4);
assert_eq!(tree.app, None);
assert_eq!(tree.game_type, GameType::Go);
assert_eq!(
tree.board_size,
Size {
width: 19,
height: 19
}
);
assert_eq!(tree.text, EXAMPLE.to_owned());
});
with_file(std::path::Path::new("test_data/print1.sgf"), |trees| {
assert_eq!(trees.len(), 1);
let tree = &trees[0];
assert_eq!(tree.file_format, 4);
assert_eq!(tree.app, None);
assert_eq!(tree.game_type, GameType::Go);
assert_eq!(
tree.board_size,
Size {
width: 19,
height: 19
}
);
});
}
/*
#[test]
fn it_parses_linebreaks() {
with_file(
std::path::Path::new("test_data/linebreak_tests.sgf"),
|tree| {},
);
}
#[test]
fn it_parses_ff4_a() {
with_file(std::path::Path::new("test_data/ff4_a.sgf"), |tree| {});
}
#[test]
fn it_parses_ff4_b() {
with_file(std::path::Path::new("test_data/ff4_b.sgf"), |tree| {});
}
#[test]
fn it_parses_ff4_ex() {
with_file(std::path::Path::new("test_data/ff4_ex.sgf"), |tree| {});
}
*/
}

480
go-sgf/src/tree.rs Normal file
View File

@ -0,0 +1,480 @@
use std::num::ParseIntError;
use nom::{
branch::alt,
bytes::complete::{escaped_transform, tag},
character::complete::{alpha1, digit1, multispace0, multispace1, none_of},
combinator::{opt, value},
multi::{many0, many1, separated_list1},
sequence::delimited,
IResult,
};
#[derive(Debug)]
pub enum ParseSizeError {
ParseIntError(ParseIntError),
InsufficientArguments,
}
impl From<ParseIntError> for ParseSizeError {
fn from(e: ParseIntError) -> Self {
Self::ParseIntError(e)
}
}
#[derive(Debug, PartialEq)]
pub struct Size {
pub width: i32,
pub height: i32,
}
impl TryFrom<&str> for Size {
type Error = ParseSizeError;
fn try_from(s: &str) -> Result<Self, Self::Error> {
let parts = s
.split(':')
.map(|v| v.parse::<i32>())
.collect::<Result<Vec<i32>, ParseIntError>>()?;
match parts[..] {
[width, height, ..] => Ok(Size { width, height }),
[dim] => Ok(Size {
width: dim,
height: dim,
}),
[] => Err(ParseSizeError::InsufficientArguments),
}
}
}
#[derive(Debug, PartialEq)]
pub struct Tree {
pub sequence: Vec<Node>,
pub sub_sequences: Vec<Tree>,
}
impl ToString for Tree {
fn to_string(&self) -> String {
let sequence = self
.sequence
.iter()
.map(|node| node.to_string())
.collect::<String>();
let subsequences = self
.sub_sequences
.iter()
.map(|seq| seq.to_string())
.collect::<String>();
format!("({}{})", sequence, subsequences)
}
}
#[derive(Debug, PartialEq)]
pub struct Node {
pub properties: Vec<Property>,
}
impl ToString for Node {
fn to_string(&self) -> String {
let props = self
.properties
.iter()
.map(|prop| prop.to_string())
.collect::<String>();
format!(";{}", props)
}
}
impl Node {
pub fn find_prop(&self, ident: &str) -> Option<Property> {
self.properties
.iter()
.find(|prop| prop.ident == ident)
.cloned()
}
}
#[derive(Clone, Debug, PartialEq)]
pub struct Property {
pub ident: String,
pub values: Vec<String>,
}
impl ToString for Property {
fn to_string(&self) -> String {
let values = self
.values
.iter()
.map(|val| format!("[{}]", val))
.collect::<String>();
format!("{}{}", self.ident, values)
}
}
pub fn parse_collection<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Vec<Tree>, E> {
separated_list1(multispace1, parse_tree)(input)
}
// note: must preserve unknown properties
// note: must fix or preserve illegally formatted game-info properties
// note: must correct or delete illegally foramtted properties, but display a warning
fn parse_tree<'a, E: nom::error::ParseError<&'a str>>(input: &'a str) -> IResult<&'a str, Tree, E> {
println!("::: parse_tree: {}", input);
let (input, _) = multispace0(input)?;
delimited(tag("("), parse_sequence, tag(")"))(input)
}
fn parse_sequence<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Tree, E> {
println!("::: parse_sequence: {}", input);
let (input, _) = multispace0(input)?;
let (input, nodes) = many1(parse_node)(input)?;
let (input, sub_sequences) = many0(parse_tree)(input)?;
Ok((
input,
Tree {
sequence: nodes,
sub_sequences,
},
))
}
fn parse_node<'a, E: nom::error::ParseError<&'a str>>(input: &'a str) -> IResult<&'a str, Node, E> {
println!("::: parse_node: {}", input);
let (input, _) = multispace0(input)?;
let (input, _) = tag(";")(input)?;
let (input, properties) = many1(parse_property)(input)?;
Ok((input, Node { properties }))
}
fn parse_property<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Property, E> {
println!(":: parse_property: {}", input);
let (input, _) = multispace0(input)?;
let (input, ident) = alpha1(input)?;
let (input, values) = many1(parse_propval)(input)?;
let values = values
.into_iter()
.map(|v| v.to_owned())
.collect::<Vec<String>>();
Ok((
input,
Property {
ident: ident.to_owned(),
values,
},
))
}
fn parse_propval<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, String, E> {
let (input, _) = multispace0(input)?;
println!("- {}", input);
let (input, _) = tag("[")(input)?;
println!("-- {}", input);
let (input, value) = parse_propval_text(input)?;
println!("--- {}", input);
let (input, _) = tag("]")(input)?;
Ok((input, value.unwrap_or(String::new())))
}
fn parse_propval_text<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Option<String>, E> {
let (input, value) = opt(escaped_transform(
none_of("\\]"),
'\\',
alt((
value("]", tag("]")),
value("\\", tag("\\")),
value("", tag("\n")),
)),
))(input)?;
Ok((input, value.map(|v| v.to_owned())))
}
pub fn parse_size<'a, E: nom::error::ParseError<&'a str>>(
input: &'a str,
) -> IResult<&'a str, Size, E> {
let (input, dimensions) = separated_list1(tag(":"), digit1)(input)?;
let (width, height) = match dimensions.as_slice() {
[width] => (width.parse::<i32>().unwrap(), width.parse::<i32>().unwrap()),
[width, height] => (
width.parse::<i32>().unwrap(),
height.parse::<i32>().unwrap(),
),
_ => (19, 19),
};
Ok((input, Size { width, height }))
}
#[cfg(test)]
mod test {
use std::{fs::File, io::Read};
use super::*;
const EXAMPLE: &'static str = "(;FF[4]C[root](;C[a];C[b](;C[c])
(;C[d];C[e]))
(;C[f](;C[g];C[h];C[i])
(;C[j])))";
#[test]
fn it_can_parse_properties() {
let (_, prop) = parse_property::<nom::error::VerboseError<&str>>("C[a]").unwrap();
assert_eq!(
prop,
Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()]
}
);
let (_, prop) = parse_property::<nom::error::VerboseError<&str>>("C[a][b][c]").unwrap();
assert_eq!(
prop,
Property {
ident: "C".to_owned(),
values: vec!["a".to_owned(), "b".to_owned(), "c".to_owned()]
}
);
}
#[test]
fn it_can_parse_a_standalone_node() {
let (_, node) = parse_node::<nom::error::VerboseError<&str>>(";B[ab]").unwrap();
assert_eq!(
node,
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
}
);
let (_, node) =
parse_node::<nom::error::VerboseError<&str>>(";B[ab];W[dp];B[pq]C[some comments]")
.unwrap();
assert_eq!(
node,
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
}
);
}
#[test]
fn it_can_parse_a_simple_sequence() {
let (_, sequence) =
parse_tree::<nom::error::VerboseError<&str>>("(;B[ab];W[dp];B[pq]C[some comments])")
.unwrap();
assert_eq!(
sequence,
Tree {
sequence: vec![
Node {
properties: vec![Property {
ident: "B".to_owned(),
values: vec!["ab".to_owned()]
}]
},
Node {
properties: vec![Property {
ident: "W".to_owned(),
values: vec!["dp".to_owned()]
}]
},
Node {
properties: vec![
Property {
ident: "B".to_owned(),
values: vec!["pq".to_owned()]
},
Property {
ident: "C".to_owned(),
values: vec!["some comments".to_owned()]
}
]
}
],
sub_sequences: vec![],
}
);
}
#[test]
fn it_can_parse_a_sequence_with_subsequences() {
let text = "(;C[a];C[b](;C[c])(;C[d];C[e]))";
let (_, sequence) = parse_tree::<nom::error::VerboseError<&str>>(text).unwrap();
let main_sequence = vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()],
}],
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["b".to_owned()],
}],
},
];
let subsequence_1 = Tree {
sequence: vec![Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["c".to_owned()],
}],
}],
sub_sequences: vec![],
};
let subsequence_2 = Tree {
sequence: vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["d".to_owned()],
}],
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["e".to_owned()],
}],
},
],
sub_sequences: vec![],
};
assert_eq!(
sequence,
Tree {
sequence: main_sequence,
sub_sequences: vec![subsequence_1, subsequence_2],
}
);
}
#[test]
fn it_can_parse_example_1() {
let (_, ex_tree) = parse_tree::<nom::error::VerboseError<&str>>(EXAMPLE).unwrap();
assert_eq!(ex_tree.sequence.len(), 1);
assert_eq!(ex_tree.sequence[0].properties.len(), 2);
assert_eq!(
ex_tree.sequence[0].properties[0],
Property {
ident: "FF".to_owned(),
values: vec!["4".to_owned()]
}
);
assert_eq!(ex_tree.sub_sequences.len(), 2);
assert_eq!(ex_tree.sub_sequences[0].sequence.len(), 2);
assert_eq!(
ex_tree.sub_sequences[0].sequence,
vec![
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["a".to_owned()]
}]
},
Node {
properties: vec![Property {
ident: "C".to_owned(),
values: vec!["b".to_owned()]
}]
},
]
);
assert_eq!(ex_tree.sub_sequences[0].sub_sequences.len(), 2);
}
#[test]
fn it_can_regenerate_the_tree() {
let (_, tree1) = parse_tree::<nom::error::VerboseError<&str>>(EXAMPLE).unwrap();
assert_eq!(
tree1.to_string(),
"(;FF[4]C[root](;C[a];C[b](;C[c])(;C[d];C[e]))(;C[f](;C[g];C[h];C[i])(;C[j])))"
);
let (_, tree2) = parse_tree::<nom::error::VerboseError<&str>>(&tree1.to_string()).unwrap();
assert_eq!(tree1, tree2);
}
#[test]
fn it_parses_propvals() {
let (_, propval) = parse_propval::<nom::error::VerboseError<&str>>("[]").unwrap();
assert_eq!(propval, "".to_owned());
let (_, propval) =
parse_propval::<nom::error::VerboseError<&str>>("[normal propval]").unwrap();
assert_eq!(propval, "normal propval".to_owned());
let (_, propval) =
parse_propval::<nom::error::VerboseError<&str>>(r"[need an [escape\] in the propval]")
.unwrap();
assert_eq!(propval, "need an [escape] in the propval".to_owned());
}
#[test]
fn it_parses_propvals_with_hard_linebreaks() {
let (_, propval) = parse_propval_text::<nom::error::VerboseError<&str>>(
"There are hard linebreaks & soft linebreaks.
Soft linebreaks...",
)
.unwrap();
assert_eq!(
propval,
Some(
"There are hard linebreaks & soft linebreaks.
Soft linebreaks..."
.to_owned()
)
);
}
#[test]
fn it_parses_propvals_with_escaped_closing_brackets() {
let (_, propval) =
parse_propval_text::<nom::error::VerboseError<&str>>(r"escaped closing \] bracket")
.unwrap();
assert_eq!(
propval,
Some(r"escaped closing ] bracket".to_owned()).to_owned()
);
}
#[test]
fn it_parses_propvals_with_soft_linebreaks() {
let (_, propval) = parse_propval_text::<nom::error::VerboseError<&str>>(
r"Soft linebreaks are linebreaks preceeded by '\\' like this one >o\
k<. Hard line breaks are all other linebreaks.",
)
.unwrap();
assert_eq!(
propval,
Some("Soft linebreaks are linebreaks preceeded by '\\' like this one >ok<. Hard line breaks are all other linebreaks.".to_owned())
.to_owned()
);
}
}