add glyph id mapping

2025-03-02 16:34:28 +01:00 · 2025-03-02 16:34:28 +01:00 · db07f3772b
parent 4b1d8718d6
commit db07f3772b
7 changed files with 396 additions and 18 deletions
--- a/Cargo.lock
+++ b/Cargo.lock
@ -106,13 +106,24 @@ version = "1.0.3"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "5b63caa9aa9397e2d9480a9b13673856c78d8ac123288526c37d7839f2a86990"

+[[package]]
+name = "deranged"
+version = "0.3.11"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b42b6fa04a440b495c8b04d0e71b707c585f83cb9cb28cf8cd0d976c315e31b4"
+dependencies = [
+ "powerfmt",
+]
+
 [[package]]
 name = "font-explorer"
 version = "0.1.0"
 dependencies = [
 "bincode",
 "clap",
+ "log",
 "serde",
+ "simplelog",
 ]

 [[package]]
@ -127,6 +138,45 @@ version = "1.70.1"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "7943c866cc5cd64cbc25b2e01621d07fa8eb2a1a23160ee81ce38704e97b8ecf"

+[[package]]
+name = "itoa"
+version = "1.0.14"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "d75a2a4b1b190afb6f5425f10f6a8f959d2ea0b9c2b1d79553551850539e4674"
+
+[[package]]
+name = "libc"
+version = "0.2.170"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "875b3680cb2f8f71bdcf9a30f38d48282f5d3c95cbf9b3fa57269bb5d5c06828"
+
+[[package]]
+name = "log"
+version = "0.4.26"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "30bde2b3dc3671ae49d8e2e9f044c7c005836e7a023ee57cffa25ab82764bb9e"
+
+[[package]]
+name = "num-conv"
+version = "0.1.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "51d515d32fb182ee37cda2ccdcb92950d6a3c2893aa280e540671c2cd0f3b1d9"
+
+[[package]]
+name = "num_threads"
+version = "0.1.7"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "5c7398b9c8b70908f6371f47ed36737907c87c52af34c268fed0bf0ceb92ead9"
+dependencies = [
+ "libc",
+]
+
+[[package]]
+name = "powerfmt"
+version = "0.2.0"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "439ee305def115ba05938db6eb1644ff94165c5ab5e9420d1c1bcedbba909391"
+
 [[package]]
 name = "proc-macro2"
 version = "1.0.89"
@ -165,6 +215,17 @@ dependencies = [
 "syn",
 ]

+[[package]]
+name = "simplelog"
+version = "0.12.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "16257adbfaef1ee58b1363bdc0664c9b8e1e30aed86049635fb5f147d065a9c0"
+dependencies = [
+ "log",
+ "termcolor",
+ "time",
+]
+
 [[package]]
 name = "strsim"
 version = "0.11.1"
@ -182,6 +243,48 @@ dependencies = [
 "unicode-ident",
 ]

+[[package]]
+name = "termcolor"
+version = "1.4.1"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "06794f8f6c5c898b3275aebefa6b8a1cb24cd2c6c79397ab15774837a0bc5755"
+dependencies = [
+ "winapi-util",
+]
+
+[[package]]
+name = "time"
+version = "0.3.37"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "35e7868883861bd0e56d9ac6efcaaca0d6d5d82a2a7ec8209ff492c07cf37b21"
+dependencies = [
+ "deranged",
+ "itoa",
+ "libc",
+ "num-conv",
+ "num_threads",
+ "powerfmt",
+ "serde",
+ "time-core",
+ "time-macros",
+]
+
+[[package]]
+name = "time-core"
+version = "0.1.2"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "ef927ca75afb808a4d64dd374f00a2adf8d0fcff8e7b184af886c3c87ec4a3f3"
+
+[[package]]
+name = "time-macros"
+version = "0.2.19"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "2834e6017e3e5e4b9834939793b282bc03b37a3336245fa820e35e233e2a85de"
+dependencies = [
+ "num-conv",
+ "time-core",
+]
+
 [[package]]
 name = "unicode-ident"
 version = "1.0.13"
@ -194,6 +297,15 @@ version = "0.2.2"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "06abde3611657adf66d383f00b093d7faecc7fa57071cce2578660c9f1010821"

+[[package]]
+name = "winapi-util"
+version = "0.1.9"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "cf221c93e13a30d793f7645a0e7762c55d169dbb0a49671918a2319d289b10bb"
+dependencies = [
+ "windows-sys",
+]
+
 [[package]]
 name = "windows-sys"
 version = "0.59.0"
--- a/Cargo.toml
+++ b/Cargo.toml
@ -6,4 +6,6 @@ edition = "2021"
 [dependencies]
 bincode = "1.3.3"
 clap = { version = "4.5.21", features = ["derive"] }
+log = "0.4.26"
 serde = { version = "1.0.215", features = ["derive", "serde_derive"] }
+simplelog = { version = "0.12.2", features = ["termcolor"] }
--- a/src/font.rs
+++ b/src/font.rs
@ -1,29 +1,56 @@
 mod table_directory;
+mod character_map;
+mod search;

-use std::io::{BufReader, Read, Seek};
+use std::{fs::File, io::{BufReader, Read}};

-use bincode::{Options, Result};
+use bincode::Options;
+use character_map::CharacterMap;
+use log::{debug, info};
 use table_directory::TableDirectory;

-fn deserialize<R: Read + Seek, T: serde::de::DeserializeOwned>(reader: R) -> Result<T> {
+fn deserialize<R: Read, T: serde::de::DeserializeOwned>(reader: R) -> bincode::Result<T> {
 	bincode::options().with_big_endian().with_fixint_encoding().deserialize_from::<R, T>(reader)
 }

 #[derive(Debug)]
-pub struct Font<R: Read + Seek> {
-	data_source: BufReader<R>,
-	table_directory: TableDirectory
+pub struct Font {
+	data_source: BufReader<File>,
+	table_directory: TableDirectory,
+	character_map: CharacterMap
 }
 	
-impl<R: Read + Seek> Font<R> {
-	pub fn new(reader: R) -> Result<Self> {
-		let mut buf_reader = BufReader::new(reader);
+impl Font {
+	pub fn new(file: File) -> bincode::Result<Self> {
+		let mut buf_reader = BufReader::new(file.try_clone()?);
 		let table_directory = TableDirectory::new(&mut buf_reader)?;

+		let Some(cmap) = table_directory.get_table("cmap") else {
+			return Err(Box::new(bincode::ErrorKind::Custom("Missing table 'cmap'".into())));
+		};
+
+		let character_map = CharacterMap::new(file.try_clone()?, cmap)?;
+		debug!("{character_map:#?}");
+
 		Ok(Font {
 			data_source: buf_reader,
-			table_directory: table_directory
+			table_directory,
+			character_map
 		})
 	}
+
+	pub fn get_data_for(&mut self, glyph: char) -> Result<(), &'static str> {
+		let Some(table) = self.table_directory.get_table("glyf") else {
+			return Err("Failed to get table. It doesn't exist");
+		};
+
+		let Some(glyph_id) = self.character_map.get_char(glyph) else {
+			return Err("Failed to get glyph id for char");
+		};
+
+		info!("glyph id for char '{glyph}' is {glyph_id}");
+
+		Ok(())
+	}
 }

--- a/src/font/character_map.rs
+++ b/src/font/character_map.rs
@ -0,0 +1,161 @@
+use std::{fs::File, io::{BufReader, Read, Seek, SeekFrom}};
+
+use bincode::{ErrorKind, Result};
+use log::{debug, error, warn};
+use serde::{de::DeserializeOwned, Deserialize};
+
+use super::{deserialize, table_directory::TableDirectoryRecord, search::SearchParameters};
+
+#[derive(Debug, Deserialize, Copy, Clone)]
+struct SegmentToDeltaHeader {
+	_length: u16,
+	_language: u16,
+}
+
+#[derive(Debug)]
+struct SegmentToDelta {
+	data_source: BufReader<File>,
+
+	header: SegmentToDeltaHeader,
+	search_params: SearchParameters,
+
+	end_code: Vec<u16>,
+	start_code: Vec<u16>,
+	id_delta: Vec<i16>,
+	id_range_offset: Vec<u16>,
+
+	glyph_table: u32
+}
+
+impl SegmentToDelta {
+	fn get_next_vec<R: Read + Seek, T: DeserializeOwned>(reader: &mut R, seg_count: u16) -> Result<Vec<T>> {
+		let mut values: Vec<T> = vec!();
+		for _ in 0..seg_count {
+			values.push(deserialize(reader.by_ref())?);
+		}
+
+		Ok(values)
+	}
+
+	fn new<R: Read + Seek>(reader: &mut R, file: File) -> Result<SegmentToDelta> {
+		let header: SegmentToDeltaHeader = deserialize(reader.by_ref())?;
+		let search_params: SearchParameters = deserialize(reader.by_ref())?;
+
+		let end_code = Self::get_next_vec(reader, search_params.seg_count_x2 >> 1)?;
+		reader.seek_relative(2)?;
+		let start_code = Self::get_next_vec(reader, search_params.seg_count_x2 >> 1)?;
+		let id_delta = Self::get_next_vec(reader, search_params.seg_count_x2 >> 1)?;
+		let id_range_offset = Self::get_next_vec(reader, search_params.seg_count_x2 >> 1)?;
+
+		Ok(SegmentToDelta {
+			data_source: BufReader::new(file),
+			glyph_table: reader.stream_position()? as u32,
+			header,
+			search_params,
+			end_code,
+			start_code,
+			id_delta,
+			id_range_offset
+		})
+	}
+
+	fn get_glyph_id(&mut self, c: u16) -> Option<u16> {
+		// Find index of first endcode that is greater or equal to c
+		let index =self.search_params.find_first_goe(self.end_code.as_slice(), c)?;
+		debug!("First endcode that is >={c} has index {index} (={})", self.end_code[index]);
+
+		if self.start_code[index] > c {
+			return None
+		}
+
+		let id_range_offset = self.id_range_offset[index];
+		debug!("range offset = {id_range_offset}");
+
+		if id_range_offset == 0 {
+			debug!("delta = {}", self.id_delta[index]);
+			return Some((c as i16 + self.id_delta[index]) as u16 % u16::MAX);
+		}
+
+		let glyph_map_offset = ((id_range_offset >> 1) as usize) + ((c - self.start_code[index]) as usize) - index;
+		if let Err(error) = self.data_source.seek(SeekFrom::Start(self.glyph_table as u64 + glyph_map_offset as u64)) {
+			error!("{error}");
+			return None;
+		}
+
+		if let Ok(glyph) = deserialize::<_, u16>(self.data_source.by_ref()) {
+			return Some(glyph);
+		}
+
+		return None;
+	}
+}
+
+#[derive(Debug)]
+pub struct CharacterMap {
+	_table_start: u32,
+	_length: u32,
+
+	mapping_table: SegmentToDelta
+}
+
+#[derive(Debug, Deserialize)]
+struct EncodingRecord {
+	platform: u16,
+	encoding: u16,
+	offset: u32,
+}
+
+impl CharacterMap {
+	fn find_unicode_table<R: Read + Seek>(reader: &mut R, num_tables: u16) -> Option<EncodingRecord> {
+		for _ in 0..num_tables {
+			let Ok(table) = deserialize::<_, EncodingRecord>(reader.by_ref()) else {
+				return None;
+			};
+
+			if table.platform == 0 && table.encoding == 3 {
+				return Some(table);
+			}
+		}
+
+		None
+	}
+
+	pub fn new(file: File, cmap: TableDirectoryRecord) -> Result<CharacterMap> {
+		let mut reader = BufReader::new(file.try_clone()?);
+		reader.seek(SeekFrom::Start(cmap.offset as u64))?;
+
+		reader.seek_relative(2)?; // Skip version because i dont care
+		let num_tables: u16 = deserialize(reader.by_ref())?;
+
+		let Some(table) = Self::find_unicode_table(&mut reader, num_tables) else {
+			return Err(Box::new(ErrorKind::Custom("No encoding for Unicode 2.0 BMP".into())));
+		};
+
+		reader.seek(SeekFrom::Start((cmap.offset + table.offset) as u64))?;
+		let format: u16 = deserialize(reader.by_ref())?;
+
+		if format != 4 {
+			todo!();
+		}
+
+		Ok(CharacterMap {
+			mapping_table: SegmentToDelta::new(reader.by_ref(), file)?,
+
+			_table_start: cmap.offset,
+			_length: cmap.length,
+
+		})
+	}
+	
+	pub fn get_char(&mut self, c: char) -> Option<u16> {
+		let code = c as u32;
+		if code > (u16::MAX as u32) {
+			warn!("Codepoint {code} for '{c}' is not a valid codepoint");
+			return None;
+		}
+
+		let code = (code & 0xFFFF) as u16;
+		debug!("Getting glyph id for codepoint {code}");
+		return self.mapping_table.get_glyph_id(code);
+	}
+}
--- a/src/font/search.rs
+++ b/src/font/search.rs
@ -0,0 +1,46 @@
+use std::fmt::Debug;
+
+use log::debug;
+use serde::Deserialize;
+
+#[derive(Debug, Copy, Clone, Deserialize, Default)]
+pub struct SearchParameters {
+	pub seg_count_x2: u16,		// num of elements x2
+	pub search_range: u16,		// largest power of two smaller than num of elements
+	pub entry_selector: u16,	// max levels in binary search
+	pub _range_shift: u16		// remaining elements in array
+}
+
+impl SearchParameters {
+	fn binary_search(&self, array: &[u16], target: u16) -> Option<usize> {
+		let mut index = (self.search_range >> 2) as usize;
+		for level in 1..=self.entry_selector {
+			let end_code = array[index];
+
+			if end_code == target {
+				return Some(index);
+			}
+
+			if end_code < target {
+				index += (self.search_range >> (level + 2)) as usize;
+			} else {
+				index -= (self.search_range >> (level + 2)) as usize;
+			}
+		}
+
+		Some(index + 1)
+	}
+
+	pub fn find_first_goe(&self, array: &[u16], target: u16) -> Option<usize> {
+		if target > array[(self.seg_count_x2 as usize >> 1) - 1] {
+			return None;
+		}
+
+		if target > array[(self.search_range as usize >> 1) - 1] {
+			debug!("{}", array[(self.search_range as usize >> 1) - 1]);
+			todo!()
+		}
+
+		self.binary_search(array, target)
+	}
+}
--- a/src/font/table_directory.rs
+++ b/src/font/table_directory.rs
@ -11,8 +11,8 @@ struct TableDirectoryHeader {
 	num_tables: u16,
 }

-#[derive(Deserialize, Debug)]
-struct TableDirectoryRecord {
+#[derive(Deserialize, Debug, Copy, Clone)]
+pub struct TableDirectoryRecord {
 	pub table_tag: u32,
 	pub checksum: u32,
 	pub offset: u32,
@ -36,4 +36,13 @@ impl TableDirectory {

 		Ok(TableDirectory(tables))
 	}
+
+	pub fn get_table(&self, table: &str) -> Option<TableDirectoryRecord> {
+		let table_id = table.as_bytes().first_chunk::<4>().unwrap().iter().fold(0u32, |running, &val| {
+			(running << 8) | (val as u32)
+		});
+
+		self.0.get(&table_id).copied()
+	}
+
 }
--- a/src/main.rs
+++ b/src/main.rs
@ -1,24 +1,45 @@
 mod font;

-use std::{fs::File, io::Result};
+use std::{error::Error, fs::File};

 use clap::Parser;
 use font::Font;
+use log::{error, LevelFilter};
+use simplelog::{ColorChoice, Config, TermLogger, TerminalMode};

 #[derive(Parser, Debug)]
 #[command(version, about, long_about = None)]
 struct Args {
-    file: String
+    file: String,
+    glyph: char
 }

-fn main() -> Result<()> {
+fn main() -> Result<(), Box<dyn Error>> {
+    TermLogger::init(
+        LevelFilter::Debug,
+        Config::default(),
+        TerminalMode::Mixed,
+        ColorChoice::Auto
+    )?;
+
    let args = Args::parse();
+
    let font = Font::new(File::open(args.file)?);
-    let Ok(font) = font else {
-        panic!("{}", font.unwrap_err().to_string());
+    let Ok(mut font) = font else {
+        let err = font.unwrap_err().to_string();
+        error!("{err}");
+        return Err(String::from(err).into());
    };

-    dbg!(font);
+    // dbg!(font);
+
+    match font.get_data_for(args.glyph) {
+        Ok(()) => {},
+        Err(err) => {
+            error!("{err}");
+            return Err(String::from(err).into());
+        }
+    };

    Ok(())
 }