add remaining encoding config options + presets for x86 and z80

Cargo.lock

@@ -117,7 +117,7 @@ dependencies = [
 
 [[package]]
 name = "upkr"
-version = "0.2.0"
+version = "0.2.0-pre1"
 dependencies = [
  "anyhow",
  "cdivsufsort",

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "upkr"
-version = "0.2.0"
+version = "0.2.0-pre1"
 edition = "2021"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html

src/context_state.rs

@@ -1,4 +1,7 @@
-use crate::rans::{PROB_BITS, ONE_PROB};
+use crate::{
+    rans::{ONE_PROB, PROB_BITS},
+    Config,
+};
 
 const INIT_PROB: u16 = 1 << (PROB_BITS - 1);
 const UPDATE_RATE: u32 = 4;
@@ -7,6 +10,8 @@ const UPDATE_ADD: u32 = 8;
 #[derive(Clone)]
 pub struct ContextState {
     contexts: Vec<u8>,
+    invert_bit_encoding: bool,
+    simplified_prob_update: bool,
 }
 
 pub struct Context<'a> {
@@ -15,9 +20,11 @@ pub struct Context<'a> {
 }
 
 impl ContextState {
-    pub fn new(size: usize) -> ContextState {
+    pub fn new(size: usize, config: &Config) -> ContextState {
         ContextState {
             contexts: vec![INIT_PROB as u8; size],
+            invert_bit_encoding: config.invert_bit_encoding,
+            simplified_prob_update: config.simplified_prob_update,
         }
     }
 
@@ -33,10 +40,21 @@ impl<'a> Context<'a> {
 
     pub fn update(&mut self, bit: bool) {
         let old = self.state.contexts[self.index];
-        self.state.contexts[self.index] = if bit {
-            old + ((ONE_PROB - old as u32 + UPDATE_ADD) >> UPDATE_RATE) as u8
+
+        self.state.contexts[self.index] = if self.state.simplified_prob_update {
+            let offset = if bit ^ self.state.invert_bit_encoding {
+                ONE_PROB as i32 >> UPDATE_RATE
+            } else {
+                0
+            };
+
+            (offset + old as i32 - ((old as i32 + UPDATE_ADD as i32) >> UPDATE_RATE)) as u8
         } else {
-            old - ((old as u32 + UPDATE_ADD) >> UPDATE_RATE) as u8
+            if bit ^ self.state.invert_bit_encoding {
+                old + ((ONE_PROB - old as u32 + UPDATE_ADD) >> UPDATE_RATE) as u8
+            } else {
+                old - ((old as u32 + UPDATE_ADD) >> UPDATE_RATE) as u8
+            }
         };
     }
 }

src/greedy_packer.rs

@@ -9,8 +9,8 @@ pub fn pack(
     mut progress_callback: Option<ProgressCallback>,
 ) -> Vec<u8> {
     let mut match_finder = MatchFinder::new(data);
-    let mut rans_coder = RansCoder::new(config.use_bitstream);
-    let mut state = lz::CoderState::new(config.parity_contexts);
+    let mut rans_coder = RansCoder::new(config);
+    let mut state = lz::CoderState::new(config);
 
     let mut pos = 0;
     while pos < data.len() {

src/lib.rs

@@ -13,10 +13,13 @@ pub struct Config {
     pub use_bitstream: bool,
     pub parity_contexts: usize,
 
-    pub invert_probs: bool,
+    pub invert_bit_encoding: bool,
     pub is_match_bit: bool,
     pub new_offset_bit: bool,
     pub continue_value_bit: bool,
+
+    pub bitstream_is_big_endian: bool,
+    pub simplified_prob_update: bool,
 }
 
 impl Default for Config {
@@ -25,10 +28,13 @@ impl Default for Config {
             use_bitstream: false,
             parity_contexts: 1,
 
-            invert_probs: false,
+            invert_bit_encoding: false,
             is_match_bit: true,
             new_offset_bit: true,
             continue_value_bit: true,
+
+            bitstream_is_big_endian: false,
+            simplified_prob_update: false,
         }
     }
 }

src/lz.rs

@@ -102,20 +102,20 @@ fn encode_length(
 #[derive(Clone)]
 pub struct CoderState {
     contexts: ContextState,
-    parity_contexts: usize,
     last_offset: u32,
     prev_was_match: bool,
     pos: usize,
+    parity_contexts: usize,
 }
 
 impl CoderState {
-    pub fn new(parity_contexts: usize) -> CoderState {
+    pub fn new(config: &Config) -> CoderState {
         CoderState {
-            contexts: ContextState::new((1 + 255) * parity_contexts + 1 + 64 + 64),
+            contexts: ContextState::new((1 + 255) * config.parity_contexts + 1 + 64 + 64, config),
             last_offset: 0,
-            parity_contexts,
             prev_was_match: false,
             pos: 0,
+            parity_contexts: config.parity_contexts,
         }
     }
 
@@ -125,8 +125,8 @@ impl CoderState {
 }
 
 pub fn unpack(packed_data: &[u8], config: Config) -> Vec<u8> {
-    let mut decoder = RansDecoder::new(packed_data, config.use_bitstream);
-    let mut contexts = ContextState::new((1 + 255) * config.parity_contexts + 1 + 64 + 64);
+    let mut decoder = RansDecoder::new(packed_data, &config);
+    let mut contexts = ContextState::new((1 + 255) * config.parity_contexts + 1 + 64 + 64, &config);
     let mut result = vec![];
     let mut offset = 0;
     let mut prev_was_match = false;

src/main.rs

@@ -22,7 +22,24 @@ fn main() -> Result<()> {
             Long("invert-is-match-bit") => config.is_match_bit = false,
             Long("invert-new-offset-bit") => config.new_offset_bit = false,
             Long("invert-continue-value-bit") => config.continue_value_bit = false,
-            Long("invert-probs") => config.invert_probs = true,
+            Long("invert-bit-encoding") => config.invert_bit_encoding = true,
+            Long("simplified-prob-update") => config.simplified_prob_update = true,
+            Long("big-endian-bitstream") => {
+                config.use_bitstream = true;
+                config.bitstream_is_big_endian = true;
+            }
+
+            Long("z80") => {
+                config.use_bitstream = true;
+                config.bitstream_is_big_endian = true;
+                config.invert_bit_encoding = true;
+                config.simplified_prob_update = true;
+            }
+            Long("x86") => {
+                config.use_bitstream = true;
+                config.continue_value_bit = false;
+                config.is_match_bit = false;
+            }
 
             Short('u') | Long("unpack") => unpack = true,
             Short('l') | Long("level") => level = parser.value()?.parse()?,
@@ -112,6 +129,10 @@ fn print_help(exit_code: i32) -> ! {
     eprintln!(" -l, --level N       compression level 0-9");
     eprintln!(" -u, --unpack        unpack infile");
     eprintln!();
+    eprintln!("Config presets for specific unpackers:");
+    eprintln!(" --z80               --big-endian-bitstream --invert-bit-encoding --simplified-prob-update");
+    eprintln!(" --x86               --bitstream --invert-is-match-bit --invert-continue-value-bit");
+    eprintln!();
     eprintln!("Config options (need to match when packing/unpacking):");
     eprintln!(" -b, --bitstream     bitstream mode");
     eprintln!(" -p, --parity N      use N (2/4) parity contexts");
@@ -121,6 +142,8 @@ fn print_help(exit_code: i32) -> ! {
     eprintln!(" --invert-is-match-bit");
     eprintln!(" --invert-new-offset-bit");
     eprintln!(" --invert-continue-value-bit");
-    eprintln!(" --invert-probs");
+    eprintln!(" --invert-bit-encoding");
+    eprintln!(" --simplified-prob-update");
+    eprintln!(" --big-endian-bitstream   (implies --bitstream)");
     process::exit(exit_code);
 }

src/parsing_packer.rs

@@ -18,8 +18,8 @@ pub fn pack(
         ops.push(link.op);
         parse = link.prev.clone();
     }
-    let mut state = lz::CoderState::new(config.parity_contexts);
-    let mut coder = RansCoder::new(config.use_bitstream);
+    let mut state = lz::CoderState::new(config);
+    let mut coder = RansCoder::new(config);
     for op in ops.into_iter().rev() {
         op.encode(&mut coder, &mut state, config);
     }
@@ -136,13 +136,13 @@ fn parse(
         0,
         Arrival {
             parse: None,
-            state: lz::CoderState::new(encoding_config.parity_contexts),
+            state: lz::CoderState::new(encoding_config),
             cost: 0.0,
         },
         max_arrivals,
     );
 
-    let cost_counter = &mut CostCounter::new();
+    let cost_counter = &mut CostCounter::new(encoding_config);
     let mut best_per_offset = HashMap::new();
     for pos in 0..data.len() {
         let match_length = |offset: usize| {

src/rans.rs

@@ -1,4 +1,4 @@
-use crate::context_state::Context;
+use crate::{context_state::Context, Config};
 
 pub const PROB_BITS: u32 = 8;
 pub const ONE_PROB: u32 = 1 << PROB_BITS;
@@ -15,20 +15,25 @@ pub trait EntropyCoder {
 pub struct RansCoder {
     bits: Vec<u16>,
     use_bitstream: bool,
+    bitstream_is_big_endian: bool,
+    invert_bit_encoding: bool,
 }
 
 impl EntropyCoder for RansCoder {
     fn encode_bit(&mut self, bit: bool, prob: u16) {
         assert!(prob < 32768);
-        self.bits.push(prob | ((bit as u16) << 15));
+        self.bits
+            .push(prob | (((bit ^ self.invert_bit_encoding) as u16) << 15));
     }
 }
 
 impl RansCoder {
-    pub fn new(use_bitstream: bool) -> RansCoder {
+    pub fn new(config: &Config) -> RansCoder {
         RansCoder {
             bits: Vec::new(),
-            use_bitstream,
+            use_bitstream: config.use_bitstream,
+            bitstream_is_big_endian: config.bitstream_is_big_endian,
+            invert_bit_encoding: config.invert_bit_encoding,
         }
     }
 
@@ -38,18 +43,31 @@ impl RansCoder {
         let mut state = 1 << l_bits;
 
         let mut byte = 0u8;
-        let mut bit = 8;
+        let mut bit = if self.bitstream_is_big_endian { 0 } else { 8 };
         let mut flush_state: Box<dyn FnMut(&mut u32)> = if self.use_bitstream {
-            Box::new(|state: &mut u32| {
-                bit -= 1;
-                byte |= ((*state & 1) as u8) << bit;
-                if bit == 0 {
-                    buffer.push(byte);
-                    byte = 0;
-                    bit = 8;
-                }
-                *state >>= 1;
-            })
+            if self.bitstream_is_big_endian {
+                Box::new(|state: &mut u32| {
+                    byte |= ((*state & 1) as u8) << bit;
+                    bit += 1;
+                    if bit == 8 {
+                        buffer.push(byte);
+                        byte = 0;
+                        bit = 0;
+                    }
+                    *state >>= 1;
+                })
+            } else {
+                Box::new(|state: &mut u32| {
+                    bit -= 1;
+                    byte |= ((*state & 1) as u8) << bit;
+                    if bit == 0 {
+                        buffer.push(byte);
+                        byte = 0;
+                        bit = 8;
+                    }
+                    *state >>= 1;
+                })
+            }
         } else {
             Box::new(|state: &mut u32| {
                 buffer.push(*state as u8);
@@ -91,10 +109,11 @@ impl RansCoder {
 pub struct CostCounter {
     cost: f64,
     log2_table: Vec<f64>,
+    invert_bit_encoding: bool,
 }
 
 impl CostCounter {
-    pub fn new() -> CostCounter {
+    pub fn new(config: &Config) -> CostCounter {
         let log2_table = (0..ONE_PROB)
             .map(|prob| {
                 let inv_prob = ONE_PROB as f64 / prob as f64;
@@ -104,6 +123,7 @@ impl CostCounter {
         CostCounter {
             cost: 0.0,
             log2_table,
+            invert_bit_encoding: config.invert_bit_encoding,
         }
     }
 
@@ -118,7 +138,7 @@ impl CostCounter {
 
 impl EntropyCoder for CostCounter {
     fn encode_bit(&mut self, bit: bool, prob: u16) {
-        let prob = if bit {
+        let prob = if bit ^ self.invert_bit_encoding {
             prob as u32
         } else {
             ONE_PROB - prob as u32
@@ -133,18 +153,22 @@ pub struct RansDecoder<'a> {
     use_bitstream: bool,
     byte: u8,
     bits_left: u8,
+    invert_bit_encoding: bool,
+    bitstream_is_big_endian: bool,
 }
 
 const PROB_MASK: u32 = ONE_PROB - 1;
 
 impl<'a> RansDecoder<'a> {
-    pub fn new(data: &'a [u8], use_bitstream: bool) -> RansDecoder<'a> {
+    pub fn new(data: &'a [u8], config: &Config) -> RansDecoder<'a> {
         RansDecoder {
             data,
             state: 0,
-            use_bitstream,
+            use_bitstream: config.use_bitstream,
             byte: 0,
             bits_left: 0,
+            invert_bit_encoding: config.invert_bit_encoding,
+            bitstream_is_big_endian: config.bitstream_is_big_endian,
         }
     }
 
@@ -163,8 +187,13 @@ impl<'a> RansDecoder<'a> {
                     self.data = &self.data[1..];
                     self.bits_left = 8;
                 }
-                self.state = (self.state << 1) | (self.byte & 1) as u32;
-                self.byte >>= 1;
+                if self.bitstream_is_big_endian {
+                    self.state = (self.state << 1) | (self.byte >> 7) as u32;
+                    self.byte <<= 1;
+                } else {
+                    self.state = (self.state << 1) | (self.byte & 1) as u32;
+                    self.byte >>= 1;
+                }
                 self.bits_left -= 1;
             }
         } else {
@@ -183,6 +212,6 @@ impl<'a> RansDecoder<'a> {
         };
         self.state = prob * (self.state >> PROB_BITS) + (self.state & PROB_MASK) - start;
 
-        bit
+        bit ^ self.invert_bit_encoding
     }
 }