Merge pull request #7 from jnises/nises/window-penalty

Use window size to determine age penalty

.github/workflows/test.yml → .github/workflows/CI.yml

@@ -1,4 +1,4 @@
-name: Test
+name: CI
 
 on:
   push:
@@ -14,23 +14,29 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-      - uses: actions/checkout@v2
-      - uses: actions-rs/toolchain@v1
+      - name: checkout
+        uses: actions/checkout@v2
+      - name: install toolchain
+        uses: actions-rs/toolchain@v1
         with:
           profile: minimal
           toolchain: stable
           components: clippy, rustfmt
-      - uses: actions-rs/cargo@v1
+      - name: check
+        uses: actions-rs/cargo@v1
         with:
           command: check
-      - uses: actions-rs/cargo@v1
+      - name: test
+        uses: actions-rs/cargo@v1
         with:
           command: test
-      - uses: actions-rs/cargo@v1
+      - name: formatting
+        uses: actions-rs/cargo@v1
         with:
           command: fmt
           args: --all -- --check
-      - uses: actions-rs/cargo@v1
+      - name: clippy
+        uses: actions-rs/cargo@v1
         with:
           command: clippy
           args: -- -D warnings

Cargo.toml

@@ -1,6 +1,6 @@
 [package]
 name = "colight"
-version = "0.3.0"
+version = "0.4.0"
 edition = "2021"
 
 [dependencies]

README.md

@@ -3,11 +3,11 @@ Colight
 
 ![screenshot](docs/screenshot.webp)
 
-Command line utility that colors text based on how compressible it is.
+Command line utility that highlights text based on how compressible it is.
 Compressibility is determined using a lz77-like algorithm.
 The idea is that compressibility is a good indicator of whether a piece of text is interesting or not.
 
-It only works in some terminals.
+It only works in true-color terminals.
 
 ## Usage
 

src/main.rs

@@ -17,19 +17,16 @@ use crate::{
 #[derive(Parser, Debug)]
 #[command(author, version, about, long_about = None)]
 struct Args {
-    /// How long back in history to look for matches
+    /// How long back in history to look for matches.
+    ///
+    /// The further back in history a match is found, the lower its compressibility score is.
     #[arg(long, default_value_t = 1024)]
     window_size: usize,
-
-    /// How much to penalize age of matches, higher values will make older matches less likely to be highlighted
-    #[arg(long, default_value_t = 0.01)]
-    age_penalty: f32,
 }
 
 fn main() -> anyhow::Result<()> {
     let args = Args::parse();
     assert!(args.window_size > 0);
-    assert!(args.age_penalty >= 0.0);
     let stdin = std::io::stdin();
     let si = stdin.lock();
     let stdout = StandardStream::stdout(ColorChoice::Auto);
@@ -43,18 +40,17 @@ where
     O: WriteColor,
     I: Read,
 {
-    let Args {
-        window_size,
-        age_penalty,
-    } = args;
+    let Args { window_size } = args;
     // refcell so we can reset on scope exit
     let soc = RefCell::new(so);
     defer! {
         soc.borrow_mut().reset().unwrap();
     }
-    let pr = |buffered: VecDeque<u8>, age: Option<usize>| -> anyhow::Result<()> {
+    let pr = |buffered: VecDeque<u8>, score: f32| -> anyhow::Result<()> {
+        // score of 1 means buffered is uncompressible, 0 means it is fully compressible
+        debug_assert!((0.0..=1.0).contains(&score));
         if !buffered.is_empty() {
-            let score = 1f32 / (buffered.len() as f32 + age.unwrap_or(0) as f32 * age_penalty);
+            //let score = 1f32 / (buffered.len() as f32 + age.unwrap_or(0) as f32 * age_penalty);
             let (a, b) = buffered.as_slices();
             for line in a
                 .split_inclusive(|&c| c == b'\n')
@@ -73,17 +69,19 @@ where
     loop {
         let mut byte_buf = [0; 1];
         if color_stripper.read_exact(&mut byte_buf).is_err() {
-            pr(searcher.flush(), None)?;
+            pr(searcher.flush(), 0.0)?;
             break;
         };
         // keep going until we find no more matches
         match searcher.search(byte_buf[0]) {
             SearchState::Buffering => {}
-            SearchState::Flushed {
-                buffer: last_found_needle,
-                age,
-            } => {
-                pr(last_found_needle, age)?;
+            SearchState::Flushed { buffer, age } => {
+                let score = 1.0
+                    / 1f32.max(
+                        buffer.len() as f32
+                            / (1.0 - (age.unwrap_or(0) as f32 / window_size as f32)),
+                    );
+                pr(buffer, score)?;
             }
         }
     }
@@ -133,10 +131,7 @@ mod tests {
 [Mon Mar 1 09:20:01 CET 2021] start new app: /Applications/app.app",
             ),
             &mut s,
-            Args {
-                window_size: 1024,
-                age_penalty: 0.0,
-            },
+            Args { window_size: 1024 },
         )
         .unwrap();
     }
@@ -153,10 +148,7 @@ ab
 ",
             ),
             &mut s,
-            Args {
-                window_size: 1024,
-                age_penalty: 0.0,
-            },
+            Args { window_size: 1024 },
         )
         .unwrap();
     }