fix: increase signal in git diff, git log, and json filters (#621)

- git diff: raise max_hunk_lines from 30 to 100 (LLMs need full hunks)
- git log: show 3 body lines instead of 1 (preserves BREAKING CHANGE, migration notes)
- json: show values by default (LLMs need values for config debugging), add --schema for types-only

Tested with phi4:14b on local LLM — all 3 fixes improve comprehension.

Signed-off-by: Patrick szymkowiak <patrick.szymkowiak@innovtech.eu>

Author: pszymkowiak

src/filters/stat.toml

@@ -1,21 +1,17 @@
 [filters.stat]
-description = "Compact stat output — strip blank lines"
+description = "Compact stat output — strip device/inode/birth noise"
 match_command = "^stat\\b"
 strip_ansi = true
 strip_lines_matching = [
   "^\\s*$",
+  "^\\s*Device:",
+  "^\\s*Birth:",
 ]
-max_lines = 30
+truncate_lines_at = 120
+max_lines = 20
 
 [[tests.stat]]
-name = "macOS stat output kept"
-input = """
-16777234 8690244974 -rw-r--r-- 1 patrick staff 0 12345 "Mar 10 12:00:00 2026" "Mar 10 11:00:00 2026" "Mar 10 11:00:00 2026" "Mar  9 10:00:00 2026" 4096 24 0 file.txt
-"""
-expected = "16777234 8690244974 -rw-r--r-- 1 patrick staff 0 12345 \"Mar 10 12:00:00 2026\" \"Mar 10 11:00:00 2026\" \"Mar 10 11:00:00 2026\" \"Mar  9 10:00:00 2026\" 4096 24 0 file.txt"
-
-[[tests.stat]]
-name = "linux stat output kept"
+name = "linux stat output strips device and birth"
 input = """
   File: main.rs
   Size: 12345           Blocks: 24         IO Block: 4096   regular file
@@ -26,7 +22,21 @@ Modify: 2026-03-10 11:00:00.000000000 +0100
 Change: 2026-03-10 11:00:00.000000000 +0100
  Birth: 2026-03-09 10:00:00.000000000 +0100
 """
-expected = "  File: main.rs\n  Size: 12345           Blocks: 24         IO Block: 4096   regular file\nDevice: 801h/2049d      Inode: 1234567     Links: 1\nAccess: (0644/-rw-r--r--)  Uid: ( 1000/ patrick)   Gid: ( 1000/ patrick)\nAccess: 2026-03-10 12:00:00.000000000 +0100\nModify: 2026-03-10 11:00:00.000000000 +0100\nChange: 2026-03-10 11:00:00.000000000 +0100\n Birth: 2026-03-09 10:00:00.000000000 +0100"
+expected = "  File: main.rs\n  Size: 12345           Blocks: 24         IO Block: 4096   regular file\nAccess: (0644/-rw-r--r--)  Uid: ( 1000/ patrick)   Gid: ( 1000/ patrick)\nAccess: 2026-03-10 12:00:00.000000000 +0100\nModify: 2026-03-10 11:00:00.000000000 +0100\nChange: 2026-03-10 11:00:00.000000000 +0100"
+
+[[tests.stat]]
+name = "macOS stat -x strips device and birth"
+input = """
+  File: "main.rs"
+  Size: 82848        FileType: Regular File
+  Mode: (0644/-rw-r--r--)         Uid: (  501/ patrick)  Gid: (   20/   staff)
+Device: 1,15   Inode: 66302332    Links: 1
+Access: Wed Mar 18 21:21:15 2026
+Modify: Wed Mar 18 20:56:11 2026
+Change: Wed Mar 18 20:56:11 2026
+ Birth: Wed Mar 18 20:56:11 2026
+"""
+expected = "  File: \"main.rs\"\n  Size: 82848        FileType: Regular File\n  Mode: (0644/-rw-r--r--)         Uid: (  501/ patrick)  Gid: (   20/   staff)\nAccess: Wed Mar 18 21:21:15 2026\nModify: Wed Mar 18 20:56:11 2026\nChange: Wed Mar 18 20:56:11 2026"
 
 [[tests.stat]]
 name = "empty input passes through"

src/git.rs

@@ -297,7 +297,7 @@ pub(crate) fn compact_diff(diff: &str, max_lines: usize) -> String {
     let mut removed = 0;
     let mut in_hunk = false;
     let mut hunk_lines = 0;
-    let max_hunk_lines = 30;
+    let max_hunk_lines = 100;
     let mut was_truncated = false;
 
     for line in diff.lines() {
@@ -532,17 +532,25 @@ fn filter_log_output(
             Some(h) => truncate_line(h.trim(), truncate_width),
             None => continue,
         };
-        // Remaining lines are the body — keep first non-empty line only
-        let body_line = lines.map(|l| l.trim()).find(|l| {
-            !l.is_empty() && !l.starts_with("Signed-off-by:") && !l.starts_with("Co-authored-by:")
-        });
-
-        match body_line {
-            Some(body) => {
-                let truncated_body = truncate_line(body, truncate_width);
-                result.push(format!("{}\n  {}", header, truncated_body));
+        // Remaining lines are the body — keep up to 3 non-empty, non-trailer lines
+        let body_lines: Vec<&str> = lines
+            .map(|l| l.trim())
+            .filter(|l| {
+                !l.is_empty()
+                    && !l.starts_with("Signed-off-by:")
+                    && !l.starts_with("Co-authored-by:")
+            })
+            .take(3)
+            .collect();
+
+        if body_lines.is_empty() {
+            result.push(header);
+        } else {
+            let mut entry = header;
+            for body in &body_lines {
+                entry.push_str(&format!("\n  {}", truncate_line(body, truncate_width)));
             }
-            None => result.push(header),
+            result.push(entry);
         }
     }
 

src/json_cmd.rs

@@ -33,8 +33,8 @@ fn validate_json_extension(file: &Path) -> Result<()> {
     Ok(())
 }
 
-/// Show JSON structure without values
-pub fn run(file: &Path, max_depth: usize, verbose: u8) -> Result<()> {
+/// Show JSON (compact with values, or schema-only with --schema)
+pub fn run(file: &Path, max_depth: usize, schema_only: bool, verbose: u8) -> Result<()> {
     validate_json_extension(file)?;
     let timer = tracking::TimedExecution::start();
 
@@ -45,19 +45,23 @@ pub fn run(file: &Path, max_depth: usize, verbose: u8) -> Result<()> {
     let content = fs::read_to_string(file)
         .with_context(|| format!("Failed to read file: {}", file.display()))?;
 
-    let schema = filter_json_string(&content, max_depth)?;
-    println!("{}", schema);
+    let output = if schema_only {
+        filter_json_string(&content, max_depth)?
+    } else {
+        filter_json_compact(&content, max_depth)?
+    };
+    println!("{}", output);
     timer.track(
         &format!("cat {}", file.display()),
         "rtk json",
         &content,
-        &schema,
+        &output,
     );
     Ok(())
 }
 
-/// Show JSON structure from stdin
-pub fn run_stdin(max_depth: usize, verbose: u8) -> Result<()> {
+/// Show JSON from stdin
+pub fn run_stdin(max_depth: usize, schema_only: bool, verbose: u8) -> Result<()> {
     let timer = tracking::TimedExecution::start();
 
     if verbose > 0 {
@@ -70,13 +74,107 @@ pub fn run_stdin(max_depth: usize, verbose: u8) -> Result<()> {
         .read_to_string(&mut content)
         .context("Failed to read from stdin")?;
 
-    let schema = filter_json_string(&content, max_depth)?;
-    println!("{}", schema);
-    timer.track("cat - (stdin)", "rtk json -", &content, &schema);
+    let output = if schema_only {
+        filter_json_string(&content, max_depth)?
+    } else {
+        filter_json_compact(&content, max_depth)?
+    };
+    println!("{}", output);
+    timer.track("cat - (stdin)", "rtk json -", &content, &output);
     Ok(())
 }
 
-/// Parse a JSON string and return its schema representation.
+/// Parse a JSON string and return compact representation with values preserved.
+/// Long strings are truncated, arrays are summarized.
+pub fn filter_json_compact(json_str: &str, max_depth: usize) -> Result<String> {
+    let value: Value = serde_json::from_str(json_str).context("Failed to parse JSON")?;
+    Ok(compact_json(&value, 0, max_depth))
+}
+
+fn compact_json(value: &Value, depth: usize, max_depth: usize) -> String {
+    let indent = "  ".repeat(depth);
+
+    if depth > max_depth {
+        return format!("{}...", indent);
+    }
+
+    match value {
+        Value::Null => format!("{}null", indent),
+        Value::Bool(b) => format!("{}{}", indent, b),
+        Value::Number(n) => format!("{}{}", indent, n),
+        Value::String(s) => {
+            if s.len() > 80 {
+                format!("{}\"{}...\"", indent, &s[..77])
+            } else {
+                format!("{}\"{}\"", indent, s)
+            }
+        }
+        Value::Array(arr) => {
+            if arr.is_empty() {
+                format!("{}[]", indent)
+            } else if arr.len() > 5 {
+                let first = compact_json(&arr[0], depth + 1, max_depth);
+                format!("{}[{}, ... +{} more]", indent, first.trim(), arr.len() - 1)
+            } else {
+                let items: Vec<String> = arr
+                    .iter()
+                    .map(|v| compact_json(v, depth + 1, max_depth))
+                    .collect();
+                let all_simple = arr.iter().all(|v| {
+                    matches!(
+                        v,
+                        Value::Null | Value::Bool(_) | Value::Number(_) | Value::String(_)
+                    )
+                });
+                if all_simple {
+                    let inline: Vec<&str> = items.iter().map(|s| s.trim()).collect();
+                    format!("{}[{}]", indent, inline.join(", "))
+                } else {
+                    let mut lines = vec![format!("{}[", indent)];
+                    for item in &items {
+                        lines.push(format!("{},", item));
+                    }
+                    lines.push(format!("{}]", indent));
+                    lines.join("\n")
+                }
+            }
+        }
+        Value::Object(map) => {
+            if map.is_empty() {
+                format!("{}{{}}", indent)
+            } else {
+                let mut lines = vec![format!("{}{{", indent)];
+                let mut keys: Vec<_> = map.keys().collect();
+                keys.sort();
+
+                for (i, key) in keys.iter().enumerate() {
+                    let val = &map[*key];
+                    let is_simple = matches!(
+                        val,
+                        Value::Null | Value::Bool(_) | Value::Number(_) | Value::String(_)
+                    );
+
+                    if is_simple {
+                        let val_str = compact_json(val, 0, max_depth);
+                        lines.push(format!("{}  {}: {}", indent, key, val_str.trim()));
+                    } else {
+                        lines.push(format!("{}  {}:", indent, key));
+                        lines.push(compact_json(val, depth + 1, max_depth));
+                    }
+
+                    if i >= 20 {
+                        lines.push(format!("{}  ... +{} more keys", indent, keys.len() - i - 1));
+                        break;
+                    }
+                }
+                lines.push(format!("{}}}", indent));
+                lines.join("\n")
+            }
+        }
+    }
+}
+
+/// Parse a JSON string and return its schema representation (types only, no values).
 /// Useful for piping JSON from other commands (e.g., `gh api`, `curl`).
 pub fn filter_json_string(json_str: &str, max_depth: usize) -> Result<String> {
     let value: Value = serde_json::from_str(json_str).context("Failed to parse JSON")?;

src/main.rs

@@ -237,13 +237,16 @@ enum Commands {
         command: Vec<String>,
     },
 
-    /// Show JSON structure without values
+    /// Show JSON (compact values, or schema-only with --schema)
     Json {
         /// JSON file
         file: PathBuf,
         /// Max depth
         #[arg(short, long, default_value = "5")]
         depth: usize,
+        /// Show structure only (strip all values)
+        #[arg(long)]
+        schema: bool,
     },
 
     /// Summarize project dependencies
@@ -387,9 +390,9 @@ enum Commands {
     Wget {
         /// URL to download
         url: String,
-        /// Output to stdout instead of file
-        #[arg(short = 'O', long)]
-        stdout: bool,
+        /// Output file (-O - for stdout)
+        #[arg(short = 'O', long = "output-document", allow_hyphen_values = true)]
+        output: Option<String>,
         /// Additional wget arguments
         #[arg(trailing_var_arg = true, allow_hyphen_values = true)]
         args: Vec<String>,
@@ -1501,11 +1504,15 @@ fn main() -> Result<()> {
             runner::run_test(&cmd, cli.verbose)?;
         }
 
-        Commands::Json { file, depth } => {
+        Commands::Json {
+            file,
+            depth,
+            schema,
+        } => {
             if file == Path::new("-") {
-                json_cmd::run_stdin(depth, cli.verbose)?;
+                json_cmd::run_stdin(depth, schema, cli.verbose)?;
             } else {
-                json_cmd::run(&file, depth, cli.verbose)?;
+                json_cmd::run(&file, depth, schema, cli.verbose)?;
             }
         }
 
@@ -1702,11 +1709,18 @@ fn main() -> Result<()> {
             }
         }
 
-        Commands::Wget { url, stdout, args } => {
-            if stdout {
+        Commands::Wget { url, output, args } => {
+            if output.as_deref() == Some("-") {
                 wget_cmd::run_stdout(&url, &args, cli.verbose)?;
             } else {
-                wget_cmd::run(&url, &args, cli.verbose)?;
+                // Pass -O <file> through to wget via args
+                let mut all_args = Vec::new();
+                if let Some(out_file) = &output {
+                    all_args.push("-O".to_string());
+                    all_args.push(out_file.clone());
+                }
+                all_args.extend(args);
+                wget_cmd::run(&url, &all_args, cli.verbose)?;
             }
         }