Merge pull request #20 from pamburus/feature/regex

new: field filter with regular expressions

Cargo.toml

@@ -4,7 +4,7 @@ categories = ["command-line-utilities"]
 description = "Utility for viewing json-formatted log files."
 keywords = ["cli", "human", "log"]
 name = "hl"
-version = "0.10.1"
+version = "0.10.2"
 edition = "2018"
 
 # See more keys and their definitions at https://doc.rust-lang.org/cargo/reference/manifest.html
@@ -34,6 +34,7 @@ itertools = "0"
 num_cpus = "1"
 once_cell = "1"
 platform-dirs = "0"
+regex = "1"
 rust-embed = "5"
 serde = { version = "1", features = ["derive"] }
 serde_json = { version = "1", features = ["raw_value"] }
@@ -94,3 +95,7 @@ harness = false
 [[bench]]
 name = "parse-and-format"
 harness = false
+
+[[bench]]
+name = "json"
+harness = false

README.md

@@ -297,7 +297,7 @@ Log viewer which translates JSON logs into pretty human-readable representation.
 ### Complete set of options and flags
 
 ```
-hl 0.10.0
+hl 0.10.2
 JSON log converter to human readable representation
 
 USAGE:
@@ -316,7 +316,7 @@ OPTIONS:
         --buffer-size <buffer-size>                          Buffer size [env: HL_BUFFER_SIZE=]  [default: 2 MiB]
         --color <color>                                      Color output options, one of { auto, always, never } [env: HL_COLOR=]  [default: auto]
     -C, --concurrency <concurrency>                          Number of processing threads [env: HL_CONCURRENCY=]
-    -f, --filter <filter>...                                 Filtering by field values in one of forms <key>=<value>, <key>~=<value>, <key>!=<value>, <key>!~=<value>
+    -f, --filter <filter>...                                 Filtering by field values in one of forms [<key>=<value>, <key>~=<value>, <key>~~=<value>, <key>!=<value>, <key>!~=<value>, <key>!~~=<value>] where ~ denotes substring match and ~~ denotes regular expression match
     -h, --hide <hide>...                                     Hide fields with the specified keys
     -e, --hide-empty-fields <hide-empty-fields>              Hide empty fields, applies for null, string, object and array fields only [env: HL_HIDE_EMPTY_FIELDS=]
         --interrupt-ignore-count <interrupt-ignore-count>    Number of interrupts to ignore, i.e. Ctrl-C (SIGINT) [env: HL_INTERRUPT_IGNORE_COUNT=]  [default: 3]

benches/json.rs

@@ -0,0 +1,46 @@
+// std imports
+use std::alloc::System;
+
+// third-party imports
+use criterion::{criterion_group, criterion_main, Criterion};
+use serde_json as json;
+use stats_alloc::{Region, StatsAlloc, INSTRUMENTED_SYSTEM};
+
+#[global_allocator]
+static GLOBAL: &StatsAlloc<System> = &INSTRUMENTED_SYSTEM;
+
+fn benchmark(c: &mut Criterion) {
+    let mut c = c.benchmark_group("json");
+
+    let mut c1 = None;
+    let mut n1 = 0;
+    c.bench_function("parse-to-str", |b| {
+        let sample = r#""test-message""#;
+        let reg = Region::new(&GLOBAL);
+        b.iter(|| {
+            assert_eq!(json::from_str::<&str>(sample).unwrap(), "test-message");
+            n1 += 1;
+        });
+        c1 = Some(reg.change());
+    });
+    println!("allocations at 1 ({:?} iterations): {:#?}", n1, c1);
+
+    let mut c2 = None;
+    let mut n2 = 0;
+    c.bench_function("parse-to-string", |b| {
+        let sample = r#""test-\"message\"""#;
+        let reg = Region::new(&GLOBAL);
+        b.iter(|| {
+            assert_eq!(
+                json::from_str::<String>(sample).unwrap(),
+                r#"test-"message""#
+            );
+            n2 += 1;
+        });
+        c2 = Some(reg.change());
+    });
+    println!("allocations at 2 ({:?} iterations): {:#?}", n2, c2);
+}
+
+criterion_group!(benches, benchmark);
+criterion_main!(benches);

src/error.rs

@@ -49,6 +49,10 @@ pub enum Error {
     Utf8Error(#[from] std::str::Utf8Error),
     #[error("failed to parse yaml: {0}")]
     YamlError(#[from] serde_yaml::Error),
+    #[error("wrong field filter format: {0}")]
+    WrongFieldFilter(String),
+    #[error("wrong regular expression: {0}")]
+    WrongRegularExpression(#[from] regex::Error),
 }
 
 /// Result is an alias for standard result with bound Error type.

src/lib.rs

@@ -18,6 +18,7 @@ mod eseq;
 mod filtering;
 mod formatting;
 mod model;
+mod pool;
 mod scanning;
 
 // conditional public modules

src/main.rs

@@ -104,7 +104,7 @@ struct Opt {
     )]
     concurrency: Option<usize>,
     //
-    /// Filtering by field values in one of forms <key>=<value>, <key>~=<value>, <key>!=<value>, <key>!~=<value>.
+    /// Filtering by field values in one of forms [<key>=<value>, <key>~=<value>, <key>~~=<value>, <key>!=<value>, <key>!~=<value>, <key>!~~=<value>] where ~ denotes substring match and ~~ denotes regular expression match.
     #[structopt(short, long, number_of_values = 1)]
     filter: Vec<String>,
     //
@@ -283,7 +283,7 @@ fn run() -> Result<()> {
     let time_format = LinuxDateFormat::new(&opt.time_format).compile();
     // Configure filter.
     let filter = hl::Filter {
-        fields: hl::FieldFilterSet::new(opt.filter),
+        fields: hl::FieldFilterSet::new(opt.filter)?,
         level: opt.level,
         since: if let Some(v) = &opt.since {
             Some(parse_time(v, &tz, &time_format)?.into())