buf_read_splitter

A stream reader with ability to read a stream until a defined pattern is reached (usually an array of [u8])

This could be a simple separator :

use std::io::Read;
use buf_read_splitter::{BufReadSplitter,MatchResult,Options,SimpleMatcher};

// To simulate a stream of this content :
let input = "First<SEP>Second<SEP>Third<SEP>Fourth<SEP>Fifth".to_string();
let mut input_reader = input.as_bytes();

// Create a reader that will end at each "<SEP>" :
let mut reader = BufReadSplitter::new(
           &mut input_reader,
           SimpleMatcher::new(b"<SEP>"),
           Options::default(),
);

// List of separate String will be listed in a Vector :
let mut words = Vec::new();

// Working variables :
let mut word = String::new();
let mut buf = vec![0u8; 100];

while {
 let sz = reader.read(&mut buf).unwrap();
 if sz > 0 {
    let to_str = String::from_utf8_lossy(&buf[..sz]);
    word.push_str(&to_str);
    true
 } else {
    words.push(word.clone());
    word.clear();
    match reader.next_part().unwrap() {  //Pass to the next part of the buffer
       Some(_) => true,     //There's a next part
       None => false,       //End of the stream
    }
 }
} {}

assert_eq!(&words[0], "First");
assert_eq!(&words[1], "Second");
assert_eq!(&words[2], "Third");
assert_eq!(&words[3], "Fourth");
assert_eq!(&words[4], "Fifth");
assert_eq!(words.len(), 5);

For more complexe pattern, the trait Matcher has to be implementing.
For example above a Matcher able to split a stream at each Mac, Unix or Windows end of line :

use buf_read_splitter::{
       MatchResult,
       Matcher,
       };

struct AllEndOfLineMatcher {
   prev_char: u8,
}
impl AllEndOfLineMatcher {
   pub fn new() -> Self {
       Self { prev_char: 0 }
   }
}
impl Matcher for AllEndOfLineMatcher {
   // This function is called at each byte read
   //   `el_buf` contains the value of the byte
   //   `pos` contains the position matched
   fn sequel(&mut self, el_buf: u8, pos: usize) -> MatchResult {
       if pos == 0 {
           if el_buf == b'\r' || el_buf == b'\n' {
               self.prev_char = el_buf;
               MatchResult::NeedNext
           } else {
               MatchResult::Mismatch
           }
       } else if pos == 1 {
           if el_buf == b'\n' && self.prev_char == b'\r' {
               //We are on \r\n
               MatchResult::Match(0, 0)
           } else {
               //Ignore the last byte (it's not a part of the end of line)
               MatchResult::Match(0, 1)
           }
       } else {
           //Unreachable
           panic!("We can't reach this code since we just manage 2 positions")
       }
   }

   // This function is called at the end of the buffer, useful to manage partial cases
   fn sequel_eos(&mut self, pos: usize) -> MatchResult {
       if pos == 0 {
           MatchResult::Match(0, 0) //Here the last char is \r or \n, at position 0
       } else {
           panic!("We can't reach this code since we just manage 2 positions")
       }
   }
}

...so the reader can be created like with this code :

let mut reader = BufReadSplitter::new(
                            &mut input_reader,
                            AllEndOfLineMatcher::new(),
                            Options::default()
                            );

The separator pattern can be changed on the fly by calling "matcher" function :

reader.matcher(SimpleMatcher::new(b"<CHANGE SEP>"))

The buffer part can be limited in size readed.
For example to limit to 100 bytes :

reader.set_limit_read(Some(100));

...and to reinitialize it to "no limit" :

reader.set_limit_read(None);

A call to ".next_part()" pass to the next part, however the end was reached or not (skips what has not been readed)\

For debug purpose, you can activate the "log" features in the Cargo.toml (slow down processing) :

[dependencies]
buf_read_splitter = {"0.4", features = ["log"] }

License: MIT

buf_read_splitter 0.4.1

buf_read_splitter