srtune/src/main.rs

#[macro_use]
extern crate failure;
#[macro_use]
extern crate log;
#[macro_use]
extern crate lazy_static;


use regex::Regex;
use std::path::Path;
use std::fs::File;
use std::io::{Read, Seek, SeekFrom};
use std::io;
use std::str::FromStr;

fn main() {
    let argv =
        clap::App::new("srtune")
            .version(env!("CARGO_PKG_VERSION"))
            .arg(clap::Arg::with_name("input")
                    .value_name("INFILE")
                     .required(true)
                     .index(1)
                    .help("Input file"),
            )
            .arg(clap::Arg::with_name("output")
                     .value_name("OUTFILE")
                     .required(true)
                     .index(2)
                     .help("Output file"),
            )
            .get_matches();

    let inf = argv.value_of("input").unwrap();
    let outf = argv.value_of("output").unwrap();

    let source = read_file(inf);

    let mut lines = source.lines();

    let mut subs = vec![];

    while let Some(x) = lines.next() {
        if x.is_empty() {
            continue;
        }

        // 236
        // 00:18:01,755 --> 00:18:03,774
        // （掃除機の音）
        // う～ん…。
        if let Ok(num) = u32::from_str(x) {
           // println!("Entry {}", num);
            let datesrow = lines.next().expect("expected date row");
            if datesrow.contains(" --> ") {
                let mut halves = datesrow.split(" --> ");
                let start = parse_time(halves.next().expect("expected two halves")).expect("invalid time");
                let end = parse_time(halves.next().expect("expected two halves")).expect("invalid time");

                //println!("{} -> {} secs", start, end);

                let mut text = vec![];
                while let Some(x) = lines.next() {
                    if x.is_empty() {
                        break;
                    }
                    text.push(x);
                }

                let text = text.join("\n");

                //println!("Lines: {}", text);

                subs.push(Subtitle {
                    num, start, end, text
                })
            }
        }
    }

    println!("Parsed {} entries.", subs.len());

    //println!("{:?}", parsed);
}

struct Subtitle {
    num : u32,
    start: f32,
    end: f32,
    text: String
}

lazy_static! {
    static ref DATE_RE: Regex = Regex::new(r"(\d+):(\d+):(\d+),(\d+)").unwrap();
}

fn parse_time(time : &str) -> Result<f32, ()> {
    // 00:18:01,755
    match DATE_RE.captures(time) {
        Some(caps) => {
            Ok(f32::from_str(caps.get(1).unwrap().as_str()).unwrap()*3600f32 +
                f32::from_str(caps.get(2).unwrap().as_str()).unwrap()*60f32 +
                f32::from_str(caps.get(3).unwrap().as_str()).unwrap()+
                f32::from_str(caps.get(4).unwrap().as_str()).unwrap() * 0.001f32)
        },
        None => Err(())
    }
}

/// Read a file to string; panics on error.
fn read_file<P: AsRef<Path>>(path: P) -> String {
    let path = path.as_ref();
    let mut file = File::open(path).expect(&format!("Could not open file: {:?}", path));

    let mut buf = String::with_capacity(file_len(&mut file).expect("Er testing file len"));
    file.read_to_string(&mut buf)
        .expect(&format!("Error reading file {:?}", path));
    buf
}

fn file_len(file : &mut File) -> io::Result<usize> {
    let old_pos = file.seek(SeekFrom::Current(0))?;
    let len = file.seek(SeekFrom::End(0))?;

    // Avoid seeking a third time when we were already at the end of the
    // stream. The branch is usually way cheaper than a seek operation.
    if old_pos != len {
        file.seek(SeekFrom::Start(old_pos))?;
    }

    Ok(len as usize)
}