|
|
|
|
@ -28,89 +28,101 @@
|
|
|
|
|
//! currently
|
|
|
|
|
|
|
|
|
|
#[derive(Debug, PartialEq)]
|
|
|
|
|
enum DataType {
|
|
|
|
|
St(String),
|
|
|
|
|
Arr(Vec<String>),
|
|
|
|
|
}
|
|
|
|
|
pub struct Action(Vec<String>);
|
|
|
|
|
|
|
|
|
|
// TODO(@ohsayan): Optmimize this function
|
|
|
|
|
fn parse_dataframe(df: Vec<u8>, sizes: Vec<usize>) -> Vec<DataType> {
|
|
|
|
|
unsafe {
|
|
|
|
|
(0..sizes.len())
|
|
|
|
|
.into_iter()
|
|
|
|
|
.scan((df.into_iter(), sizes.into_iter()), |(df, sizes), _| {
|
|
|
|
|
let cursize = match sizes.next() {
|
|
|
|
|
Some(s) => s,
|
|
|
|
|
_ => return None,
|
|
|
|
|
};
|
|
|
|
|
let curline: Vec<u8> = df.take(cursize).collect();
|
|
|
|
|
if curline.len() == 0 {
|
|
|
|
|
return None;
|
|
|
|
|
}
|
|
|
|
|
let _ = df.next();
|
|
|
|
|
match curline.get_unchecked(0) {
|
|
|
|
|
b'+' => {
|
|
|
|
|
// This is a string
|
|
|
|
|
return Some(DataType::St(
|
|
|
|
|
String::from_utf8_lossy(&curline.get_unchecked(1..)).to_string(),
|
|
|
|
|
));
|
|
|
|
|
}
|
|
|
|
|
b'&' => {
|
|
|
|
|
// This is an array
|
|
|
|
|
let mut remsize = 0;
|
|
|
|
|
let mut it = curline.into_iter().skip(1).peekable();
|
|
|
|
|
while let Some(tok) = it.next() {
|
|
|
|
|
if it.next().is_some() {
|
|
|
|
|
let s: usize = match tok.checked_sub(48) {
|
|
|
|
|
Some(x) => x.into(),
|
|
|
|
|
_ => return None,
|
|
|
|
|
};
|
|
|
|
|
remsize += s;
|
|
|
|
|
pub fn parse_df(buf: Vec<u8>, sizes: Vec<usize>, nc: usize) -> Option<Vec<Action>> {
|
|
|
|
|
let (mut i, mut pos) = (0, 0);
|
|
|
|
|
if buf.len() < 1 || sizes.len() < 1 {
|
|
|
|
|
// Having fun, eh? Why're you giving empty dataframes?
|
|
|
|
|
return None;
|
|
|
|
|
}
|
|
|
|
|
let mut commands = Vec::with_capacity(nc);
|
|
|
|
|
while i < sizes.len() && pos < buf.len() {
|
|
|
|
|
// Allocate everything first
|
|
|
|
|
unsafe {
|
|
|
|
|
let cursize = sizes.get_unchecked(0);
|
|
|
|
|
i += 1; // We've just read a line push it ahead
|
|
|
|
|
// Get the current line-> pos..pos+cursize+1
|
|
|
|
|
let curline = match buf.get(pos..pos + cursize + 1) {
|
|
|
|
|
Some(line) => line,
|
|
|
|
|
None => return None,
|
|
|
|
|
};
|
|
|
|
|
// We've read `cursize` number of elements, so skip them
|
|
|
|
|
// Also skip the newline
|
|
|
|
|
pos += cursize + 1;
|
|
|
|
|
if *curline.get_unchecked(0) == b'&' {
|
|
|
|
|
// A valid action array
|
|
|
|
|
let mut cursize = 0usize; // The number of elements in this action array
|
|
|
|
|
let mut k = 1; // Skip the '&' character in `curline`
|
|
|
|
|
while k < (curline.len() - 1) {
|
|
|
|
|
let cur_dig: usize = match curline.get_unchecked(k).checked_sub(48) {
|
|
|
|
|
Some(dig) => {
|
|
|
|
|
if dig > 9 {
|
|
|
|
|
// For the UTF8 character to be a number (0-9)
|
|
|
|
|
// `dig` must be lesser than 9, since `48` is the UTF8
|
|
|
|
|
// code for 0
|
|
|
|
|
return None;
|
|
|
|
|
} else {
|
|
|
|
|
dig.into()
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
let array_elems: Vec<String> = sizes
|
|
|
|
|
.take(remsize)
|
|
|
|
|
.into_iter()
|
|
|
|
|
.scan(0, |_, size| {
|
|
|
|
|
let tok: Vec<u8> = df.take(size).collect();
|
|
|
|
|
let _ = df.next();
|
|
|
|
|
Some(
|
|
|
|
|
String::from_utf8_lossy(&tok.get_unchecked(..size - 1))
|
|
|
|
|
.to_string(),
|
|
|
|
|
)
|
|
|
|
|
})
|
|
|
|
|
.collect();
|
|
|
|
|
// .into_iter()
|
|
|
|
|
// .map(|elemsize| {
|
|
|
|
|
// let v: Vec<u8> = df.take(elemsize).collect();
|
|
|
|
|
// let _ = df.next();
|
|
|
|
|
// String::from_utf8_lossy(&v.get_unchecked(..v.len() - 1)).to_string()
|
|
|
|
|
// })
|
|
|
|
|
// .collect();
|
|
|
|
|
return Some(DataType::Arr(array_elems));
|
|
|
|
|
}
|
|
|
|
|
_ => return None,
|
|
|
|
|
None => return None,
|
|
|
|
|
};
|
|
|
|
|
cursize = (cursize * 10) + cur_dig;
|
|
|
|
|
k += 1;
|
|
|
|
|
}
|
|
|
|
|
let mut toks: Vec<String> = sizes
|
|
|
|
|
.iter()
|
|
|
|
|
.take(cursize)
|
|
|
|
|
.map(|sz| String::with_capacity(*sz))
|
|
|
|
|
.collect();
|
|
|
|
|
let mut l = 0;
|
|
|
|
|
// We now know the array size, so let's parse it!
|
|
|
|
|
// Get all the sizes of the array elements
|
|
|
|
|
let arr_elem_sizes = match sizes.get(i..(i + cursize)) {
|
|
|
|
|
Some(sizes) => sizes,
|
|
|
|
|
None => return None,
|
|
|
|
|
};
|
|
|
|
|
i += cursize; // We've already read `cursize` items from the `sizes` array
|
|
|
|
|
arr_elem_sizes
|
|
|
|
|
.into_iter()
|
|
|
|
|
.zip(toks.iter_mut())
|
|
|
|
|
.for_each(|(size, empty_buf)| {
|
|
|
|
|
let extracted = match buf.get(pos..pos + size) {
|
|
|
|
|
Some(ex) => ex,
|
|
|
|
|
None => return (),
|
|
|
|
|
};
|
|
|
|
|
pos += size + 1; // Advance `pos` by `sz` and `1` for the newline
|
|
|
|
|
l += 1; // Move ahead
|
|
|
|
|
*empty_buf = String::from_utf8_lossy(extracted).to_string();
|
|
|
|
|
});
|
|
|
|
|
if toks.len() != cursize {
|
|
|
|
|
return None;
|
|
|
|
|
}
|
|
|
|
|
})
|
|
|
|
|
.collect()
|
|
|
|
|
// We're done with parsing the entire array, return it
|
|
|
|
|
commands.push(Action(toks));
|
|
|
|
|
} else {
|
|
|
|
|
i += 1;
|
|
|
|
|
continue;
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
}
|
|
|
|
|
Some(commands)
|
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
#[cfg(test)]
|
|
|
|
|
#[test]
|
|
|
|
|
fn test_df() {
|
|
|
|
|
let ss: Vec<usize> = vec![4, 3, 6, 7, 7];
|
|
|
|
|
let df = "+GET\n&3\n+sayan\n+foobar\n+opnsrc\n".as_bytes().to_owned();
|
|
|
|
|
let parsed = parse_dataframe(df, ss);
|
|
|
|
|
let ss: Vec<usize> = vec![2, 3, 5, 6, 6];
|
|
|
|
|
let df = "&4\nGET\nsayan\nfoobar\nopnsrc\n".as_bytes().to_owned();
|
|
|
|
|
let parsed = parse_df(df, ss, 1).unwrap();
|
|
|
|
|
assert_eq!(
|
|
|
|
|
parsed,
|
|
|
|
|
vec![
|
|
|
|
|
DataType::St("GET".to_owned()),
|
|
|
|
|
DataType::Arr(vec![
|
|
|
|
|
"sayan".to_owned(),
|
|
|
|
|
"foobar".to_owned(),
|
|
|
|
|
"opnsrc".to_owned(),
|
|
|
|
|
])
|
|
|
|
|
]
|
|
|
|
|
vec![Action(vec![
|
|
|
|
|
"GET".to_owned(),
|
|
|
|
|
"sayan".to_owned(),
|
|
|
|
|
"foobar".to_owned(),
|
|
|
|
|
"opnsrc".to_owned()
|
|
|
|
|
])]
|
|
|
|
|
);
|
|
|
|
|
}
|
|
|
|
|
|
Sayan Nandan
4 years ago