Add IdString32 #183

Merged
lucas merged 1 commit from feat/various into master 2024-07-19 11:41:44 +02:00
4 changed files with 347 additions and 167 deletions

View file

@ -0,0 +1,162 @@
use std::fmt;
use serde::{Deserializer, Serializer};
use super::Murmur32;
// This type encodes the fact that when reading in a bundle, we don't always have a dictionary
// entry for every hash in there. So we do want to have the real string available when needed,
// but at the same time retain the original hash information for when we don't.
// This is especially important when wanting to write back the read bundle, as the hashes need to
// stay the same.
// The previous system of always turning hashes into strings worked well for the purpose of
// displaying hashes, but would have made it very hard to turn a stringyfied hash back into
// an actual hash.
#[derive(Clone, Debug, Eq)]
pub enum IdString32 {
Hash(Murmur32),
String(String),
}
impl IdString32 {
pub fn to_murmur32(&self) -> Murmur32 {
match self {
Self::Hash(hash) => *hash,
Self::String(s) => Murmur32::hash(s.as_bytes()),
}
}
pub fn display(&self) -> IdString32Display {
let s = match self {
IdString32::Hash(hash) => hash.to_string(),
IdString32::String(s) => s.clone(),
};
IdString32Display(s)
}
pub fn is_string(&self) -> bool {
match self {
IdString32::Hash(_) => false,
IdString32::String(_) => true,
}
}
pub fn is_hash(&self) -> bool {
match self {
IdString32::Hash(_) => true,
IdString32::String(_) => false,
}
}
}
impl From<String> for IdString32 {
fn from(value: String) -> Self {
Self::String(value)
}
}
impl From<u32> for IdString32 {
fn from(value: u32) -> Self {
Self::Hash(value.into())
}
}
impl From<IdString32> for u32 {
fn from(value: IdString32) -> Self {
value.to_murmur32().into()
}
}
impl From<Murmur32> for IdString32 {
fn from(value: Murmur32) -> Self {
Self::Hash(value)
}
}
impl From<IdString32> for Murmur32 {
fn from(value: IdString32) -> Self {
value.to_murmur32()
}
}
impl PartialEq for IdString32 {
fn eq(&self, other: &Self) -> bool {
self.to_murmur32() == other.to_murmur32()
}
}
impl std::hash::Hash for IdString32 {
fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
state.write_u32(self.to_murmur32().into());
}
}
impl serde::Serialize for IdString32 {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: Serializer,
{
serializer.serialize_u32(self.to_murmur32().into())
}
}
struct IdString32Visitor;
impl<'de> serde::de::Visitor<'de> for IdString32Visitor {
type Value = IdString32;
fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
formatter.write_str("an u32 or a string")
}
fn visit_u32<E>(self, value: u32) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString32::Hash(value.into()))
}
fn visit_str<E>(self, v: &str) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString32::String(v.to_string()))
}
fn visit_string<E>(self, v: String) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString32::String(v))
}
}
impl<'de> serde::Deserialize<'de> for IdString32 {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
deserializer.deserialize_u32(IdString32Visitor)
}
}
pub struct IdString32Display(String);
impl std::fmt::Display for IdString32Display {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}", self.0)
}
}
impl std::fmt::UpperHex for IdString32 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::UpperHex::fmt(&self.to_murmur32(), f)
}
}
impl std::fmt::LowerHex for IdString32 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::LowerHex::fmt(&self.to_murmur32(), f)
}
}

View file

@ -0,0 +1,175 @@
use std::{fmt, path::Path};
use path_slash::PathExt as _;
use serde::{Deserializer, Serializer};
use super::Murmur64;
// This type encodes the fact that when reading in a bundle, we don't always have a dictionary
// entry for every hash in there. So we do want to have the real string available when needed,
// but at the same time retain the original hash information for when we don't.
// This is especially important when wanting to write back the read bundle, as the hashes need to
// stay the same.
// The previous system of always turning hashes into strings worked well for the purpose of
// displaying hashes, but would have made it very hard to turn a stringyfied hash back into
// an actual hash.
#[derive(Clone, Debug, Eq)]
pub enum IdString64 {
Hash(Murmur64),
String(String),
}
impl IdString64 {
pub fn to_murmur64(&self) -> Murmur64 {
match self {
Self::Hash(hash) => *hash,
Self::String(s) => Murmur64::hash(s.as_bytes()),
}
}
pub fn display(&self) -> IdString64Display {
let s = match self {
IdString64::Hash(hash) => hash.to_string(),
IdString64::String(s) => s.clone(),
};
IdString64Display(s)
}
pub fn is_string(&self) -> bool {
match self {
IdString64::Hash(_) => false,
IdString64::String(_) => true,
}
}
pub fn is_hash(&self) -> bool {
match self {
IdString64::Hash(_) => true,
IdString64::String(_) => false,
}
}
// Would love to have this as a proper `impl From`, but
// rustc will complain that it overlaps with the `impl From<Into<String>>`.
pub fn from_path(p: impl AsRef<Path>) -> Self {
Self::String(p.as_ref().to_slash_lossy().to_string())
}
}
impl From<String> for IdString64 {
fn from(value: String) -> Self {
Self::String(value)
}
}
impl From<u64> for IdString64 {
fn from(value: u64) -> Self {
Self::Hash(value.into())
}
}
impl From<Murmur64> for IdString64 {
fn from(value: Murmur64) -> Self {
Self::Hash(value)
}
}
impl From<IdString64> for Murmur64 {
fn from(value: IdString64) -> Self {
value.to_murmur64()
}
}
impl From<IdString64> for u64 {
fn from(value: IdString64) -> Self {
value.to_murmur64().into()
}
}
impl Default for IdString64 {
fn default() -> Self {
Self::Hash(0.into())
}
}
impl PartialEq for IdString64 {
fn eq(&self, other: &Self) -> bool {
self.to_murmur64() == other.to_murmur64()
}
}
impl std::hash::Hash for IdString64 {
fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
state.write_u64(self.to_murmur64().into());
}
}
impl serde::Serialize for IdString64 {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: Serializer,
{
serializer.serialize_u64(self.to_murmur64().into())
}
}
struct IdString64Visitor;
impl<'de> serde::de::Visitor<'de> for IdString64Visitor {
type Value = IdString64;
fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
formatter.write_str("an u64 or a string")
}
fn visit_u64<E>(self, value: u64) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::Hash(value.into()))
}
fn visit_str<E>(self, v: &str) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::String(v.to_string()))
}
fn visit_string<E>(self, v: String) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::String(v))
}
}
impl<'de> serde::Deserialize<'de> for IdString64 {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
deserializer.deserialize_u64(IdString64Visitor)
}
}
pub struct IdString64Display(String);
impl std::fmt::Display for IdString64Display {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}", self.0)
}
}
impl std::fmt::UpperHex for IdString64 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::UpperHex::fmt(&self.to_murmur64(), f)
}
}
impl std::fmt::LowerHex for IdString64 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::LowerHex::fmt(&self.to_murmur64(), f)
}
}

View file

@ -8,6 +8,8 @@ use serde::{Deserialize, Deserializer, Serialize, Serializer};
mod dictionary;
// Currently unused
// mod murmurhash32;
mod idstring32;
mod idstring64;
mod murmurhash64;
mod types;
mod util;
@ -15,6 +17,8 @@ mod util;
pub const SEED: u32 = 0;
pub use dictionary::{Dictionary, Entry, HashGroup};
pub use idstring32::*;
pub use idstring64::*;
pub use murmurhash64::hash;
pub use murmurhash64::hash32;
pub use murmurhash64::hash_inverse as inverse;

View file

@ -1,7 +1,3 @@
use std::path::Path;
use path_slash::PathExt;
use self::util::{parse_hex32, parse_hex64};
use super::*;
@ -154,6 +150,12 @@ impl fmt::UpperHex for Murmur32 {
}
}
impl fmt::LowerHex for Murmur32 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
fmt::LowerHex::fmt(&self.0, f)
}
}
impl fmt::Display for Murmur32 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{:08X}", self)
@ -222,166 +224,3 @@ impl<'de> Deserialize<'de> for Murmur32 {
deserializer.deserialize_any(Self(0))
}
}
// This type encodes the fact that when reading in a bundle, we don't always have a dictionary
// entry for every hash in there. So we do want to have the real string available when needed,
// but at the same time retain the original hash information for when we don't.
// This is especially important when wanting to write back the read bundle, as the hashes need to
// stay the same.
// The previous system of always turning hashes into strings worked well for the purpose of
// displaying hashes, but would have made it very hard to turn a stringyfied hash back into
// an actual hash.
#[derive(Clone, Debug, Eq)]
pub enum IdString64 {
Hash(Murmur64),
String(String),
}
impl IdString64 {
pub fn to_murmur64(&self) -> Murmur64 {
match self {
Self::Hash(hash) => *hash,
Self::String(s) => Murmur64::hash(s.as_bytes()),
}
}
pub fn display(&self) -> IdString64Display {
let s = match self {
IdString64::Hash(hash) => hash.to_string(),
IdString64::String(s) => s.clone(),
};
IdString64Display(s)
}
pub fn is_string(&self) -> bool {
match self {
IdString64::Hash(_) => false,
IdString64::String(_) => true,
}
}
pub fn is_hash(&self) -> bool {
match self {
IdString64::Hash(_) => true,
IdString64::String(_) => false,
}
}
// Would love to have this as a proper `impl From`, but
// rustc will complain that it overlaps with the `impl From<Into<String>>`.
pub fn from_path(p: impl AsRef<Path>) -> Self {
Self::String(p.as_ref().to_slash_lossy().to_string())
}
}
impl From<String> for IdString64 {
fn from(value: String) -> Self {
Self::String(value)
}
}
impl From<u64> for IdString64 {
fn from(value: u64) -> Self {
Self::Hash(value.into())
}
}
impl From<Murmur64> for IdString64 {
fn from(value: Murmur64) -> Self {
Self::Hash(value)
}
}
impl From<IdString64> for Murmur64 {
fn from(value: IdString64) -> Self {
value.to_murmur64()
}
}
impl Default for IdString64 {
fn default() -> Self {
Self::Hash(0.into())
}
}
impl PartialEq for IdString64 {
fn eq(&self, other: &Self) -> bool {
self.to_murmur64() == other.to_murmur64()
}
}
impl std::hash::Hash for IdString64 {
fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
state.write_u64(self.to_murmur64().into());
}
}
impl serde::Serialize for IdString64 {
fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
where
S: Serializer,
{
serializer.serialize_u64(self.to_murmur64().into())
}
}
struct IdString64Visitor;
impl<'de> serde::de::Visitor<'de> for IdString64Visitor {
type Value = IdString64;
fn expecting(&self, formatter: &mut fmt::Formatter) -> fmt::Result {
formatter.write_str("an u64 or a string")
}
fn visit_u64<E>(self, value: u64) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::Hash(value.into()))
}
fn visit_str<E>(self, v: &str) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::String(v.to_string()))
}
fn visit_string<E>(self, v: String) -> Result<Self::Value, E>
where
E: serde::de::Error,
{
Ok(IdString64::String(v))
}
}
impl<'de> serde::Deserialize<'de> for IdString64 {
fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
where
D: Deserializer<'de>,
{
deserializer.deserialize_u64(IdString64Visitor)
}
}
pub struct IdString64Display(String);
impl std::fmt::Display for IdString64Display {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
write!(f, "{}", self.0)
}
}
impl std::fmt::UpperHex for IdString64 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::UpperHex::fmt(&self.to_murmur64(), f)
}
}
impl std::fmt::LowerHex for IdString64 {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
std::fmt::LowerHex::fmt(&self.to_murmur64(), f)
}
}