refactor: drop hexrange

Signed-off-by: Greg Heartsfield <scsibug@imap.cc>
This commit is contained in:
Kieran 2023-11-24 12:52:25 +00:00 committed by Greg Heartsfield
parent 971889f9a6
commit 9c86f03902
4 changed files with 75 additions and 343 deletions

View File

@ -1,159 +0,0 @@
//! Utilities for searching hexadecimal
use crate::utils::is_hex;
use hex;
/// Types of hexadecimal queries.
#[derive(PartialEq, Eq, PartialOrd, Ord, Debug, Clone)]
pub enum HexSearch {
// when no range is needed, exact 32-byte
Exact(Vec<u8>),
// lower (inclusive) and upper range (exclusive)
Range(Vec<u8>, Vec<u8>),
// lower bound only, upper bound is MAX inclusive
LowerOnly(Vec<u8>),
}
/// Check if a string contains only f chars
fn is_all_fs(s: &str) -> bool {
s.chars().all(|x| x == 'f' || x == 'F')
}
/// Find the next hex sequence greater than the argument.
#[must_use]
pub fn hex_range(s: &str) -> Option<HexSearch> {
let mut hash_base = s.to_owned();
if !is_hex(&hash_base) || hash_base.len() > 64 {
return None;
}
if hash_base.len() == 64 {
return Some(HexSearch::Exact(hex::decode(&hash_base).ok()?));
}
// if s is odd, add a zero
let mut odd = hash_base.len() % 2 != 0;
if odd {
// extend the string to make it even
hash_base.push('0');
}
let base = hex::decode(hash_base).ok()?;
// check for all ff's
if is_all_fs(s) {
// there is no higher bound, we only want to search for blobs greater than this.
return Some(HexSearch::LowerOnly(base));
}
// return a range
let mut upper = base.clone();
let mut byte_len = upper.len();
// for odd strings, we made them longer, but we want to increment the upper char (+16).
// we know we can do this without overflowing because we explicitly set the bottom half to 0's.
while byte_len > 0 {
byte_len -= 1;
// check if byte can be incremented, or if we need to carry.
let b = upper[byte_len];
if b == u8::MAX {
// reset and carry
upper[byte_len] = 0;
} else if odd {
// check if first char in this byte is NOT 'f'
if b < 240 {
// bump up the first character in this byte
upper[byte_len] = b + 16;
// increment done, stop iterating through the vec
break;
}
// if it is 'f', reset the byte to 0 and do a carry
// reset and carry
upper[byte_len] = 0;
// done with odd logic, so don't repeat this
odd = false;
} else {
// bump up the first character in this byte
upper[byte_len] = b + 1;
// increment done, stop iterating
break;
}
}
Some(HexSearch::Range(base, upper))
}
#[cfg(test)]
mod tests {
use super::*;
use crate::error::Result;
#[test]
fn hex_range_exact() -> Result<()> {
let hex = "abcdef00abcdef00abcdef00abcdef00abcdef00abcdef00abcdef00abcdef00";
let r = hex_range(hex);
assert_eq!(
r,
Some(HexSearch::Exact(hex::decode(hex).expect("invalid hex")))
);
Ok(())
}
#[test]
fn hex_full_range() -> Result<()> {
let hex = "aaaa";
let hex_upper = "aaab";
let r = hex_range(hex);
assert_eq!(
r,
Some(HexSearch::Range(
hex::decode(hex).expect("invalid hex"),
hex::decode(hex_upper).expect("invalid hex")
))
);
Ok(())
}
#[test]
fn hex_full_range_odd() -> Result<()> {
let r = hex_range("abc");
assert_eq!(
r,
Some(HexSearch::Range(
hex::decode("abc0").expect("invalid hex"),
hex::decode("abd0").expect("invalid hex")
))
);
Ok(())
}
#[test]
fn hex_full_range_odd_end_f() -> Result<()> {
let r = hex_range("abf");
assert_eq!(
r,
Some(HexSearch::Range(
hex::decode("abf0").expect("invalid hex"),
hex::decode("ac00").expect("invalid hex")
))
);
Ok(())
}
#[test]
fn hex_no_upper() -> Result<()> {
let r = hex_range("ffff");
assert_eq!(
r,
Some(HexSearch::LowerOnly(
hex::decode("ffff").expect("invalid hex")
))
);
Ok(())
}
#[test]
fn hex_no_upper_odd() -> Result<()> {
let r = hex_range("fff");
assert_eq!(
r,
Some(HexSearch::LowerOnly(
hex::decode("fff0").expect("invalid hex")
))
);
Ok(())
}
}

View File

@ -6,7 +6,6 @@ pub mod db;
pub mod delegation; pub mod delegation;
pub mod error; pub mod error;
pub mod event; pub mod event;
pub mod hexrange;
pub mod info; pub mod info;
pub mod nauthz; pub mod nauthz;
pub mod nip05; pub mod nip05;

View File

@ -14,7 +14,6 @@ use sqlx::{Error, Execute, FromRow, Postgres, QueryBuilder, Row};
use std::time::{Duration, Instant}; use std::time::{Duration, Instant};
use crate::error; use crate::error;
use crate::hexrange::{hex_range, HexSearch};
use crate::repo::postgres_migration::run_migrations; use crate::repo::postgres_migration::run_migrations;
use crate::server::NostrMetrics; use crate::server::NostrMetrics;
use crate::utils::{self, is_hex, is_lower_hex}; use crate::utils::{self, is_hex, is_lower_hex};
@ -60,8 +59,7 @@ async fn cleanup_expired(conn: PostgresPool, frequency: Duration) -> Result<()>
} }
} }
} }
} };
;
} }
}); });
Ok(()) Ok(())
@ -726,6 +724,7 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
} }
let mut query = QueryBuilder::new("SELECT e.\"content\", e.created_at FROM \"event\" e WHERE "); let mut query = QueryBuilder::new("SELECT e.\"content\", e.created_at FROM \"event\" e WHERE ");
// This tracks whether we need to push a prefix AND before adding another clause // This tracks whether we need to push a prefix AND before adding another clause
let mut push_and = false; let mut push_and = false;
// Query for "authors", allowing prefix matches // Query for "authors", allowing prefix matches
@ -736,12 +735,8 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
if auth_vec.is_empty() { if auth_vec.is_empty() {
return None; return None;
} }
query.push("("); query.push("(e.pub_key in (");
// shortcut authors into "IN" query
let any_is_range = auth_vec.iter().any(|pk| pk.len() != 64);
if !any_is_range {
query.push("e.pub_key in (");
let mut pk_sep = query.separated(", "); let mut pk_sep = query.separated(", ");
for pk in auth_vec.iter() { for pk in auth_vec.iter() {
pk_sep.push_bind(hex::decode(pk).ok()); pk_sep.push_bind(hex::decode(pk).ok());
@ -751,48 +746,8 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
for pk in auth_vec.iter() { for pk in auth_vec.iter() {
pk_delegated_sep.push_bind(hex::decode(pk).ok()); pk_delegated_sep.push_bind(hex::decode(pk).ok());
} }
query.push(")");
push_and = true; push_and = true;
} else { query.push("))");
let mut range_authors = query.separated(" OR ");
for auth in auth_vec {
match hex_range(auth) {
Some(HexSearch::Exact(ex)) => {
range_authors
.push("(e.pub_key = ")
.push_bind_unseparated(ex.clone())
.push_unseparated(" OR e.delegated_by = ")
.push_bind_unseparated(ex)
.push_unseparated(")");
}
Some(HexSearch::Range(lower, upper)) => {
range_authors
.push("((e.pub_key > ")
.push_bind_unseparated(lower.clone())
.push_unseparated(" AND e.pub_key < ")
.push_bind_unseparated(upper.clone())
.push_unseparated(") OR (e.delegated_by > ")
.push_bind_unseparated(lower)
.push_unseparated(" AND e.delegated_by < ")
.push_bind_unseparated(upper)
.push_unseparated("))");
}
Some(HexSearch::LowerOnly(lower)) => {
range_authors
.push("(e.pub_key > ")
.push_bind_unseparated(lower.clone())
.push_unseparated(" OR e.delegated_by > ")
.push_bind_unseparated(lower)
.push_unseparated(")");
}
None => {
info!("Could not parse hex range from author {:?}", auth);
}
}
push_and = true;
}
}
query.push(")");
} }
// Query for Kind // Query for Kind
@ -813,7 +768,7 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
query.push(")"); query.push(")");
} }
// Query for event, allowing prefix matches // Query for event,
if let Some(id_vec) = &f.ids { if let Some(id_vec) = &f.ids {
// filter out non-hex values // filter out non-hex values
let id_vec: Vec<&String> = id_vec.iter().filter(|a| is_hex(a)).collect(); let id_vec: Vec<&String> = id_vec.iter().filter(|a| is_hex(a)).collect();
@ -827,48 +782,12 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
} }
push_and = true; push_and = true;
// shortcut ids into "IN" query
let any_is_range = id_vec.iter().any(|pk| pk.len() != 64);
if !any_is_range {
query.push("id in ("); query.push("id in (");
let mut sep = query.separated(", "); let mut sep = query.separated(", ");
for id in id_vec.iter() { for id in id_vec.iter() {
sep.push_bind(hex::decode(id).ok()); sep.push_bind(hex::decode(id).ok());
} }
query.push(")"); query.push("))");
} else {
// take each author and convert to a hex search
let mut id_query = query.separated(" OR ");
for id in id_vec {
match hex_range(id) {
Some(HexSearch::Exact(ex)) => {
id_query
.push("(id = ")
.push_bind_unseparated(ex)
.push_unseparated(")");
}
Some(HexSearch::Range(lower, upper)) => {
id_query
.push("(id > ")
.push_bind_unseparated(lower)
.push_unseparated(" AND id < ")
.push_bind_unseparated(upper)
.push_unseparated(")");
}
Some(HexSearch::LowerOnly(lower)) => {
id_query
.push("(id > ")
.push_bind_unseparated(lower)
.push_unseparated(")");
}
None => {
info!("Could not parse hex range from id {:?}", id);
}
}
}
}
query.push(")");
} }
// Query for tags // Query for tags
@ -888,7 +807,8 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
if push_or { if push_or {
query.push(" OR "); query.push(" OR ");
} }
query.push("(t.\"name\" = ") query
.push("(t.\"name\" = ")
.push_bind(key.to_string()) .push_bind(key.to_string())
.push(" AND ("); .push(" AND (");
@ -898,8 +818,7 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
query.push("value in ("); query.push("value in (");
// plain value match first // plain value match first
let mut tag_query = query.separated(", "); let mut tag_query = query.separated(", ");
for v in val.iter() for v in val.iter().filter(|v| !is_lower_hex(v)) {
.filter(|v| !is_lower_hex(v)) {
tag_query.push_bind(v.as_bytes()); tag_query.push_bind(v.as_bytes());
} }
} }
@ -910,8 +829,7 @@ fn query_from_filter(f: &ReqFilter) -> Option<QueryBuilder<Postgres>> {
query.push("value_hex in ("); query.push("value_hex in (");
// plain value match first // plain value match first
let mut tag_query = query.separated(", "); let mut tag_query = query.separated(", ");
for v in val.iter() for v in val.iter().filter(|v| v.len() % 2 == 0 && is_lower_hex(v)) {
.filter(|v| v.len() % 2 == 0 && is_lower_hex(v)) {
tag_query.push_bind(hex::decode(v).ok()); tag_query.push_bind(hex::decode(v).ok());
} }
} }
@ -988,11 +906,10 @@ impl FromRow<'_, PgRow> for VerificationRecord {
} }
} }
#[cfg(test)] #[cfg(test)]
mod tests { mod tests {
use std::collections::{HashMap, HashSet};
use super::*; use super::*;
use std::collections::{HashMap, HashSet};
#[test] #[test]
fn test_query_gen_tag_value_hex() { fn test_query_gen_tag_value_hex() {
@ -1001,11 +918,16 @@ mod tests {
kinds: Some(vec![1000]), kinds: Some(vec![1000]),
since: None, since: None,
until: None, until: None,
authors: Some(vec!["84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned()]), authors: Some(vec![
"84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned(),
]),
limit: None, limit: None,
tags: Some(HashMap::from([ tags: Some(HashMap::from([(
('p', HashSet::from(["63fe6318dc58583cfe16810f86dd09e18bfd76aabc24a0081ce2856f330504ed".to_owned()])) 'p',
])), HashSet::from([
"63fe6318dc58583cfe16810f86dd09e18bfd76aabc24a0081ce2856f330504ed".to_owned(),
]),
)])),
force_no_match: false, force_no_match: false,
}; };
@ -1020,11 +942,11 @@ mod tests {
kinds: Some(vec![1000]), kinds: Some(vec![1000]),
since: None, since: None,
until: None, until: None,
authors: Some(vec!["84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned()]), authors: Some(vec![
"84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned(),
]),
limit: None, limit: None,
tags: Some(HashMap::from([ tags: Some(HashMap::from([('d', HashSet::from(["test".to_owned()]))])),
('d', HashSet::from(["test".to_owned()]))
])),
force_no_match: false, force_no_match: false,
}; };
@ -1039,11 +961,17 @@ mod tests {
kinds: Some(vec![1000]), kinds: Some(vec![1000]),
since: None, since: None,
until: None, until: None,
authors: Some(vec!["84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned()]), authors: Some(vec![
"84de35e2584d2b144aae823c9ed0b0f3deda09648530b93d1a2a146d1dea9864".to_owned(),
]),
limit: None, limit: None,
tags: Some(HashMap::from([ tags: Some(HashMap::from([(
('d', HashSet::from(["test".to_owned(), "63fe6318dc58583cfe16810f86dd09e18bfd76aabc24a0081ce2856f330504ed".to_owned()])) 'd',
])), HashSet::from([
"test".to_owned(),
"63fe6318dc58583cfe16810f86dd09e18bfd76aabc24a0081ce2856f330504ed".to_owned(),
]),
)])),
force_no_match: false, force_no_match: false,
}; };
@ -1062,7 +990,7 @@ mod tests {
limit: None, limit: None,
tags: Some(HashMap::from([ tags: Some(HashMap::from([
('d', HashSet::from(["follow".to_owned()])), ('d', HashSet::from(["follow".to_owned()])),
('t', HashSet::from(["siamstr".to_owned()])) ('t', HashSet::from(["siamstr".to_owned()])),
])), ])),
force_no_match: false, force_no_match: false,
}; };
@ -1079,9 +1007,7 @@ mod tests {
until: None, until: None,
authors: None, authors: None,
limit: None, limit: None,
tags: Some(HashMap::from([ tags: Some(HashMap::from([('a', HashSet::new())])),
('a', HashSet::new())
])),
force_no_match: false, force_no_match: false,
}; };
assert!(query_from_filter(&filter).is_none()); assert!(query_from_filter(&filter).is_none());

View File

@ -4,8 +4,6 @@ use crate::config::Settings;
use crate::db::QueryResult; use crate::db::QueryResult;
use crate::error::{Error::SqlError, Result}; use crate::error::{Error::SqlError, Result};
use crate::event::{single_char_tagname, Event}; use crate::event::{single_char_tagname, Event};
use crate::hexrange::hex_range;
use crate::hexrange::HexSearch;
use crate::nip05::{Nip05Name, VerificationRecord}; use crate::nip05::{Nip05Name, VerificationRecord};
use crate::payment::{InvoiceInfo, InvoiceStatus}; use crate::payment::{InvoiceInfo, InvoiceStatus};
use crate::repo::sqlite_migration::{upgrade_db, STARTUP_SQL}; use crate::repo::sqlite_migration::{upgrade_db, STARTUP_SQL};
@ -994,24 +992,8 @@ fn query_from_filter(f: &ReqFilter) -> (String, Vec<Box<dyn ToSql>>, Option<Stri
// take each author and convert to a hexsearch // take each author and convert to a hexsearch
let mut auth_searches: Vec<String> = vec![]; let mut auth_searches: Vec<String> = vec![];
for auth in authvec { for auth in authvec {
match hex_range(auth) {
Some(HexSearch::Exact(ex)) => {
auth_searches.push("author=?".to_owned()); auth_searches.push("author=?".to_owned());
params.push(Box::new(ex)); params.push(Box::new(auth.clone()));
}
Some(HexSearch::Range(lower, upper)) => {
auth_searches.push("(author>? AND author<?)".to_owned());
params.push(Box::new(lower));
params.push(Box::new(upper));
}
Some(HexSearch::LowerOnly(lower)) => {
auth_searches.push("author>?".to_owned());
params.push(Box::new(lower));
}
None => {
trace!("Could not parse hex range from author {:?}", auth);
}
}
} }
if !authvec.is_empty() { if !authvec.is_empty() {
let auth_clause = format!("({})", auth_searches.join(" OR ")); let auth_clause = format!("({})", auth_searches.join(" OR "));
@ -1032,24 +1014,8 @@ fn query_from_filter(f: &ReqFilter) -> (String, Vec<Box<dyn ToSql>>, Option<Stri
// take each author and convert to a hexsearch // take each author and convert to a hexsearch
let mut id_searches: Vec<String> = vec![]; let mut id_searches: Vec<String> = vec![];
for id in idvec { for id in idvec {
match hex_range(id) {
Some(HexSearch::Exact(ex)) => {
id_searches.push("event_hash=?".to_owned()); id_searches.push("event_hash=?".to_owned());
params.push(Box::new(ex)); params.push(Box::new(id.clone()));
}
Some(HexSearch::Range(lower, upper)) => {
id_searches.push("(event_hash>? AND event_hash<?)".to_owned());
params.push(Box::new(lower));
params.push(Box::new(upper));
}
Some(HexSearch::LowerOnly(lower)) => {
id_searches.push("event_hash>?".to_owned());
params.push(Box::new(lower));
}
None => {
info!("Could not parse hex range from id {:?}", id);
}
}
} }
if idvec.is_empty() { if idvec.is_empty() {
// if the ids list was empty, we should never return // if the ids list was empty, we should never return