Fix #206 and make (most) emotes embed in comments (#209)

* Fix links not being converted when multiple emojis are in one comment

* Make (most) emotes embed within comments

* Restore the behavior that the "rewrite_urls_removes_backslashes_and_rewrites_url" test looks for

* Listen to cargo fmt and cargo clippy's suggestions as well as removing some leftover comments and code

---------

Co-authored-by: Matthew Esposito <matt@matthew.science>
This commit is contained in:
Butter Cat 2024-09-25 13:36:23 -04:00 committed by GitHub
parent 403513ac4c
commit a807002ddf
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
5 changed files with 165 additions and 8 deletions

60
Cargo.lock generated
View File

@ -888,6 +888,12 @@ dependencies = [
"cfg-if",
]
[[package]]
name = "inventory"
version = "0.3.15"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "f958d3d68f4167080a18141e10381e7634563984a537f2a49a30fd8e53ac5767"
[[package]]
name = "is-terminal"
version = "0.4.12"
@ -1255,6 +1261,7 @@ dependencies = [
"sealed_test",
"serde",
"serde_json",
"serde_json_path",
"serde_yaml",
"time",
"tokio",
@ -1565,6 +1572,59 @@ dependencies = [
"serde",
]
[[package]]
name = "serde_json_path"
version = "0.6.7"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "0bc0207b6351893eafa1e39aa9aea452abb6425ca7b02dd64faf29109e7a33ba"
dependencies = [
"inventory",
"nom",
"once_cell",
"regex",
"serde",
"serde_json",
"serde_json_path_core",
"serde_json_path_macros",
"thiserror",
]
[[package]]
name = "serde_json_path_core"
version = "0.1.6"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "a3d64fe53ce1aaa31bea2b2b46d3b6ab6a37e61854bedcbd9f174e188f3f7d79"
dependencies = [
"inventory",
"once_cell",
"serde",
"serde_json",
"thiserror",
]
[[package]]
name = "serde_json_path_macros"
version = "0.1.4"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "2a31e8177a443fd3e94917f12946ae7891dfb656e6d4c5e79b8c5d202fbcb723"
dependencies = [
"inventory",
"once_cell",
"serde_json_path_core",
"serde_json_path_macros_internal",
]
[[package]]
name = "serde_json_path_macros_internal"
version = "0.1.1"
source = "registry+https://github.com/rust-lang/crates.io-index"
checksum = "75dde5a1d2ed78dfc411fc45592f72d3694436524d3353683ecb3d22009731dc"
dependencies = [
"proc-macro2",
"quote",
"syn 2.0.68",
]
[[package]]
name = "serde_spanned"
version = "0.6.6"

View File

@ -44,6 +44,7 @@ pretty_env_logger = "0.5.0"
dotenvy = "0.15.7"
rss = "2.0.7"
arc-swap = "1.7.1"
serde_json_path = "0.6.7"
async-recursion = "1.1.1"

View File

@ -239,6 +239,9 @@ async fn main() {
app.at("/img/*path").get(|r| proxy(r, "https://i.redd.it/{path}").boxed());
app.at("/thumb/:point/:id").get(|r| proxy(r, "https://{point}.thumbs.redditmedia.com/{id}").boxed());
app.at("/emoji/:id/:name").get(|r| proxy(r, "https://emoji.redditmedia.com/{id}/{name}").boxed());
app
.at("/emote/:subreddit_id/:filename")
.get(|r| proxy(r, "https://reddit-econ-prod-assets-permanent.s3.amazonaws.com/asset-manager/{subreddit_id}/{filename}").boxed());
app
.at("/preview/:loc/award_images/:fullname/:id")
.get(|r| proxy(r, "https://{loc}view.redd.it/award_images/{fullname}/{id}").boxed());

View File

@ -4,7 +4,7 @@ use crate::config::get_setting;
use crate::server::RequestExt;
use crate::subreddit::{can_access_quarantine, quarantine};
use crate::utils::{
error, format_num, get_filters, nsfw_landing, param, parse_post, rewrite_urls, setting, template, time, val, Author, Awards, Comment, Flair, FlairPart, Post, Preferences,
error, format_num, get_filters, nsfw_landing, param, parse_post, rewrite_emotes, setting, template, time, val, Author, Awards, Comment, Flair, FlairPart, Post, Preferences,
};
use hyper::{Body, Request, Response};
@ -178,7 +178,7 @@ fn build_comment(
get_setting("REDLIB_PUSHSHIFT_FRONTEND").unwrap_or_else(|| String::from(crate::config::DEFAULT_PUSHSHIFT_FRONTEND)),
)
} else {
rewrite_urls(&val(comment, "body_html"))
rewrite_emotes(&data["media_metadata"], val(comment, "body_html"))
};
let kind = comment["kind"].as_str().unwrap_or_default().to_string();

View File

@ -12,6 +12,7 @@ use once_cell::sync::Lazy;
use regex::Regex;
use rust_embed::RustEmbed;
use serde_json::Value;
use serde_json_path::{JsonPath, JsonPathExt};
use std::collections::{HashMap, HashSet};
use std::env;
use std::str::FromStr;
@ -919,12 +920,19 @@ pub fn rewrite_urls(input_text: &str) -> String {
// Rewrite Reddit links to Redlib
REDDIT_REGEX.replace_all(input_text, r#"href="/"#)
.to_string();
text1 = REDDIT_EMOJI_REGEX
.replace_all(&text1, format_url(REDDIT_EMOJI_REGEX.find(&text1).map(|x| x.as_str()).unwrap_or_default()))
.to_string()
// Remove (html-encoded) "\" from URLs.
.replace("%5C", "")
.replace("\\_", "_");
loop {
if REDDIT_EMOJI_REGEX.find(&text1).is_none() {
break;
} else {
text1 = REDDIT_EMOJI_REGEX
.replace_all(&text1, format_url(REDDIT_EMOJI_REGEX.find(&text1).map(|x| x.as_str()).unwrap_or_default()))
.to_string()
}
}
// Remove (html-encoded) "\" from URLs.
text1 = text1.replace("%5C", "").replace("\\_", "_");
// Rewrite external media previews to Redlib
loop {
@ -980,6 +988,83 @@ pub fn rewrite_urls(input_text: &str) -> String {
}
}
// These links all follow a pattern of "https://reddit-econ-prod-assets-permanent.s3.amazonaws.com/asset-manager/SUBREDDIT_ID/RANDOM_FILENAME.png"
static REDDIT_EMOTE_LINK_REGEX: Lazy<Regex> = Lazy::new(|| Regex::new(r#"https://reddit-econ-prod-assets-permanent.s3.amazonaws.com/asset-manager/(.*)"#).unwrap());
// These all follow a pattern of '"emote|SUBREDDIT_IT|NUMBER"', we want the number
static REDDIT_EMOTE_ID_NUMBER_REGEX: Lazy<Regex> = Lazy::new(|| Regex::new(r#""emote\|.*\|(.*)""#).unwrap());
pub fn rewrite_emotes(media_metadata: &Value, comment: String) -> String {
/* Create the paths we'll use to look for our data inside the json.
Because we don't know the name of any given emote we use a wildcard to parse them. */
let link_path = JsonPath::parse("$[*].s.u").expect("valid JSON Path");
let id_path = JsonPath::parse("$[*].id").expect("valid JSON Path");
let size_path = JsonPath::parse("$[*].s.y").expect("valid JSON Path");
// Extract all of the results from those json paths
let link_nodes = media_metadata.json_path(&link_path);
let id_nodes = media_metadata.json_path(&id_path);
// Initialize our vectors
let mut id_vec = Vec::new();
let mut link_vec = Vec::new();
// Add the relevant data to each of our vectors so we can access it by number later
for current_id in id_nodes {
id_vec.push(current_id)
}
for current_link in link_nodes {
link_vec.push(current_link)
}
/* Set index to the length of link_vec.
This is one larger than we'll actually be looking at, but we correct that later */
let mut index = link_vec.len();
// Comment needs to be in scope for when we call rewrite_urls()
let mut comment = comment;
/* Loop until index hits zero.
This also prevents us from trying to do anything on an empty vector */
while index != 0 {
/* Subtract 1 from index to get the real index we should be looking at.
Then continue on each subsequent loop to continue until we hit the last entry in the vector.
This is how we get this to deal with multiple emotes in a single message and properly replace each ID with it's link */
index -= 1;
// Convert our current index in id_vec into a string so we can search through it with regex
let current_id = id_vec[index].to_string();
/* The ID number can be multiple lengths, so we capture it with regex.
We also want to only attempt anything when we get matches to avoid panicking */
if let Some(id_capture) = REDDIT_EMOTE_ID_NUMBER_REGEX.captures(&current_id) {
// Format the ID to include the colons it has in the comment text
let id = format!(":{}:", &id_capture[1]);
// Convert current link to string to search through it with the regex
let link = link_vec[index].to_string();
// Make sure we only do operations when we get matches, otherwise we panic when trying to access the first match
if let Some(link_capture) = REDDIT_EMOTE_LINK_REGEX.captures(&link) {
/* Reddit sends a size for the image based on whether it's alone or accompanied by text.
It's a good idea and makes everything look nicer, so we'll do the same. */
let size = media_metadata.json_path(&size_path).first().unwrap().to_string();
// Replace the ID we found earlier in the comment with the respective image and it's link from the regex capture
let to_replace_with = format!(
"<img loading=\"lazy\" src=\"/emote/{} width=\"{size}\" height=\"{size}\" style=\"vertical-align:text-bottom\">",
&link_capture[1]
);
// Inside the comment replace the ID we found with the string that will embed the image
comment = comment.replace(&id, &to_replace_with).to_string();
}
}
}
// Call rewrite_urls() to transform any other Reddit links
rewrite_urls(&comment)
}
// Format vote count to a string that will be displayed.
// Append `m` and `k` for millions and thousands respectively, and
// round to the nearest tenth.
@ -1301,3 +1386,11 @@ fn test_url_path_basename() {
// empty path
assert_eq!(url_path_basename("/"), "");
}
#[test]
fn test_rewriting_emotes() {
let json_input = serde_json::from_str(r#"{"emote|t5_31hpy|2028":{"e":"Image","id":"emote|t5_31hpy|2028","m":"image/png","s":{"u":"https://reddit-econ-prod-assets-permanent.s3.amazonaws.com/asset-manager/t5_31hpy/PW6WsOaLcd.png","x":60,"y":60},"status":"valid","t":"sticker"}}"#).expect("Valid JSON");
let comment_input = r#"<div class="comment_body "><div class="md"><p>:2028:</p></div></div>"#;
let output = r#"<div class="comment_body "><div class="md"><p><img loading="lazy" src="/emote/t5_31hpy/PW6WsOaLcd.png" width="60" height="60" style="vertical-align:text-bottom"></p></div></div>"#;
assert_eq!(rewrite_emotes(&json_input, comment_input.to_string()), output);
}