feat: forced migrate report

Forced migrate is a tool which creates a mapping from published nostr events to old void.cat links, this report can be used later to prompt the user to mirror their files to nostr.download

Cargo.toml

@@ -7,6 +7,10 @@ edition = "2021"
 name = "void_cat_migrate"
 required-features = ["bin-void-cat-migrate"]
 
+[[bin]]
+name = "void_cat_forced_migrate"
+required-features = ["bin-void-cat-force-migrate"]
+
 [[bin]]
 name = "route96"
 path = "src/bin/main.rs"
@@ -21,6 +25,7 @@ labels = ["nip96", "dep:candle-core", "dep:candle-nn", "dep:candle-transformers"
 nip96 = ["media-compression"]
 blossom = []
 bin-void-cat-migrate = ["dep:sqlx-postgres"]
+bin-void-cat-force-migrate = ["dep:regex", "dep:nostr-cursor"]
 torrent-v2 = []
 analytics = []
 void-cat-redirects = ["dep:sqlx-postgres"]
@@ -31,12 +36,12 @@ react-ui = []
 log = "0.4.21"
 nostr = "0.37.0"
 pretty_env_logger = "0.5.0"
-rocket = { version = "0.5.0", features = ["json"] }
+rocket = { version = "0.5.1", features = ["json"] }
 tokio = { version = "1.37.0", features = ["rt", "rt-multi-thread", "macros"] }
 base64 = "0.22.1"
 hex = { version = "0.4.3", features = ["serde"] }
 serde = { version = "1.0.198", features = ["derive"] }
-uuid = { version = "1.8.0", features = ["v4"] }
+uuid = { version = "1.8.0", features = ["v4", "serde"] }
 anyhow = "^1.0.82"
 sha2 = "0.10.8"
 sqlx = { version = "0.8.1", features = ["mysql", "runtime-tokio", "chrono", "uuid"] }
@@ -55,4 +60,6 @@ candle-nn = { git = "https://git.v0l.io/huggingface/candle.git", tag = "0.8.1",
 candle-transformers = { git = "https://git.v0l.io/huggingface/candle.git", tag = "0.8.1", optional = true }
 sqlx-postgres = { version = "0.8.2", optional = true, features = ["chrono", "uuid"] }
 http-range-header = { version = "0.4.2", optional = true }
+nostr-cursor = { git = "https://git.v0l.io/Kieran/nostr_backup_proc.git", branch = "main", optional = true }
+regex = { version = "1.11.1", optional = true }
 

src/bin/void_cat_forced_migrate.rs

@@ -0,0 +1,70 @@
+use clap::Parser;
+use log::{info, warn};
+use nostr::serde_json;
+use nostr_cursor::cursor::NostrCursor;
+use regex::Regex;
+use rocket::futures::StreamExt;
+use std::collections::{HashMap, HashSet};
+use std::path::PathBuf;
+use tokio::fs::File;
+use tokio::io::AsyncWriteExt;
+use uuid::Uuid;
+
+#[derive(Parser, Debug)]
+#[command(version, about)]
+struct ProgramArgs {
+    /// Directory pointing to archives to scan
+    #[arg(short, long)]
+    pub archive: PathBuf,
+
+    /// Output path .csv
+    #[arg(short, long)]
+    pub output: PathBuf,
+}
+
+#[tokio::main]
+async fn main() -> Result<(), anyhow::Error> {
+    pretty_env_logger::init();
+
+    let args: ProgramArgs = ProgramArgs::parse();
+
+    let mut report: HashMap<String, HashSet<Uuid>> = HashMap::new();
+
+    let mut binding = NostrCursor::new(args.archive);
+    let mut cursor = Box::pin(binding.walk());
+    let matcher = Regex::new(r"void\.cat/d/(\w+)")?;
+    while let Some(Ok(e)) = cursor.next().await {
+        if e.content.contains("void.cat") {
+            let links = matcher.captures_iter(&e.content).collect::<Vec<_>>();
+            for link in links {
+                let g = link.get(1).unwrap().as_str();
+                let base58 = if let Ok(b) = nostr::bitcoin::base58::decode(g) {
+                    b
+                } else {
+                    warn!("Invalid base58 id {}", g);
+                    continue;
+                };
+                let uuid = if let Ok(u) = Uuid::from_slice_le(base58.as_slice()) {
+                    u
+                } else {
+                    warn!("Invalid uuid {}", g);
+                    continue;
+                };
+                info!("Got link: {} => {}", g, e.pubkey);
+                if let Some(ur) = report.get_mut(&e.pubkey) {
+                    ur.insert(uuid);
+                } else {
+                    report.insert(e.pubkey.clone(), HashSet::from([uuid]));
+                }
+            }
+        }
+    }
+
+    let json = serde_json::to_string(&report)?;
+    File::create(args.output)
+        .await?
+        .write_all(json.as_bytes())
+        .await?;
+
+    Ok(())
+}