1 files changed, 184 insertions, 0 deletions
diff --git a/crates/yt_dlp/src/post_processors/dearrow.rs b/crates/yt_dlp/src/post_processors/dearrow.rs
new file mode 100644
index 0000000..3cac745
--- /dev/null
+++ b/crates/yt_dlp/src/post_processors/dearrow.rs
@@ -0,0 +1,184 @@
+// yt - A fully featured command line YouTube client
+//
+// Copyright (C) 2025 Benedikt Peetz <benedikt.peetz@b-peetz.de>
+// SPDX-License-Identifier: GPL-3.0-or-later
+//
+// This file is part of Yt.
+//
+// You should have received a copy of the License along with this program.
+// If not, see <https://www.gnu.org/licenses/gpl-3.0.txt>.
+
+use curl::easy::Easy;
+use log::{error, info, warn};
+use rustpython::vm::{
+    PyRef, VirtualMachine,
+    builtins::{PyDict, PyStr},
+};
+use serde::{Deserialize, Serialize};
+
+use crate::{pydict_cast, pydict_get, wrap_post_processor};
+
+wrap_post_processor!("DeArrow", unwrapped_process, process);
+
+/// # Errors
+/// If the API access fails.
+pub fn unwrapped_process(info: PyRef<PyDict>, vm: &VirtualMachine) -> Result<PyRef<PyDict>, Error> {
+    if pydict_get!(@vm, info, "extractor_key", PyStr).as_str() != "Youtube" {
+        warn!("DeArrow: Extractor did not match, exiting.");
+        return Ok(info);
+    }
+
+    let mut output: DeArrowApi = {
+        let output_bytes = {
+            let mut dst = Vec::new();
+
+            let mut easy = Easy::new();
+            easy.url(
+                format!(
+                    "https://sponsor.ajay.app/api/branding?videoID={}",
+                    pydict_get!(@vm, info, "id", PyStr).as_str()
+                )
+                .as_str(),
+            )?;
+
+            let mut transfer = easy.transfer();
+            transfer.write_function(|data| {
+                dst.extend_from_slice(data);
+                Ok(data.len())
+            })?;
+            transfer.perform()?;
+            drop(transfer);
+
+            dst
+        };
+
+        serde_json::from_slice(&output_bytes)?
+    };
+
+    // We pop the titles, so we need this vector reversed.
+    output.titles.reverse();
+
+    let title_len = output.titles.len();
+    let mut iterator = output.titles.clone();
+    let selected = loop {
+        let Some(title) = iterator.pop() else {
+            break false;
+        };
+
+        if (title.locked || title.votes < 1) && title_len > 1 {
+            info!(
+                "DeArrow: Skipping title {:#?}, as it is not good enough",
+                title.value
+            );
+            // Skip titles that are not “good” enough.
+            continue;
+        }
+
+        update_title(&info, &title.value, vm);
+
+        break true;
+    };
+
+    if !selected && title_len != 0 {
+        // No title was selected, even though we had some titles.
+        // Just pick the first one in this case.
+        update_title(&info, &output.titles[0].value, vm);
+    }
+
+    Ok(info)
+}
+
+#[derive(thiserror::Error, Debug)]
+pub enum Error {
+    #[error("Failed to access the DeArrow api: {0}")]
+    Get(#[from] curl::Error),
+
+    #[error("Failed to deserialize a api json return object: {0}")]
+    Deserialize(#[from] serde_json::Error),
+}
+
+fn update_title(info: &PyRef<PyDict>, new_title: &str, vm: &VirtualMachine) {
+    assert!(!info.contains_key("original_title", vm));
+
+    if let Ok(old_title) = info.get_item("title", vm) {
+        warn!(
+            "DeArrow: Updating title from {:#?} to {:#?}",
+            pydict_cast!(@ref old_title, PyStr).as_str(),
+            new_title
+        );
+
+        info.set_item("original_title", old_title, vm)
+            .expect("We checked, it is a new key");
+    } else {
+        warn!("DeArrow: Setting title to {new_title:#?}");
+    }
+
+    let cleaned_title = {
+        // NOTE(@bpeetz): DeArrow uses `>` as a “Don't format the next word” mark.
+        // They should be removed, if one does not use a auto-formatter. <2025-06-16>
+        new_title.replace('>', "")
+    };
+
+    info.set_item("title", vm.new_pyobj(cleaned_title), vm)
+        .expect("This should work?");
+}
+
+#[derive(Serialize, Deserialize)]
+/// See: <https://wiki.sponsor.ajay.app/w/API_Docs/DeArrow>
+struct DeArrowApi {
+    titles: Vec<Title>,
+    thumbnails: Vec<Thumbnail>,
+
+    #[serde(alias = "randomTime")]
+    random_time: Option<f64>,
+
+    #[serde(alias = "videoDuration")]
+    video_duration: Option<f64>,
+
+    #[serde(alias = "casualVotes")]
+    casual_votes: Vec<CasualVote>,
+}
+
+#[derive(Serialize, Deserialize)]
+struct CasualVote {
+    id: String,
+    count: u32,
+    title: String,
+}
+
+#[derive(Serialize, Deserialize, Clone)]
+struct Title {
+    /// Note: Titles will sometimes contain > before a word.
+    /// This tells the auto-formatter to not format a word.
+    /// If you have no auto-formatter, you can ignore this and replace it with an empty string
+    #[serde(alias = "title")]
+    value: String,
+
+    original: bool,
+    votes: u64,
+    locked: bool,
+
+    #[serde(alias = "UUID")]
+    uuid: String,
+
+    /// only present if requested
+    #[serde(alias = "userID")]
+    user_id: Option<String>,
+}
+
+#[derive(Serialize, Deserialize)]
+struct Thumbnail {
+    // null if original is true
+    timestamp: Option<f64>,
+
+    original: bool,
+    votes: u64,
+    locked: bool,
+
+    #[serde(alias = "UUID")]
+    uuid: String,
+
+    /// only present if requested
+    #[serde(alias = "userID")]
+    user_id: Option<String>,
+}