fix(yt_dlp/post_processors): Register in python

We need to tell yt_dlp about our post processors, as they would
otherwise not take full effect. For example, changing the title would
previously only have changed the title in the *in-memory* info json, the
actual file on disk (video and .info.json) would still have the old
title, as yt_dlp did not know about our post processor.

Registering it via their api also has the upside of being able to
determine when to run.

2 files changed, 152 insertions, 26 deletions

diff --git a/crates/yt_dlp/src/post_processors/dearrow.rs b/crates/yt_dlp/src/post_processors/dearrow.rs
index 7dc6bbb..77c7ab9 100644
--- a/crates/yt_dlp/src/post_processors/dearrow.rs
+++ b/crates/yt_dlp/src/post_processors/dearrow.rs
@@ -16,22 +16,21 @@ use rustpython::vm::{
 };
 use serde::{Deserialize, Serialize};
 
-use crate::{InfoJson, json_get};
+use crate::{pydict_cast, pydict_get, wrap_post_processor};
 
-use super::PostProcessor;
+wrap_post_processor!("DeArrow", unwrapped_process, process);
 
-#[derive(Debug, Clone, Copy)]
-pub struct DeArrowPP;
-
-impl PostProcessor for DeArrowPP {
-    fn extractors(&self) -> &'static [&'static str] {
-        &["Youtube"]
+/// # 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);
     }
 
-    fn process(&self, mut info: InfoJson) -> Result<InfoJson, super::Error> {
-       let mut output: DeArrowApi = {
-           let output_bytes = {
-               let mut dst = Vec::new();
+    let mut output: DeArrowApi = {
+        let output_bytes = {
+            let mut dst = Vec::new();
 
             let mut easy = Easy::new();
             easy.url(
@@ -88,6 +87,41 @@ impl PostProcessor for DeArrowPP {
     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 {
diff --git a/crates/yt_dlp/src/post_processors/mod.rs b/crates/yt_dlp/src/post_processors/mod.rs
index 65801c2..575dc45 100644
--- a/crates/yt_dlp/src/post_processors/mod.rs
+++ b/crates/yt_dlp/src/post_processors/mod.rs
@@ -8,23 +8,115 @@
 // 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 crate::InfoJson;
-
 pub mod dearrow;
 
-pub trait PostProcessor: std::fmt::Debug + Send {
-    /// Process a [`InfoJson`] object and return the updated one.
-    ///
-    /// # Errors
-    /// If the processing steps failed.
-    fn process(&self, info: InfoJson) -> Result<InfoJson, Error>;
+#[macro_export]
+macro_rules! pydict_get {
+    (@$vm:expr, $value:expr, $name:literal, $into:ident) => {{
+        match $value.get_item($name, $vm) {
+            Ok(val) => $crate::pydict_cast!(val, $into),
+            Err(_) => panic!(
+                concat!(
+                    "Expected '",
+                    $name,
+                    "' to be a key for the'",
+                    stringify!($value),
+                    "' py dictionary: {:#?}"
+                ),
+                $value
+            ),
+        }
+    }};
+}
 
-    /// The supported extractors for this post processor
-    fn extractors(&self) -> &'static [&'static str];
+#[macro_export]
+macro_rules! pydict_cast {
+    ($value:expr, $into:ident) => {{
+        match $value.downcast::<$into>() {
+            Ok(result) => result,
+            Err(val) => panic!(
+                concat!(
+                    "Expected to be able to downcast value ({:#?}) as ",
+                    stringify!($into)
+                ),
+                val
+            ),
+        }
+    }};
+    (@ref $value:expr, $into:ident) => {{
+        match $value.downcast_ref::<$into>() {
+            Some(result) => result,
+            None => panic!(
+                concat!(
+                    "Expected to be able to downcast value ({:#?}) as ",
+                    stringify!($into)
+                ),
+                $value
+            ),
+        }
+    }};
 }
 
-#[derive(thiserror::Error, Debug)]
-pub enum Error {
-    #[error("Failed to access a api: {0}")]
-    Get(#[from] reqwest::Error),
+#[macro_export]
+macro_rules! wrap_post_processor {
+    ($name:literal, $unwrap:ident, $wrapped:ident) => {
+        use $crate::progress_hook::__priv::vm;
+
+        /// # Errors
+        /// - If the underlying function returns an error.
+        /// - If python operations fail.
+        pub fn $wrapped(vm: &vm::VirtualMachine) -> vm::PyResult<vm::PyObjectRef> {
+            fn actual_processor(
+                mut input: vm::function::FuncArgs,
+                vm: &vm::VirtualMachine,
+            ) -> vm::PyResult<vm::PyRef<vm::builtins::PyDict>> {
+                let input = input
+                    .args
+                    .remove(0)
+                    .downcast::<vm::builtins::PyDict>()
+                    .expect("Should be a py dict");
+
+                let output = match unwrapped_process(input, vm) {
+                    Ok(ok) => ok,
+                    Err(err) => {
+                        return Err(vm.new_runtime_error(err.to_string()));
+                    }
+                };
+
+                Ok(output)
+            }
+
+            let scope = vm.new_scope_with_builtins();
+
+            scope.globals.set_item(
+                "actual_processor",
+                vm.new_function("actual_processor", actual_processor).into(),
+                vm,
+            )?;
+
+            let local_scope = scope.clone();
+            vm.run_code_string(
+                local_scope,
+                format!(
+                    "
+import yt_dlp
+
+class {}(yt_dlp.postprocessor.PostProcessor):
+    def run(self, info):
+        info = actual_processor(info)
+        return [], info
+
+inst = {}()
+",
+                    $name, $name
+                ).as_str(),
+                "<embedded post processor initializing code>".to_owned(),
+            )?;
+
+            Ok(scope
+                .globals
+                .get_item("inst", vm)
+                .expect("We just declared it"))
+        }
+    };
 }