diff options
28 files changed, 1216 insertions, 2019 deletions
diff --git a/Cargo.toml b/Cargo.toml index 403c8f1..222ef7e 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -59,6 +59,10 @@ codegen-units = 1 panic = "abort" split-debuginfo = "off" +[profile.dev] +# Otherwise, yt_dlp is just too slow +opt-level = 2 + [workspace.lints.rust] # rustc lint groups https://doc.rust-lang.org/rustc/lints/groups.html warnings = "warn" diff --git a/crates/yt_dlp/.cargo/config.toml b/crates/yt_dlp/.cargo/config.toml deleted file mode 100644 index d84f14d..0000000 --- a/crates/yt_dlp/.cargo/config.toml +++ /dev/null @@ -1,12 +0,0 @@ -# yt - A fully featured command line YouTube client -# -# Copyright (C) 2024 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>. - -[env] -PYO3_PYTHON = "/nix/store/7xzk119acyws2c4ysygdv66l0grxkr39-python3-3.11.9-env/bin/python3" diff --git a/crates/yt_dlp/Cargo.toml b/crates/yt_dlp/Cargo.toml index b80c70f..ddd5f9b 100644 --- a/crates/yt_dlp/Cargo.toml +++ b/crates/yt_dlp/Cargo.toml @@ -10,7 +10,7 @@ [package] name = "yt_dlp" -description = "A wrapper around the python yt_dlp library" +description = "A rust fii wrapper library for the python yt_dlp library" keywords = [] categories = [] version.workspace = true @@ -19,19 +19,16 @@ authors.workspace = true license.workspace = true repository.workspace = true rust-version.workspace = true -publish = false +publish = true [dependencies] -pyo3 = { version = "0.24.0", features = ["auto-initialize"] } -bytes.workspace = true +indexmap = { version = "2.9.0", default-features = false } log.workspace = true -serde.workspace = true +rustpython = { git = "https://github.com/RustPython/RustPython.git", features = ["threading", "stdlib", "stdio", "importlib", "ssl"], default-features = false } serde_json.workspace = true +thiserror = "2.0.12" url.workspace = true -[dev-dependencies] -tokio.workspace = true - [lints] workspace = true diff --git a/crates/yt_dlp/src/duration.rs b/crates/yt_dlp/src/duration.rs deleted file mode 100644 index 19181a5..0000000 --- a/crates/yt_dlp/src/duration.rs +++ /dev/null @@ -1,78 +0,0 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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>. - -// TODO: This file should be de-duplicated with the same file in the 'yt' crate <2024-06-25> - -#[derive(Debug, Clone, Copy)] -pub struct Duration { - time: u32, -} - -impl From<&str> for Duration { - fn from(v: &str) -> Self { - let buf: Vec<_> = v.split(':').take(2).collect(); - Self { - time: (buf[0].parse::<u32>().expect("Should be a number") * 60) - + buf[1].parse::<u32>().expect("Should be a number"), - } - } -} - -impl From<Option<f64>> for Duration { - fn from(value: Option<f64>) -> Self { - Self { - #[allow( - clippy::cast_possible_truncation, - clippy::cast_precision_loss, - clippy::cast_sign_loss - )] - time: value.unwrap_or(0.0).ceil() as u32, - } - } -} - -impl std::fmt::Display for Duration { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> Result<(), std::fmt::Error> { - const SECOND: u32 = 1; - const MINUTE: u32 = 60 * SECOND; - const HOUR: u32 = 60 * MINUTE; - - let base_hour = self.time - (self.time % HOUR); - let base_min = (self.time % HOUR) - ((self.time % HOUR) % MINUTE); - let base_sec = (self.time % HOUR) % MINUTE; - - let h = base_hour / HOUR; - let m = base_min / MINUTE; - let s = base_sec / SECOND; - - if self.time == 0 { - write!(f, "0s") - } else if h > 0 { - write!(f, "{h}h {m}m") - } else { - write!(f, "{m}m {s}s") - } - } -} -#[cfg(test)] -mod test { - use super::Duration; - - #[test] - fn test_display_duration_1h() { - let dur = Duration { time: 60 * 60 }; - assert_eq!("1h 0m".to_owned(), dur.to_string()); - } - #[test] - fn test_display_duration_30min() { - let dur = Duration { time: 60 * 30 }; - assert_eq!("30m 0s".to_owned(), dur.to_string()); - } -} diff --git a/crates/yt_dlp/src/error.rs b/crates/yt_dlp/src/error.rs deleted file mode 100644 index 3881f0b..0000000 --- a/crates/yt_dlp/src/error.rs +++ /dev/null @@ -1,68 +0,0 @@ -// 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 std::{fmt::Display, io}; - -use pyo3::Python; - -#[derive(Debug)] -#[allow(clippy::module_name_repetitions)] -pub enum YtDlpError { - ResponseParseError { - error: serde_json::error::Error, - }, - PythonError { - error: Box<pyo3::PyErr>, - kind: String, - }, - IoError { - error: io::Error, - }, -} - -impl std::error::Error for YtDlpError {} - -impl Display for YtDlpError { - fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { - match self { - YtDlpError::ResponseParseError { error } => write!( - f, - include_str!("./python_json_decode_failed.error_msg"), - error - ), - YtDlpError::PythonError { error, kind: _ } => write!(f, "Python error: {error}"), - YtDlpError::IoError { error } => write!(f, "Io error: {error}"), - } - } -} - -impl From<serde_json::error::Error> for YtDlpError { - fn from(value: serde_json::error::Error) -> Self { - Self::ResponseParseError { error: value } - } -} - -impl From<pyo3::PyErr> for YtDlpError { - fn from(value: pyo3::PyErr) -> Self { - Python::with_gil(|py| { - let kind = value.get_type(py).to_string(); - Self::PythonError { - error: Box::new(value), - kind, - } - }) - } -} - -impl From<io::Error> for YtDlpError { - fn from(value: io::Error) -> Self { - Self::IoError { error: value } - } -} diff --git a/crates/yt_dlp/src/lib.rs b/crates/yt_dlp/src/lib.rs index c6d9290..34b8a5d 100644 --- a/crates/yt_dlp/src/lib.rs +++ b/crates/yt_dlp/src/lib.rs @@ -1,549 +1,541 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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>. - -// The pyo3 `pyfunction` proc-macros call unsafe functions internally, which trigger this lint. -#![allow(unsafe_op_in_unsafe_fn)] -#![allow(clippy::missing_errors_doc)] - -use std::io::stderr; -use std::{env, process}; -use std::{fs::File, io::Write}; - -use std::{path::PathBuf, sync::Once}; - -use crate::{duration::Duration, logging::setup_logging, wrapper::info_json::InfoJson}; - -use bytes::Bytes; -use error::YtDlpError; -use log::{Level, debug, info, log_enabled}; -use pyo3::types::{PyString, PyTuple, PyTupleMethods}; -use pyo3::{ - Bound, PyAny, PyResult, Python, pyfunction, - types::{PyAnyMethods, PyDict, PyDictMethods, PyList, PyListMethods, PyModule}, - wrap_pyfunction, +//! The `yt_dlp` interface is completely contained in the [`YoutubeDL`] structure. + +use std::io::Write; +use std::mem; +use std::{env, fs::File, path::PathBuf}; + +use indexmap::IndexMap; +use log::{Level, debug, error, info, log_enabled}; +use logging::setup_logging; +use rustpython::vm::builtins::PyList; +use rustpython::{ + InterpreterConfig, + vm::{ + self, Interpreter, PyObjectRef, PyRef, VirtualMachine, + builtins::{PyBaseException, PyDict, PyStr}, + function::{FuncArgs, KwArgs, PosArgs}, + }, }; -use serde::Serialize; -use serde_json::{Map, Value}; use url::Url; -pub mod duration; -pub mod error; -pub mod logging; -pub mod wrapper; +mod logging; +pub mod progress_hook; -#[cfg(test)] -mod tests; - -/// Synchronisation helper, to ensure that we don't setup the logger multiple times -static SYNC_OBJ: Once = Once::new(); +#[macro_export] +macro_rules! json_get { + ($value:expr, $name:literal, $into:ident) => { + $crate::json_cast!($value.get($name).expect("Should exist"), $into) + }; +} -/// Add a logger to the yt-dlp options. -/// If you have an logger set (i.e. for rust), than this will log to rust -/// -/// # Panics -/// This should never panic. -pub fn add_logger_and_sig_handler<'a>( - opts: Bound<'a, PyDict>, - py: Python<'_>, -) -> PyResult<Bound<'a, PyDict>> { - /// Is the specified record to be logged? Returns false for no, - /// true for yes. Filters can either modify log records in-place or - /// return a completely different record instance which will replace - /// the original log record in any future processing of the event. - #[pyfunction] - fn filter_error_log(_py: Python<'_>, record: &Bound<'_, PyAny>) -> bool { - // Filter out all error logs (they are propagated as rust errors) - let levelname: String = record - .getattr("levelname") - .expect("This should exist") - .extract() - .expect("This should be a String"); - - let return_value = levelname.as_str() != "ERROR"; - - if log_enabled!(Level::Debug) && !return_value { - let message: String = record - .call_method0("getMessage") - .expect("This method exists") - .extract() - .expect("The message is a string"); - - debug!("Swollowed error message: '{message}'"); - } - return_value - } +#[macro_export] +macro_rules! json_cast { + ($value:expr, $into:ident) => { + $value.$into().expect(concat!( + "Should be able to cast value into ", + stringify!($into) + )) + }; +} - setup_logging(py, "yt_dlp")?; - - let logging = PyModule::import(py, "logging")?; - let ytdl_logger = logging.call_method1("getLogger", ("yt_dlp",))?; - - // Ensure that all events are logged by setting the log level to NOTSET (we filter on rust's side) - // Also use this static, to ensure that we don't configure the logger every time - SYNC_OBJ.call_once(|| { - // Disable the SIGINT (Ctrl+C) handler, python installs. - // This allows the user to actually stop the application with Ctrl+C. - // This is here because it can only be run in the main thread and this was here already. - py.run( - c"\ -import signal -signal.signal(signal.SIGINT, signal.SIG_DFL)", - None, - None, - ) - .expect("This code should always work"); - - let config_opts = PyDict::new(py); - config_opts - .set_item("level", 0) - .expect("Setting this item should always work"); - - logging - .call_method("basicConfig", (), Some(&config_opts)) - .expect("This method exists"); - }); - - ytdl_logger.call_method1( - "addFilter", - (wrap_pyfunction!(filter_error_log, py).expect("This function can be wrapped"),), - )?; - - // This was taken from `ytcc`, I don't think it is still applicable - // ytdl_logger.setattr("propagate", false)?; - // let logging_null_handler = logging.call_method0("NullHandler")?; - // ytdl_logger.setattr("addHandler", logging_null_handler)?; - - opts.set_item("logger", ytdl_logger).expect("Should work"); - - Ok(opts) +/// The core of the `yt_dlp` interface. +pub struct YoutubeDL { + interpreter: Interpreter, + youtube_dl_class: PyObjectRef, + yt_dlp_module: PyObjectRef, + options: serde_json::Map<String, serde_json::Value>, } -#[pyfunction] -#[allow(clippy::too_many_lines)] -#[allow(clippy::missing_panics_doc)] -#[allow(clippy::items_after_statements)] -#[allow( - clippy::cast_possible_truncation, - clippy::cast_sign_loss, - clippy::cast_precision_loss -)] -pub fn progress_hook(py: Python<'_>, input: &Bound<'_, PyDict>) -> PyResult<()> { - // Only add the handler, if the log-level is higher than Debug (this avoids covering debug - // messages). - if log_enabled!(Level::Debug) { - return Ok(()); +impl std::fmt::Debug for YoutubeDL { + fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result { + // TODO(@bpeetz): Use something useful here. <2025-06-13> + f.write_str("YoutubeDL") } +} - // ANSI ESCAPE CODES Wrappers {{{ - // see: https://en.wikipedia.org/wiki/ANSI_escape_code#Control_Sequence_Introducer_commands - const CSI: &str = "\x1b["; - fn clear_whole_line() { - eprint!("{CSI}2K"); - } - fn move_to_col(x: usize) { - eprint!("{CSI}{x}G"); - } - // }}} - - let input: Map<String, Value> = serde_json::from_str(&json_dumps( - py, - input - .downcast::<PyAny>() - .expect("Will always work") - .to_owned(), - )?) - .expect("python's json is valid"); - - macro_rules! get { - (@interrogate $item:ident, $type_fun:ident, $get_fun:ident, $name:expr) => {{ - let a = $item.get($name).expect(concat!( - "The field '", - stringify!($name), - "' should exist." - )); - - if a.$type_fun() { - a.$get_fun().expect( - "The should have been checked in the if guard, so unpacking here is fine", - ) - } else { - panic!( - "Value {} => \n{}\n is not of type: {}", - $name, - a, - stringify!($type_fun) - ); +impl YoutubeDL { + /// Construct this instance from options. + /// + /// # Panics + /// If `yt_dlp` changed their interface. + /// + /// # Errors + /// If a python call fails. + pub fn from_options(mut options: YoutubeDLOptions) -> Result<Self, build::Error> { + let mut settings = vm::Settings::default(); + if let Ok(python_path) = env::var("PYTHONPATH") { + for path in python_path.split(':') { + settings.path_list.push(path.to_owned()); } - }}; + } else { + error!( + "No PYTHONPATH found or invalid utf8. \ + This means, that you probably did not \ + supply the yt_dlp!" + ); + } - ($type_fun:ident, $get_fun:ident, $name1:expr, $name2:expr) => {{ - let a = get! {@interrogate input, is_object, as_object, $name1}; - let b = get! {@interrogate a, $type_fun, $get_fun, $name2}; - b - }}; + settings.install_signal_handlers = false; - ($type_fun:ident, $get_fun:ident, $name:expr) => {{ - get! {@interrogate input, $type_fun, $get_fun, $name} - }}; - } + // NOTE(@bpeetz): Another value leads to an internal codegen error. <2025-06-13> + settings.optimize = 0; - macro_rules! default_get { - (@interrogate $item:ident, $default:expr, $get_fun:ident, $name:expr) => {{ - let a = if let Some(field) = $item.get($name) { - field.$get_fun().unwrap_or($default) - } else { - $default - }; - a - }}; - - ($get_fun:ident, $default:expr, $name1:expr, $name2:expr) => {{ - let a = get! {@interrogate input, is_object, as_object, $name1}; - let b = default_get! {@interrogate a, $default, $get_fun, $name2}; - b - }}; - - ($get_fun:ident, $default:expr, $name:expr) => {{ - default_get! {@interrogate input, $default, $get_fun, $name} - }}; - } + settings.isolated = true; - macro_rules! c { - ($color:expr, $format:expr) => { - format!("\x1b[{}m{}\x1b[0m", $color, $format) - }; - } + let interpreter = InterpreterConfig::new() + .init_stdlib() + .settings(settings) + .interpreter(); - fn format_bytes(bytes: u64) -> String { - let bytes = Bytes::new(bytes); - bytes.to_string() - } + let output_options = options.options.clone(); - fn format_speed(speed: f64) -> String { - #[allow(clippy::cast_possible_truncation, clippy::cast_sign_loss)] - let bytes = Bytes::new(speed.floor() as u64); - format!("{bytes}/s") - } + let (yt_dlp_module, youtube_dl_class) = match interpreter.enter(|vm| { + let yt_dlp_module = vm.import("yt_dlp", 0)?; + let class = yt_dlp_module.get_attr("YoutubeDL", vm)?; - let get_title = || -> String { - match get! {is_string, as_str, "info_dict", "ext"} { - "vtt" => { - format!( - "Subtitles ({})", - default_get! {as_str, "<No Subtitle Language>", "info_dict", "name"} - ) + let maybe_hook = mem::take(&mut options.progress_hook); + let opts = options.into_py_dict(vm); + if let Some(function) = maybe_hook { + opts.get_or_insert(vm, vm.new_pyobj("progress_hooks"), || { + let hook: PyObjectRef = vm.new_function("progress_hook", function).into(); + vm.new_pyobj(vec![hook]) + }) + .expect("Should work?"); } - "webm" | "mp4" | "mp3" | "m4a" => { - default_get! { as_str, "<No title>", "info_dict", "title"}.to_owned() - } - other => panic!("The extension '{other}' is not yet implemented"), - } - }; - match get! {is_string, as_str, "status"} { - "downloading" => { - let elapsed = default_get! {as_f64, 0.0f64, "elapsed"}; - let eta = default_get! {as_f64, 0.0, "eta"}; - let speed = default_get! {as_f64, 0.0, "speed"}; - - let downloaded_bytes = get! {is_u64, as_u64, "downloaded_bytes"}; - let (total_bytes, bytes_is_estimate): (u64, &'static str) = { - let total_bytes = default_get!(as_u64, 0, "total_bytes"); - if total_bytes == 0 { - let maybe_estimate = default_get!(as_u64, 0, "total_bytes_estimate"); - - if maybe_estimate == 0 { - // The download speed should be in bytes per second and the eta in seconds. - // Thus multiplying them gets us the raw bytes (which were estimated by `yt_dlp`, from their `info.json`) - let bytes_still_needed = (speed * eta).ceil() as u64; - - (downloaded_bytes + bytes_still_needed, "~") - } else { - (maybe_estimate, "~") + { + // Unconditionally set a logger. + // Otherwise, yt_dlp will log to stderr. + + /// Is the specified record to be logged? Returns false for no, + /// true for yes. Filters can either modify log records in-place or + /// return a completely different record instance which will replace + /// the original log record in any future processing of the event. + fn filter_error_log(mut input: FuncArgs, vm: &VirtualMachine) -> bool { + let record = input.args.remove(0); + + // Filter out all error logs (they are propagated as rust errors) + let levelname: PyRef<PyStr> = record + .get_attr("levelname", vm) + .expect("This should exist") + .downcast() + .expect("This should be a String"); + + let return_value = levelname.as_str() != "ERROR"; + + if log_enabled!(Level::Debug) && !return_value { + let message: String = { + let get_message = record.get_attr("getMessage", vm).expect("Is set"); + let message: PyRef<PyStr> = get_message + .call((), vm) + .expect("Can be called") + .downcast() + .expect("Downcasting works"); + + message.as_str().to_owned() + }; + + debug!("Swollowed error message: '{message}'"); } - } else { - (total_bytes, "") + return_value } - }; - let percent: f64 = { - if total_bytes == 0 { - 100.0 - } else { - (downloaded_bytes as f64 / total_bytes as f64) * 100.0 + + let logging = setup_logging(vm, "yt_dlp")?; + let ytdl_logger = { + let get_logger = logging.get_item("getLogger", vm)?; + get_logger.call(("yt_dlp",), vm)? + }; + + { + let args = FuncArgs::new( + PosArgs::new(vec![]), + KwArgs::new({ + let mut map = IndexMap::new(); + // Ensure that all events are logged by setting + // the log level to NOTSET (we filter on rust's side) + map.insert("level".to_owned(), vm.new_pyobj(0)); + map + }), + ); + + let basic_config = logging.get_item("basicConfig", vm)?; + basic_config.call(args, vm)?; } - }; - clear_whole_line(); - move_to_col(1); - - eprint!( - "'{}' [{}/{} at {}] -> [{} of {}{} {}] ", - c!("34;1", get_title()), - c!("33;1", Duration::from(Some(elapsed))), - c!("33;1", Duration::from(Some(eta))), - c!("32;1", format_speed(speed)), - c!("31;1", format_bytes(downloaded_bytes)), - c!("31;1", bytes_is_estimate), - c!("31;1", format_bytes(total_bytes)), - c!("36;1", format!("{:.02}%", percent)) - ); - stderr().flush()?; - } - "finished" => { - eprintln!("-> Finished downloading."); - } - "error" => { - // TODO: This should probably return an Err. But I'm not so sure where the error would - // bubble up to (i.e., who would catch it) <2025-01-21> - eprintln!("-> Error while downloading: {}", get_title()); - process::exit(1); - } - other => unreachable!("'{other}' should not be a valid state!"), - }; + { + let add_filter = ytdl_logger.get_attr("addFilter", vm)?; + add_filter.call( + (vm.new_function("yt_dlp_error_filter", filter_error_log),), + vm, + )?; + } - Ok(()) -} + opts.set_item("logger", ytdl_logger, vm)?; + } -pub fn add_hooks<'a>(opts: Bound<'a, PyDict>, py: Python<'_>) -> PyResult<Bound<'a, PyDict>> { - if let Some(hooks) = opts.get_item("progress_hooks")? { - let hooks = hooks.downcast::<PyList>()?; - hooks.append(wrap_pyfunction!(progress_hook, py)?)?; + let youtube_dl_class = class.call((opts,), vm)?; - opts.set_item("progress_hooks", hooks)?; - } else { - // No hooks are set yet - let hooks_list = PyList::new(py, &[wrap_pyfunction!(progress_hook, py)?])?; + Ok::<_, PyRef<PyBaseException>>((yt_dlp_module, youtube_dl_class)) + }) { + Ok(ok) => ok, + Err(err) => { + interpreter.finalize(Some(err)); + return Err(build::Error::Python); + } + }; - opts.set_item("progress_hooks", hooks_list)?; + Ok(Self { + interpreter, + youtube_dl_class, + yt_dlp_module, + options: output_options, + }) } - Ok(opts) -} - -/// Take the result of the ie (may be modified) and resolve all unresolved -/// references (URLs, playlist items). -/// -/// It will also download the videos if 'download'. -/// Returns the resolved `ie_result`. -#[allow(clippy::missing_panics_doc)] -pub fn process_ie_result( - yt_dlp_opts: &Map<String, Value>, - ie_result: InfoJson, - download: bool, -) -> Result<InfoJson, YtDlpError> { - Python::with_gil(|py| -> Result<InfoJson, YtDlpError> { - let opts = json_map_to_py_dict(yt_dlp_opts, py)?; - - let instance = get_yt_dlp(py, opts)?; - - let args = { - let ie_result = json_loads_str(py, ie_result)?; - (ie_result,) - }; + /// # Panics + /// + /// If `yt_dlp` changed their location or type of `__version__`. + pub fn version(&self) -> String { + let str_ref: PyRef<PyStr> = self.interpreter.enter_and_expect( + |vm| { + let version_module = self.yt_dlp_module.get_attr("version", vm)?; + let version = version_module.get_attr("__version__", vm)?; + let version = version.downcast().expect("This should always be a string"); + Ok(version) + }, + "yt_dlp version location has changed", + ); + str_ref.to_string() + } - let kwargs = PyDict::new(py); - kwargs.set_item("download", download)?; + /// Download a given list of URLs. + /// Returns the paths they were downloaded to. + /// + /// # Errors + /// If one of the downloads error. + pub fn download(&self, urls: &[Url]) -> Result<Vec<PathBuf>, extract_info::Error> { + let mut out_paths = Vec::with_capacity(urls.len()); + + for url in urls { + info!("Started downloading url: '{url}'"); + let info_json = self.extract_info(url, true, true)?; + + // Try to work around yt-dlp type weirdness + let result_string = if let Some(filename) = info_json.get("filename") { + PathBuf::from(json_cast!(filename, as_str)) + } else { + PathBuf::from(json_get!( + json_cast!( + json_get!(info_json, "requested_downloads", as_array)[0], + as_object + ), + "filename", + as_str + )) + }; - let result = instance - .call_method("process_ie_result", args, Some(&kwargs))? - .downcast_into::<PyDict>() - .expect("This is a dict"); + out_paths.push(result_string); + info!("Finished downloading url"); + } - let result_str = json_dumps(py, result.into_any())?; + Ok(out_paths) + } - serde_json::from_str(&result_str).map_err(Into::into) - }) -} + /// `extract_info(self, url, download=True, ie_key=None, extra_info=None, process=True, force_generic_extractor=False)` + /// + /// Extract and return the information dictionary of the URL + /// + /// Arguments: + /// - `url` URL to extract + /// + /// Keyword arguments: + /// :`download` Whether to download videos + /// :`process` Whether to resolve all unresolved references (URLs, playlist items). + /// Must be True for download to work + /// + /// # Panics + /// If expectations about python fail to hold. + /// + /// # Errors + /// If python operations fail. + pub fn extract_info( + &self, + url: &Url, + download: bool, + process: bool, + ) -> Result<InfoJson, extract_info::Error> { + match self.interpreter.enter(|vm| { + let pos_args = PosArgs::new(vec![vm.new_pyobj(url.to_string())]); + + let kw_args = KwArgs::new({ + let mut map = IndexMap::new(); + map.insert("download".to_owned(), vm.new_pyobj(download)); + map.insert("process".to_owned(), vm.new_pyobj(process)); + map + }); + + let fun_args = FuncArgs::new(pos_args, kw_args); + + let inner = self.youtube_dl_class.get_attr("extract_info", vm)?; + let result = inner + .call_with_args(fun_args, vm)? + .downcast::<PyDict>() + .expect("This is a dict"); + + // Resolve the generator object + if let Ok(generator) = result.get_item("entries", vm) { + if generator.payload_is::<PyList>() { + // already resolved. Do nothing + } else { + let max_backlog = self.options.get("playlistend").map_or(10000, |value| { + usize::try_from(value.as_u64().expect("Works")).expect("Should work") + }); + + let mut out = vec![]; + let next = generator.get_attr("__next__", vm)?; + while let Ok(output) = next.call((), vm) { + out.push(output); + + if out.len() == max_backlog { + break; + } + } + result.set_item("entries", vm.new_pyobj(out), vm)?; + } + } -/// `extract_info(self, url, download=True, ie_key=None, extra_info=None, process=True, force_generic_extractor=False)` -/// -/// Extract and return the information dictionary of the URL -/// -/// Arguments: -/// @param url URL to extract -/// -/// Keyword arguments: -/// @param download Whether to download videos -/// @param process Whether to resolve all unresolved references (URLs, playlist items). -/// Must be True for download to work -/// @param `ie_key` Use only the extractor with this key -/// -/// @param `extra_info` Dictionary containing the extra values to add to the info (For internal use only) -/// @`force_generic_extractor` Force using the generic extractor (Deprecated; use `ie_key`='Generic') -#[allow(clippy::missing_panics_doc)] -pub fn extract_info( - yt_dlp_opts: &Map<String, Value>, - url: &Url, - download: bool, - process: bool, -) -> Result<InfoJson, YtDlpError> { - Python::with_gil(|py| -> Result<InfoJson, YtDlpError> { - let opts = json_map_to_py_dict(yt_dlp_opts, py)?; - - let instance = get_yt_dlp(py, opts)?; - let args = (url.as_str(),); - - let kwargs = PyDict::new(py); - kwargs.set_item("download", download)?; - kwargs.set_item("process", process)?; - - let result = instance - .call_method("extract_info", args, Some(&kwargs))? - .downcast_into::<PyDict>() - .expect("This is a dict"); - - // Resolve the generator object - if let Some(generator) = result.get_item("entries")? { - if generator.is_instance_of::<PyList>() { - // already resolved. Do nothing - } else { - let max_backlog = yt_dlp_opts.get("playlistend").map_or(10000, |value| { - usize::try_from(value.as_u64().expect("Works")).expect("Should work") - }); + let result = { + let sanitize = self.youtube_dl_class.get_attr("sanitize_info", vm)?; + let value = sanitize.call((result,), vm)?; - let mut out = vec![]; - while let Ok(output) = generator.call_method0("__next__") { - out.push(output); + value.downcast::<PyDict>().expect("This should stay a dict") + }; - if out.len() == max_backlog { - break; - } + let result_json = json_dumps(result, vm); + + if let Ok(confirm) = env::var("YT_STORE_INFO_JSON") { + if confirm == "yes" { + let mut file = File::create("output.info.json").unwrap(); + write!( + file, + "{}", + serde_json::to_string_pretty(&serde_json::Value::Object( + result_json.clone() + )) + .expect("Valid json") + ) + .unwrap(); } - result.set_item("entries", out)?; + } + + Ok::<_, PyRef<PyBaseException>>(result_json) + }) { + Ok(ok) => Ok(ok), + Err(err) => { + self.interpreter.enter(|vm| { + vm.print_exception(err); + }); + Err(extract_info::Error::Python) } } + } + + /// Take the (potentially modified) result of the information extractor (i.e., + /// [`Self::extract_info`] with `process` and `download` set to false) + /// and resolve all unresolved references (URLs, + /// playlist items). + /// + /// It will also download the videos if 'download' is true. + /// Returns the resolved `ie_result`. + /// + /// # Panics + /// If expectations about python fail to hold. + /// + /// # Errors + /// If python operations fail. + pub fn process_ie_result( + &self, + ie_result: InfoJson, + download: bool, + ) -> Result<InfoJson, process_ie_result::Error> { + match self.interpreter.enter(|vm| { + let pos_args = PosArgs::new(vec![vm.new_pyobj(json_loads(ie_result, vm))]); + + let kw_args = KwArgs::new({ + let mut map = IndexMap::new(); + map.insert("download".to_owned(), vm.new_pyobj(download)); + map + }); + + let fun_args = FuncArgs::new(pos_args, kw_args); + + let inner = self.youtube_dl_class.get_attr("process_ie_result", vm)?; + let result = inner + .call_with_args(fun_args, vm)? + .downcast::<PyDict>() + .expect("This is a dict"); + + let result = { + let sanitize = self.youtube_dl_class.get_attr("sanitize_info", vm)?; + let value = sanitize.call((result,), vm)?; + + value.downcast::<PyDict>().expect("This should stay a dict") + }; - let result_str = json_dumps(py, result.into_any())?; + let result_json = json_dumps(result, vm); - if let Ok(confirm) = env::var("YT_STORE_INFO_JSON") { - if confirm == "yes" { - let mut file = File::create("output.info.json")?; - write!(file, "{result_str}").unwrap(); + Ok::<_, PyRef<PyBaseException>>(result_json) + }) { + Ok(ok) => Ok(ok), + Err(err) => { + self.interpreter.enter(|vm| { + vm.print_exception(err); + }); + Err(process_ie_result::Error::Python) } } - - serde_json::from_str(&result_str).map_err(Into::into) - }) + } } -/// # Panics -/// Only if python fails to return a valid URL. -pub fn unsmuggle_url(smug_url: &Url) -> PyResult<Url> { - Python::with_gil(|py| { - let utils = get_yt_dlp_utils(py)?; - let url = utils - .call_method1("unsmuggle_url", (smug_url.as_str(),))? - .downcast::<PyTuple>()? - .get_item(0)?; - - let url: Url = url - .downcast::<PyString>()? - .to_string() - .parse() - .expect("Python should be able to return a valid url"); - - Ok(url) - }) +#[allow(missing_docs)] +pub mod process_ie_result { + #[derive(Debug, thiserror::Error, Clone, Copy)] + pub enum Error { + #[error("Python threw an exception")] + Python, + } } - -/// Download a given list of URLs. -/// Returns the paths they were downloaded to. -/// -/// # Panics -/// Only if `yt_dlp` changes their `info_json` schema. -pub fn download( - urls: &[Url], - download_options: &Map<String, Value>, -) -> Result<Vec<PathBuf>, YtDlpError> { - let mut out_paths = Vec::with_capacity(urls.len()); - - for url in urls { - info!("Started downloading url: '{}'", url); - let info_json = extract_info(download_options, url, true, true)?; - - // Try to work around yt-dlp type weirdness - let result_string = if let Some(filename) = info_json.filename { - filename - } else { - info_json.requested_downloads.expect("This must exist")[0] - .filename - .clone() - }; - - out_paths.push(result_string); - info!("Finished downloading url: '{}'", url); +#[allow(missing_docs)] +pub mod extract_info { + #[derive(Debug, thiserror::Error, Clone, Copy)] + pub enum Error { + #[error("Python threw an exception")] + Python, } - - Ok(out_paths) } -fn json_map_to_py_dict<'a>( - map: &Map<String, Value>, - py: Python<'a>, -) -> PyResult<Bound<'a, PyDict>> { - let json_string = serde_json::to_string(&map).expect("This must always work"); +pub type InfoJson = serde_json::Map<String, serde_json::Value>; +pub type ProgressHookFunction = fn(input: FuncArgs, vm: &VirtualMachine); - let python_dict = json_loads(py, json_string)?; - - Ok(python_dict) +/// Options, that are used to customize the download behaviour. +/// +/// In the future, this might get a Builder api. +/// +/// See `help(yt_dlp.YoutubeDL())` from python for a full list of available options. +#[derive(Default, Debug)] +pub struct YoutubeDLOptions { + options: serde_json::Map<String, serde_json::Value>, + progress_hook: Option<ProgressHookFunction>, } -fn json_dumps(py: Python<'_>, input: Bound<'_, PyAny>) -> PyResult<String> { - // json.dumps(yt_dlp.sanitize_info(input)) +impl YoutubeDLOptions { + #[must_use] + pub fn new() -> Self { + Self { + options: serde_json::Map::new(), + progress_hook: None, + } + } - let yt_dlp = get_yt_dlp(py, PyDict::new(py))?; - let sanitized_result = yt_dlp.call_method1("sanitize_info", (input,))?; + #[must_use] + pub fn set(self, key: impl Into<String>, value: impl Into<serde_json::Value>) -> Self { + let mut options = self.options; + options.insert(key.into(), value.into()); - let json = PyModule::import(py, "json")?; - let dumps = json.getattr("dumps")?; + Self { + options, + progress_hook: self.progress_hook, + } + } - let output = dumps.call1((sanitized_result,))?; + #[must_use] + pub fn with_progress_hook(self, progress_hook: ProgressHookFunction) -> Self { + if let Some(_previous_hook) = self.progress_hook { + todo!() + } else { + Self { + options: self.options, + progress_hook: Some(progress_hook), + } + } + } - let output_str = output.extract::<String>()?; + /// # Errors + /// If the underlying [`YoutubeDL::from_options`] errors. + pub fn build(self) -> Result<YoutubeDL, build::Error> { + YoutubeDL::from_options(self) + } - Ok(output_str) -} + #[must_use] + pub fn from_json_options(options: serde_json::Map<String, serde_json::Value>) -> Self { + Self { + options, + progress_hook: None, + } + } -fn json_loads_str<T: Serialize>(py: Python<'_>, input: T) -> PyResult<Bound<'_, PyDict>> { - let string = serde_json::to_string(&input).expect("Correct json must be pased"); + #[must_use] + pub fn get(&self, key: &str) -> Option<&serde_json::Value> { + self.options.get(key) + } - json_loads(py, string) + fn into_py_dict(self, vm: &VirtualMachine) -> PyRef<PyDict> { + json_loads(self.options, vm) + } } -fn json_loads(py: Python<'_>, input: String) -> PyResult<Bound<'_, PyDict>> { - // json.loads(input) - - let json = PyModule::import(py, "json")?; - let dumps = json.getattr("loads")?; +#[allow(missing_docs)] +pub mod build { + #[derive(Debug, thiserror::Error)] + pub enum Error { + #[error("Python threw an exception")] + Python, - let output = dumps.call1((input,))?; - - Ok(output - .downcast::<PyDict>() - .expect("This should always be a PyDict") - .clone()) + #[error("Io error: {0}")] + Io(#[from] std::io::Error), + } } -fn get_yt_dlp_utils(py: Python<'_>) -> PyResult<Bound<'_, PyAny>> { - let yt_dlp = PyModule::import(py, "yt_dlp")?; - let utils = yt_dlp.getattr("utils")?; - - Ok(utils) +fn json_loads( + input: serde_json::Map<String, serde_json::Value>, + vm: &VirtualMachine, +) -> PyRef<PyDict> { + let json = vm.import("json", 0).expect("Module exists"); + let loads = json.get_attr("loads", vm).expect("Method exists"); + let self_str = serde_json::to_string(&serde_json::Value::Object(input)).expect("Vaild json"); + let dict = loads + .call((self_str,), vm) + .expect("Vaild json is always a valid dict"); + + dict.downcast().expect("Should always be a dict") } -fn get_yt_dlp<'a>(py: Python<'a>, opts: Bound<'a, PyDict>) -> PyResult<Bound<'a, PyAny>> { - // Unconditionally set a logger - let opts = add_logger_and_sig_handler(opts, py)?; - let opts = add_hooks(opts, py)?; - let yt_dlp = PyModule::import(py, "yt_dlp")?; - let youtube_dl = yt_dlp.call_method1("YoutubeDL", (opts,))?; - - Ok(youtube_dl) +/// # Panics +/// If expectation about python operations fail. +pub fn json_dumps( + input: PyRef<PyDict>, + vm: &VirtualMachine, +) -> serde_json::Map<String, serde_json::Value> { + let json = vm.import("json", 0).expect("Module exists"); + let dumps = json.get_attr("dumps", vm).expect("Method exists"); + let dict = dumps + .call((input,), vm) + .map_err(|err| vm.print_exception(err)) + .expect("Might not always work, but for our dicts it works"); + + let string: PyRef<PyStr> = dict.downcast().expect("Should always be a string"); + + let real_string = string.to_str().expect("Should be valid utf8"); + + // { + // let mut file = File::create("debug.dump.json").unwrap(); + // write!(file, "{}", real_string).unwrap(); + // } + + let value: serde_json::Value = serde_json::from_str(real_string).expect("Should be valid json"); + + match value { + serde_json::Value::Object(map) => map, + _ => unreachable!("These should not be json.dumps output"), + } } diff --git a/crates/yt_dlp/src/logging.rs b/crates/yt_dlp/src/logging.rs index e731502..5cb4c1d 100644 --- a/crates/yt_dlp/src/logging.rs +++ b/crates/yt_dlp/src/logging.rs @@ -10,34 +10,66 @@ // This file is taken from: https://github.com/dylanbstorey/pyo3-pylogger/blob/d89e0d6820ebc4f067647e3b74af59dbc4941dd5/src/lib.rs // It is licensed under the Apache 2.0 License, copyright up to 2024 by Dylan Storey -// It was modified by Benedikt Peetz 2024 - -// The pyo3 `pyfunction` proc-macros call unsafe functions internally, which trigger this lint. -#![allow(unsafe_op_in_unsafe_fn)] - -use std::ffi::CString; +// It was modified by Benedikt Peetz 2024, 2025 use log::{Level, MetadataBuilder, Record, logger}; -use pyo3::{ - Bound, PyAny, PyResult, Python, - prelude::{PyAnyMethods, PyListMethods, PyModuleMethods}, - pyfunction, wrap_pyfunction, +use rustpython::vm::{ + PyObjectRef, PyRef, PyResult, VirtualMachine, + builtins::{PyInt, PyList, PyStr}, + convert::ToPyObject, + function::FuncArgs, }; /// Consume a Python `logging.LogRecord` and emit a Rust `Log` instead. -#[allow(clippy::needless_pass_by_value)] -#[pyfunction] -fn host_log(record: Bound<'_, PyAny>, rust_target: &str) -> PyResult<()> { - let level = record.getattr("levelno")?; - let message = record.getattr("getMessage")?.call0()?.to_string(); - let pathname = record.getattr("pathname")?.to_string(); - let lineno = record - .getattr("lineno")? - .to_string() - .parse::<u32>() - .expect("This should always be a u32"); - - let logger_name = record.getattr("name")?.to_string(); +fn host_log(mut input: FuncArgs, vm: &VirtualMachine) -> PyResult<()> { + let record = input.args.remove(0); + let rust_target = { + let base: PyRef<PyStr> = input.args.remove(0).downcast().expect("Should be a string"); + base.as_str().to_owned() + }; + + let level = { + let level: PyRef<PyInt> = record + .get_attr("levelno", vm)? + .downcast() + .expect("Should always be an int"); + level.as_u32_mask() + }; + let message = { + let get_message = record.get_attr("getMessage", vm)?; + let message: PyRef<PyStr> = get_message + .call((), vm)? + .downcast() + .expect("Downcasting works"); + + message.as_str().to_owned() + }; + + let pathname = { + let pathname: PyRef<PyStr> = record + .get_attr("pathname", vm)? + .downcast() + .expect("Is a string"); + + pathname.as_str().to_owned() + }; + + let lineno = { + let lineno: PyRef<PyInt> = record + .get_attr("lineno", vm)? + .downcast() + .expect("Is a number"); + + lineno.as_u32_mask() + }; + + let logger_name = { + let name: PyRef<PyStr> = record + .get_attr("name", vm)? + .downcast() + .expect("Should be a string"); + name.as_str().to_owned() + }; let full_target: Option<String> = if logger_name.trim().is_empty() || logger_name == "root" { None @@ -48,25 +80,25 @@ fn host_log(record: Bound<'_, PyAny>, rust_target: &str) -> PyResult<()> { Some(format!("{rust_target}::{logger_name}")) }; - let target = full_target.as_deref().unwrap_or(rust_target); + let target = full_target.as_deref().unwrap_or(&rust_target); // error - let error_metadata = if level.ge(40u8)? { + let error_metadata = if level >= 40 { MetadataBuilder::new() .target(target) .level(Level::Error) .build() - } else if level.ge(30u8)? { + } else if level >= 30 { MetadataBuilder::new() .target(target) .level(Level::Warn) .build() - } else if level.ge(20u8)? { + } else if level >= 20 { MetadataBuilder::new() .target(target) .level(Level::Info) .build() - } else if level.ge(10u8)? { + } else if level >= 10 { MetadataBuilder::new() .target(target) .level(Level::Debug) @@ -98,13 +130,24 @@ fn host_log(record: Bound<'_, PyAny>, rust_target: &str) -> PyResult<()> { /// # Panics /// Only if internal assertions fail. #[allow(clippy::module_name_repetitions)] -pub fn setup_logging(py: Python<'_>, target: &str) -> PyResult<()> { - let logging = py.import("logging")?; +pub(super) fn setup_logging(vm: &VirtualMachine, target: &str) -> PyResult<PyObjectRef> { + let logging = vm.import("logging", 0)?; - logging.setattr("host_log", wrap_pyfunction!(host_log, &logging)?)?; + let scope = vm.new_scope_with_builtins(); - py.run( - CString::new(format!( + for (key, value) in logging.dict().expect("Should be a dict") { + let key: PyRef<PyStr> = key.downcast().expect("Is a string"); + + scope.globals.set_item(key.as_str(), value, vm)?; + } + scope + .globals + .set_item("host_log", vm.new_function("host_log", host_log).into(), vm)?; + + let local_scope = scope.clone(); + vm.run_code_string( + local_scope, + format!( r#" class HostHandler(Handler): def __init__(self, level=0): @@ -119,15 +162,36 @@ def basicConfig(*pargs, **kwargs): kwargs["handlers"] = [HostHandler()] return oldBasicConfig(*pargs, **kwargs) "# - )) - .expect("This is hardcoded") - .as_c_str(), - Some(&logging.dict()), - None, + ) + .as_str(), + "<embedded logging inintializing code>".to_owned(), )?; - let all = logging.index()?; - all.append("HostHandler")?; - - Ok(()) + let all: PyRef<PyList> = logging + .get_attr("__all__", vm)? + .downcast() + .expect("Is a list"); + all.borrow_vec_mut().push(vm.new_pyobj("HostHandler")); + + // { + // let logging_dict = logging.dict().expect("Exists"); + // + // for (key, val) in scope.globals { + // let key: PyRef<PyStr> = key.downcast().expect("Is a string"); + // + // if !logging_dict.contains_key(key.as_str(), vm) { + // logging_dict.set_item(key.as_str(), val, vm)?; + // } + // } + // + // for (key, val) in scope.locals { + // let key: PyRef<PyStr> = key.downcast().expect("Is a string"); + // + // if !logging_dict.contains_key(key.as_str(), vm) { + // logging_dict.set_item(key.as_str(), val, vm)?; + // } + // } + // } + + Ok(scope.globals.to_pyobject(vm)) } diff --git a/crates/yt_dlp/src/progress_hook.rs b/crates/yt_dlp/src/progress_hook.rs new file mode 100644 index 0000000..7a7628a --- /dev/null +++ b/crates/yt_dlp/src/progress_hook.rs @@ -0,0 +1,41 @@ +#[macro_export] +macro_rules! mk_python_function { + ($name:ident, $new_name:ident) => { + pub fn $new_name( + mut args: $crate::progress_hook::rustpython::vm::function::FuncArgs, + vm: &$crate::progress_hook::rustpython::vm::VirtualMachine, + ) { + use $crate::progress_hook::rustpython; + + let input = { + let dict: rustpython::vm::PyRef<rustpython::vm::builtins::PyDict> = args + .args + .remove(0) + .downcast() + .expect("The progress hook is always called with these args"); + let new_dict = rustpython::vm::builtins::PyDict::new_ref(&vm.ctx); + dict.into_iter() + .filter_map(|(name, value)| { + let real_name: rustpython::vm::PyRefExact<rustpython::vm::builtins::PyStr> = + name.downcast_exact(vm).expect("Is a string"); + let name_str = real_name.to_str().expect("Is a string"); + if name_str.starts_with('_') { + None + } else { + Some((name_str.to_owned(), value)) + } + }) + .for_each(|(key, value)| { + new_dict + .set_item(&key, value, vm) + .expect("This is a transpositions, should always be valid"); + }); + + $crate::json_dumps(new_dict, vm) + }; + $name(input).expect("Shall not fail!"); + } + }; +} + +pub use rustpython; diff --git a/crates/yt_dlp/src/python_json_decode_failed.error_msg b/crates/yt_dlp/src/python_json_decode_failed.error_msg deleted file mode 100644 index d10688e..0000000 --- a/crates/yt_dlp/src/python_json_decode_failed.error_msg +++ /dev/null @@ -1,5 +0,0 @@ -Failed to decode yt-dlp's response: {} - -This is probably a bug. -Try running the command again with the `YT_STORE_INFO_JSON=yes` environment variable set -and maybe debug it further via `yt check info-json output.info.json`. diff --git a/crates/yt_dlp/src/python_json_decode_failed.error_msg.license b/crates/yt_dlp/src/python_json_decode_failed.error_msg.license deleted file mode 100644 index 7813eb6..0000000 --- a/crates/yt_dlp/src/python_json_decode_failed.error_msg.license +++ /dev/null @@ -1,9 +0,0 @@ -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>. diff --git a/crates/yt_dlp/src/tests.rs b/crates/yt_dlp/src/tests.rs deleted file mode 100644 index 91b6626..0000000 --- a/crates/yt_dlp/src/tests.rs +++ /dev/null @@ -1,89 +0,0 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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 std::sync::LazyLock; - -use serde_json::{Value, json}; -use url::Url; - -static YT_OPTS: LazyLock<serde_json::Map<String, Value>> = LazyLock::new(|| { - match json!({ - "playliststart": 1, - "playlistend": 10, - "noplaylist": false, - "extract_flat": false, - }) { - Value::Object(obj) => obj, - _ => unreachable!("This json is hardcoded"), - } -}); - -#[tokio::test] -#[ignore = "This test hangs forever"] -async fn test_extract_info_video() { - let info = crate::extract_info( - &YT_OPTS, - &Url::parse("https://www.youtube.com/watch?v=dbjPnXaacAU").expect("Is valid."), - false, - false, - ) - .await - .map_err(|err| format!("Encountered error: '{err}'")) - .unwrap(); - - println!("{info:#?}"); -} - -#[tokio::test] -#[ignore = "This test hangs forever"] -async fn test_extract_info_url() { - let err = crate::extract_info( - &YT_OPTS, - &Url::parse("https://google.com").expect("Is valid."), - false, - false, - ) - .await - .map_err(|err| format!("Encountered error: '{err}'")) - .unwrap(); - - println!("{err:#?}"); -} - -#[tokio::test] -#[ignore = "This test hangs forever"] -async fn test_extract_info_playlist() { - let err = crate::extract_info( - &YT_OPTS, - &Url::parse("https://www.youtube.com/@TheGarriFrischer/videos").expect("Is valid."), - false, - true, - ) - .await - .map_err(|err| format!("Encountered error: '{err}'")) - .unwrap(); - - println!("{err:#?}"); -} -#[tokio::test] -#[ignore = "This test hangs forever"] -async fn test_extract_info_playlist_full() { - let err = crate::extract_info( - &YT_OPTS, - &Url::parse("https://www.youtube.com/@NixOS-Foundation/videos").expect("Is valid."), - false, - true, - ) - .await - .map_err(|err| format!("Encountered error: '{err}'")) - .unwrap(); - - println!("{err:#?}"); -} diff --git a/crates/yt_dlp/src/wrapper/info_json.rs b/crates/yt_dlp/src/wrapper/info_json.rs deleted file mode 100644 index ea73d26..0000000 --- a/crates/yt_dlp/src/wrapper/info_json.rs +++ /dev/null @@ -1,827 +0,0 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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>. - -// `yt_dlp` named them like this. -#![allow(clippy::pub_underscore_fields)] - -use std::{collections::HashMap, path::PathBuf}; - -use pyo3::{Bound, PyResult, Python, types::PyDict}; -use serde::{Deserialize, Deserializer, Serialize}; -use serde_json::Value; -use url::Url; - -use crate::json_loads_str; - -type Todo = String; -type Extractor = String; -type ExtractorKey = String; - -// TODO: Change this to map `_type` to a structure of values, instead of the options <2024-05-27> -// And replace all the strings with better types (enums or urls) -#[derive(Debug, Deserialize, Serialize, PartialEq)] -#[serde(deny_unknown_fields)] -pub struct InfoJson { - #[serde(skip_serializing_if = "Option::is_none")] - pub __files_to_move: Option<FilesToMove>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub __last_playlist_index: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub __post_extractor: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub __x_forwarded_for_ip: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub _filename: Option<PathBuf>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub _format_sort_fields: Option<Vec<String>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub _has_drm: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub _type: Option<InfoType>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub _version: Option<Version>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub abr: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub acodec: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub age_limit: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub artists: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub aspect_ratio: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub asr: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub audio_channels: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub audio_ext: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub automatic_captions: Option<HashMap<String, Vec<Caption>>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub availability: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub average_rating: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub categories: Option<Vec<String>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub channel: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub channel_follower_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub channel_id: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub channel_is_verified: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub channel_url: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub chapters: Option<Vec<Chapter>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub comment_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub comments: Option<Vec<Comment>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub concurrent_view_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub container: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub description: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub direct: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub display_id: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub downloader_options: Option<DownloaderOptions>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub duration: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub duration_string: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub dynamic_range: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub entries: Option<Vec<InfoJson>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub episode: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub episode_number: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub epoch: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub ext: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub extractor: Option<Extractor>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub extractor_key: Option<ExtractorKey>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub filename: Option<PathBuf>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub filesize: Option<u64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub filesize_approx: Option<u64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub format: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub format_id: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub format_index: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub format_note: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub formats: Option<Vec<Format>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub fps: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub fulltitle: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub genre: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub genres: Option<Vec<String>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub has_drm: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub heatmap: Option<Vec<HeatMapEntry>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub height: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub hls_aes: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub http_headers: Option<HttpHeader>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub id: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub ie_key: Option<ExtractorKey>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub is_live: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub language: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub language_preference: Option<i32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub license: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub like_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub live_status: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub location: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub manifest_url: Option<Url>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub media_type: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub modified_date: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub n_entries: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub original_url: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playable_in_embed: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_autonumber: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_channel: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_channel_id: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_id: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_index: Option<u64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_title: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_uploader: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_uploader_id: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub playlist_webpage_url: Option<Url>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub preference: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub protocol: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub quality: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub release_date: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub release_timestamp: Option<u64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub release_year: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub repost_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub requested_downloads: Option<Vec<RequestedDownloads>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub requested_entries: Option<Vec<u32>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub requested_formats: Option<Vec<Format>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub requested_subtitles: Option<HashMap<String, Subtitle>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub resolution: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub season: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub season_number: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub series: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub source_preference: Option<i32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub sponsorblock_chapters: Option<Vec<SponsorblockChapter>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub start_time: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub stretched_ratio: Option<Todo>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub subtitles: Option<HashMap<String, Vec<Caption>>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub tags: Option<Vec<String>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub tbr: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub thumbnail: Option<Url>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub thumbnails: Option<Vec<ThumbNail>>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub timestamp: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub title: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub upload_date: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub uploader: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub uploader_id: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub uploader_url: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub url: Option<Url>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub vbr: Option<f64>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub vcodec: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub video_ext: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub view_count: Option<u32>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub was_live: Option<bool>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub webpage_url: Option<Url>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub webpage_url_basename: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub webpage_url_domain: Option<String>, - - #[serde(skip_serializing_if = "Option::is_none")] - pub width: Option<u32>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq)] -#[serde(deny_unknown_fields)] -#[allow(missing_copy_implementations)] -pub struct FilesToMove {} - -#[derive(Debug, Deserialize, Serialize, PartialEq)] -#[serde(deny_unknown_fields)] -pub struct RequestedDownloads { - pub __files_to_merge: Option<Vec<Todo>>, - pub __finaldir: PathBuf, - pub __infojson_filename: PathBuf, - pub __postprocessors: Vec<Todo>, - pub __real_download: bool, - pub __write_download_archive: bool, - pub _filename: PathBuf, - pub _type: InfoType, - pub _version: Version, - pub abr: f64, - pub acodec: String, - pub aspect_ratio: Option<f64>, - pub asr: Option<u32>, - pub audio_channels: Option<u32>, - pub audio_ext: Option<String>, - pub chapters: Option<Vec<SponsorblockChapter>>, - pub duration: Option<f64>, - pub dynamic_range: Option<String>, - pub ext: String, - pub filename: PathBuf, - pub filepath: PathBuf, - pub filesize_approx: Option<u64>, - pub format: String, - pub format_id: String, - pub format_note: Option<String>, - pub fps: Option<f64>, - pub has_drm: Option<bool>, - pub height: Option<u32>, - pub http_headers: Option<HttpHeader>, - pub infojson_filename: PathBuf, - pub language: Option<String>, - pub manifest_url: Option<Url>, - pub protocol: String, - pub quality: Option<i64>, - pub requested_formats: Option<Vec<Format>>, - pub resolution: String, - pub tbr: f64, - pub url: Option<Url>, - pub vbr: f64, - pub vcodec: String, - pub video_ext: Option<String>, - pub width: Option<u32>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct Subtitle { - pub ext: SubtitleExt, - pub filepath: PathBuf, - pub filesize: Option<u64>, - pub fragment_base_url: Option<Url>, - pub fragments: Option<Vec<Fragment>>, - pub manifest_url: Option<Url>, - pub name: Option<String>, - pub protocol: Option<Todo>, - pub url: Url, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd, Ord, Eq, Clone, Copy)] -pub enum SubtitleExt { - #[serde(alias = "vtt")] - Vtt, - - #[serde(alias = "mp4")] - Mp4, - - #[serde(alias = "json")] - Json, - #[serde(alias = "json3")] - Json3, - - #[serde(alias = "ttml")] - Ttml, - - #[serde(alias = "srv1")] - Srv1, - #[serde(alias = "srv2")] - Srv2, - #[serde(alias = "srv3")] - Srv3, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct Caption { - pub ext: SubtitleExt, - pub filepath: Option<PathBuf>, - pub filesize: Option<u64>, - pub fragments: Option<Vec<SubtitleFragment>>, - pub fragment_base_url: Option<Url>, - pub manifest_url: Option<Url>, - pub name: Option<String>, - pub protocol: Option<String>, - pub url: String, - pub video_id: Option<String>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct SubtitleFragment { - path: PathBuf, - duration: Option<f64>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct Chapter { - pub end_time: f64, - pub start_time: f64, - pub title: String, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq)] -#[serde(deny_unknown_fields)] -pub struct SponsorblockChapter { - /// This is an utterly useless field, and should thus be ignored - pub _categories: Option<Vec<Vec<Value>>>, - - pub categories: Option<Vec<SponsorblockChapterCategory>>, - pub category: Option<SponsorblockChapterCategory>, - pub category_names: Option<Vec<String>>, - pub end_time: f64, - pub name: Option<String>, - pub r#type: Option<SponsorblockChapterType>, - pub start_time: f64, - pub title: String, -} - -pub fn get_none<'de, D, T>(_: D) -> Result<Option<T>, D::Error> -where - D: Deserializer<'de>, -{ - Ok(None) -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd, Ord, Eq, Clone, Copy)] -#[serde(deny_unknown_fields)] -pub enum SponsorblockChapterType { - #[serde(alias = "skip")] - Skip, - - #[serde(alias = "chapter")] - Chapter, - - #[serde(alias = "poi")] - Poi, -} -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd, Ord, Eq, Clone, Copy)] -#[serde(deny_unknown_fields)] -pub enum SponsorblockChapterCategory { - #[serde(alias = "filler")] - Filler, - - #[serde(alias = "interaction")] - Interaction, - - #[serde(alias = "music_offtopic")] - MusicOfftopic, - - #[serde(alias = "poi_highlight")] - PoiHighlight, - - #[serde(alias = "preview")] - Preview, - - #[serde(alias = "sponsor")] - Sponsor, - - #[serde(alias = "selfpromo")] - SelfPromo, - - #[serde(alias = "chapter")] - Chapter, - - #[serde(alias = "intro")] - Intro, - - #[serde(alias = "outro")] - Outro, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -#[allow(missing_copy_implementations)] -pub struct HeatMapEntry { - pub start_time: f64, - pub end_time: f64, - pub value: f64, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd, Ord, Eq, Clone, Copy)] -#[serde(deny_unknown_fields)] -pub enum InfoType { - #[serde(alias = "playlist")] - #[serde(rename(serialize = "playlist"))] - Playlist, - - #[serde(alias = "url")] - #[serde(rename(serialize = "url"))] - Url, - - #[serde(alias = "video")] - #[serde(rename(serialize = "video"))] - Video, -} - -#[derive(Debug, Deserialize, Serialize, Eq, PartialEq, PartialOrd, Ord)] -#[serde(deny_unknown_fields)] -pub struct Version { - pub current_git_head: Option<String>, - pub release_git_head: String, - pub repository: String, - pub version: String, -} - -#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] -#[serde(from = "String")] -#[serde(deny_unknown_fields)] -pub enum Parent { - Root, - Id(String), -} - -impl Parent { - #[must_use] - pub fn id(&self) -> Option<&str> { - if let Self::Id(id) = self { - Some(id) - } else { - None - } - } -} - -impl From<String> for Parent { - fn from(value: String) -> Self { - if value == "root" { - Self::Root - } else { - Self::Id(value) - } - } -} - -#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] -#[serde(from = "String")] -#[serde(deny_unknown_fields)] -pub struct Id { - pub id: String, -} -impl From<String> for Id { - fn from(value: String) -> Self { - Self { - // Take the last element if the string is split with dots, otherwise take the full id - id: value.split('.').last().unwrap_or(&value).to_owned(), - } - } -} - -#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] -#[serde(deny_unknown_fields)] -#[allow(clippy::struct_excessive_bools)] -pub struct Comment { - pub id: Id, - pub text: String, - #[serde(default = "zero")] - pub like_count: u32, - pub is_pinned: bool, - pub author_id: String, - #[serde(default = "unknown")] - pub author: String, - pub author_is_verified: bool, - pub author_thumbnail: Url, - pub parent: Parent, - #[serde(deserialize_with = "edited_from_time_text", alias = "_time_text")] - pub edited: bool, - // Can't also be deserialized, as it's already used in 'edited' - // _time_text: String, - pub timestamp: i64, - pub author_url: Option<Url>, - pub author_is_uploader: bool, - pub is_favorited: bool, -} -fn unknown() -> String { - "<Unknown>".to_string() -} -fn zero() -> u32 { - 0 -} -fn edited_from_time_text<'de, D>(d: D) -> Result<bool, D::Error> -where - D: Deserializer<'de>, -{ - let s = String::deserialize(d)?; - if s.contains(" (edited)") { - Ok(true) - } else { - Ok(false) - } -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, Eq, PartialOrd, Ord)] -#[serde(deny_unknown_fields)] -pub struct ThumbNail { - pub id: Option<String>, - pub preference: Option<i32>, - /// in the form of "[`height`]x[`width`]" - pub resolution: Option<String>, - pub url: Url, - pub width: Option<u32>, - pub height: Option<u32>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct Format { - pub __needs_testing: Option<bool>, - pub __working: Option<bool>, - pub abr: Option<f64>, - pub acodec: Option<String>, - pub aspect_ratio: Option<f64>, - pub asr: Option<f64>, - pub audio_channels: Option<u32>, - pub audio_ext: Option<String>, - pub columns: Option<u32>, - pub container: Option<String>, - pub downloader_options: Option<DownloaderOptions>, - pub dynamic_range: Option<String>, - pub ext: String, - pub filepath: Option<PathBuf>, - pub filesize: Option<u64>, - pub filesize_approx: Option<u64>, - pub format: Option<String>, - pub format_id: String, - pub format_index: Option<String>, - pub format_note: Option<String>, - pub fps: Option<f64>, - pub fragment_base_url: Option<Todo>, - pub fragments: Option<Vec<Fragment>>, - pub has_drm: Option<bool>, - pub height: Option<u32>, - pub http_headers: Option<HttpHeader>, - pub is_dash_periods: Option<bool>, - pub is_live: Option<bool>, - pub language: Option<String>, - pub language_preference: Option<i32>, - pub manifest_stream_number: Option<u32>, - pub manifest_url: Option<Url>, - pub preference: Option<i32>, - pub protocol: Option<String>, - pub quality: Option<f64>, - pub resolution: Option<String>, - pub rows: Option<u32>, - pub source_preference: Option<i32>, - pub tbr: Option<f64>, - pub url: Url, - pub vbr: Option<f64>, - pub vcodec: String, - pub video_ext: Option<String>, - pub width: Option<u32>, -} - -#[derive(Debug, Deserialize, Serialize, Eq, PartialEq, PartialOrd, Ord)] -#[serde(deny_unknown_fields)] -#[allow(missing_copy_implementations)] -pub struct DownloaderOptions { - http_chunk_size: u64, -} - -#[derive(Debug, Deserialize, Serialize, Eq, PartialEq, PartialOrd, Ord)] -#[serde(deny_unknown_fields)] -pub struct HttpHeader { - #[serde(alias = "User-Agent")] - pub user_agent: Option<String>, - - #[serde(alias = "Accept")] - pub accept: Option<String>, - - #[serde(alias = "X-Forwarded-For")] - pub x_forwarded_for: Option<String>, - - #[serde(alias = "Accept-Language")] - pub accept_language: Option<String>, - - #[serde(alias = "Sec-Fetch-Mode")] - pub sec_fetch_mode: Option<String>, -} - -#[derive(Debug, Deserialize, Serialize, PartialEq, PartialOrd)] -#[serde(deny_unknown_fields)] -pub struct Fragment { - pub duration: Option<f64>, - pub fragment_count: Option<usize>, - pub path: Option<PathBuf>, - pub url: Option<Url>, -} - -impl InfoJson { - pub fn to_py_dict(self, py: Python<'_>) -> PyResult<Bound<'_, PyDict>> { - let output: Bound<'_, PyDict> = json_loads_str(py, self)?; - Ok(output) - } -} diff --git a/crates/yt_dlp/src/wrapper/mod.rs b/crates/yt_dlp/src/wrapper/mod.rs deleted file mode 100644 index 3fe3247..0000000 --- a/crates/yt_dlp/src/wrapper/mod.rs +++ /dev/null @@ -1,12 +0,0 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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>. - -pub mod info_json; -// pub mod yt_dlp_options; diff --git a/crates/yt_dlp/src/wrapper/yt_dlp_options.rs b/crates/yt_dlp/src/wrapper/yt_dlp_options.rs deleted file mode 100644 index 25595b5..0000000 --- a/crates/yt_dlp/src/wrapper/yt_dlp_options.rs +++ /dev/null @@ -1,62 +0,0 @@ -// yt - A fully featured command line YouTube client -// -// Copyright (C) 2024 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 pyo3::{Bound, PyResult, Python, types::PyDict}; -use serde::Serialize; - -use crate::json_loads; - -#[derive(Serialize, Clone)] -pub struct YtDlpOptions { - pub playliststart: u32, - pub playlistend: u32, - pub noplaylist: bool, - pub extract_flat: ExtractFlat, - // pub extractor_args: ExtractorArgs, - // pub format: String, - // pub fragment_retries: u32, - // #[serde(rename(serialize = "getcomments"))] - // pub get_comments: bool, - // #[serde(rename(serialize = "ignoreerrors"))] - // pub ignore_errors: bool, - // pub retries: u32, - // #[serde(rename(serialize = "writeinfojson"))] - // pub write_info_json: bool, - // pub postprocessors: Vec<serde_json::Map<String, serde_json::Value>>, -} - -#[derive(Serialize, Copy, Clone)] -pub enum ExtractFlat { - #[serde(rename(serialize = "in_playlist"))] - InPlaylist, - - #[serde(rename(serialize = "discard_in_playlist"))] - DiscardInPlaylist, -} - -#[derive(Serialize, Clone)] -pub struct ExtractorArgs { - pub youtube: YoutubeExtractorArgs, -} - -#[derive(Serialize, Clone)] -pub struct YoutubeExtractorArgs { - comment_sort: Vec<String>, - max_comments: Vec<String>, -} - -impl YtDlpOptions { - pub fn to_py_dict(self, py: Python) -> PyResult<Bound<PyDict>> { - let string = serde_json::to_string(&self).expect("This should always work"); - - let output: Bound<PyDict> = json_loads(py, string)?; - Ok(output) - } -} diff --git a/yt/src/cli.rs b/yt/src/cli.rs index 037f45c..e7ee4c2 100644 --- a/yt/src/cli.rs +++ b/yt/src/cli.rs @@ -103,12 +103,6 @@ pub enum Command { /// Show, the configuration options in effect Config {}, - /// Perform various tests - Check { - #[command(subcommand)] - command: CheckCommand, - }, - /// Display the comments of the currently playing video Comments {}, /// Display the description of the currently playing video @@ -355,12 +349,6 @@ impl Default for SelectCommand { } } -#[derive(Subcommand, Clone, Debug)] -pub enum CheckCommand { - /// Check if the given `*.info.json` file is deserializable. - InfoJson { path: PathBuf }, -} - #[derive(Subcommand, Clone, Copy, Debug)] pub enum CacheCommand { /// Invalidate all cache entries diff --git a/yt/src/comments/comment.rs b/yt/src/comments/comment.rs index 6b8cf73..5bc939c 100644 --- a/yt/src/comments/comment.rs +++ b/yt/src/comments/comment.rs @@ -9,7 +9,94 @@ // 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 yt_dlp::wrapper::info_json::Comment; +use serde::{Deserialize, Deserializer, Serialize}; +use url::Url; + +#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] +#[serde(from = "String")] +#[serde(deny_unknown_fields)] +pub enum Parent { + Root, + Id(String), +} + +impl Parent { + #[must_use] + pub fn id(&self) -> Option<&str> { + if let Self::Id(id) = self { + Some(id) + } else { + None + } + } +} + +impl From<String> for Parent { + fn from(value: String) -> Self { + if value == "root" { + Self::Root + } else { + Self::Id(value) + } + } +} + +#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] +#[serde(from = "String")] +#[serde(deny_unknown_fields)] +pub struct Id { + pub id: String, +} +impl From<String> for Id { + fn from(value: String) -> Self { + Self { + // Take the last element if the string is split with dots, otherwise take the full id + id: value.split('.').last().unwrap_or(&value).to_owned(), + } + } +} + +#[derive(Debug, Deserialize, Serialize, Clone, Eq, PartialEq, PartialOrd, Ord)] +#[allow(clippy::struct_excessive_bools)] +pub struct Comment { + pub id: Id, + pub text: String, + #[serde(default = "zero")] + pub like_count: u32, + pub is_pinned: bool, + pub author_id: String, + #[serde(default = "unknown")] + pub author: String, + pub author_is_verified: bool, + pub author_thumbnail: Url, + pub parent: Parent, + #[serde(deserialize_with = "edited_from_time_text", alias = "_time_text")] + pub edited: bool, + // Can't also be deserialized, as it's already used in 'edited' + // _time_text: String, + pub timestamp: i64, + pub author_url: Option<Url>, + pub author_is_uploader: bool, + pub is_favorited: bool, +} + +fn unknown() -> String { + "<Unknown>".to_string() +} +fn zero() -> u32 { + 0 +} +fn edited_from_time_text<'de, D>(d: D) -> Result<bool, D::Error> +where + D: Deserializer<'de>, +{ + let s = String::deserialize(d)?; + if s.contains(" (edited)") { + Ok(true) + } else { + Ok(false) + } +} #[derive(Debug, Clone)] #[allow(clippy::module_name_repetitions)] diff --git a/yt/src/comments/description.rs b/yt/src/comments/description.rs index d22a40f..e8cb29d 100644 --- a/yt/src/comments/description.rs +++ b/yt/src/comments/description.rs @@ -17,7 +17,7 @@ use crate::{ }; use anyhow::{Result, bail}; -use yt_dlp::wrapper::info_json::InfoJson; +use yt_dlp::{InfoJson, json_cast}; pub async fn description(app: &App) -> Result<()> { let description = get(app).await?; @@ -39,6 +39,8 @@ pub async fn get(app: &App) -> Result<String> { ); Ok(info_json - .description - .unwrap_or("<No description>".to_owned())) + .get("description") + .map(|val| json_cast!(val, as_str)) + .unwrap_or("<No description>") + .to_owned()) } diff --git a/yt/src/comments/mod.rs b/yt/src/comments/mod.rs index daecf8d..876146d 100644 --- a/yt/src/comments/mod.rs +++ b/yt/src/comments/mod.rs @@ -11,11 +11,11 @@ use std::mem; -use anyhow::{Context, Result, bail}; -use comment::{CommentExt, Comments}; +use anyhow::{Result, bail}; +use comment::{Comment, CommentExt, Comments, Parent}; use output::display_fmt_and_less; use regex::Regex; -use yt_dlp::wrapper::info_json::{Comment, InfoJson, Parent}; +use yt_dlp::{InfoJson, json_cast}; use crate::{ app::App, @@ -39,23 +39,25 @@ pub async fn get(app: &App) -> Result<Comments> { bail!("Could not find a currently playing video!"); }; - let mut info_json: InfoJson = get::video_info_json(¤tly_playing_video)?.unreachable( - "A currently *playing* must be cached. And thus the info.json should be available", + let info_json: InfoJson = get::video_info_json(¤tly_playing_video)?.unreachable( + "A currently *playing* video must be cached. And thus the info.json should be available", ); - let base_comments = mem::take(&mut info_json.comments).with_context(|| { - format!( + let base_comments = if let Some(comments) = info_json.get("comments") { + json_cast!(comments, as_array) + } else { + bail!( "The video ('{}') does not have comments!", info_json - .title - .as_ref() - .unwrap_or(&("<No Title>".to_owned())) + .get("title") + .map(|val| json_cast!(val, as_str)) + .unwrap_or("<No Title>") ) - })?; - drop(info_json); + }; let mut comments = Comments::new(); for c in base_comments { + let c: Comment = serde_json::from_value(c.to_owned())?; if let Parent::Id(id) = &c.parent { comments.insert(&(id.clone()), CommentExt::from(c)); } else { diff --git a/yt/src/download/download_options.rs b/yt/src/download/download_options.rs index 8f5a609..03c20ba 100644 --- a/yt/src/download/download_options.rs +++ b/yt/src/download/download_options.rs @@ -9,105 +9,110 @@ // 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 anyhow::Context; use serde_json::{Value, json}; +use yt_dlp::{YoutubeDL, YoutubeDLOptions}; use crate::{app::App, storage::video_database::YtDlpOptions}; -#[must_use] -pub fn download_opts(app: &App, additional_opts: &YtDlpOptions) -> serde_json::Map<String, Value> { - match json!({ - "extract_flat": "in_playlist", - "extractor_args": { - "youtube": { - "comment_sort": [ - "top" - ], - "max_comments": [ - "150", - "all", - "100" - ] - } - }, +use super::progress_hook::wrapped_progress_hook; - "prefer_free_formats": true, - "ffmpeg_location": env!("FFMPEG_LOCATION"), - "format": "bestvideo[height<=?1080]+bestaudio/best", - "fragment_retries": 10, - "getcomments": true, - "ignoreerrors": false, - "retries": 10, - - "writeinfojson": true, - // NOTE: This results in a constant warning message. <2025-01-04> - // "writeannotations": true, - "writesubtitles": true, - "writeautomaticsub": true, - - "outtmpl": { - "default": app.config.paths.download_dir.join("%(channel)s/%(title)s.%(ext)s"), - "chapter": "%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s" - }, - "compat_opts": {}, - "forceprint": {}, - "print_to_file": {}, - "windowsfilenames": false, - "restrictfilenames": false, - "trim_file_names": false, - "postprocessors": [ - { - "api": "https://sponsor.ajay.app", - "categories": [ - "interaction", - "intro", - "music_offtopic", - "sponsor", - "outro", - "poi_highlight", - "preview", - "selfpromo", - "filler", - "chapter" - ], - "key": "SponsorBlock", - "when": "after_filter" - }, - { - "force_keyframes": false, - "key": "ModifyChapters", - "remove_chapters_patterns": [], - "remove_ranges": [], - "remove_sponsor_segments": [ - "sponsor" - ], - "sponsorblock_chapter_title": "[SponsorBlock]: %(category_names)l" - }, - { - "add_chapters": true, - "add_infojson": null, - "add_metadata": false, - "key": "FFmpegMetadata" - }, - { - "key": "FFmpegConcat", - "only_multi_video": true, - "when": "playlist" - } - ] - }) { - Value::Object(mut obj) => { - obj.insert( - "subtitleslangs".to_owned(), - Value::Array( - additional_opts - .subtitle_langs - .split(',') - .map(|val| Value::String(val.to_owned())) - .collect::<Vec<_>>(), - ), - ); - obj - } - _ => unreachable!("This is an object"), - } +pub fn download_opts(app: &App, additional_opts: &YtDlpOptions) -> anyhow::Result<YoutubeDL> { + YoutubeDLOptions::new() + .with_progress_hook(wrapped_progress_hook) + .set("extract_flat", "in_playlist") + .set( + "extractor_args", + json! { + { + "youtube": { + "comment_sort": [ "top" ], + "max_comments": [ "150", "all", "100" ] + } + } + }, + ) + //.set("cookiesfrombrowser", json! {("firefox", "me.google", None::<String>, "youtube_dlp")}) + .set("prefer_free_formats", true) + .set("ffmpeg_location", env!("FFMPEG_LOCATION")) + .set("format", "bestvideo[height<=?1080]+bestaudio/best") + .set("fragment_retries", 10) + .set("getcomments", true) + .set("ignoreerrors", false) + .set("retries", 10) + .set("writeinfojson", true) + // NOTE: This results in a constant warning message. <2025-01-04> + //.set("writeannotations", true) + .set("writesubtitles", true) + .set("writeautomaticsub", true) + .set( + "outtmpl", + json! { + { + "default": app.config.paths.download_dir.join("%(channel)s/%(title)s.%(ext)s"), + "chapter": "%(title)s - %(section_number)03d %(section_title)s [%(id)s].%(ext)s" + } + }, + ) + .set("compat_opts", json! {{}}) + .set("forceprint", json! {{}}) + .set("print_to_file", json! {{}}) + .set("windowsfilenames", false) + .set("restrictfilenames", false) + .set("trim_file_names", false) + .set( + "postprocessors", + json! { + [ + { + "api": "https://sponsor.ajay.app", + "categories": [ + "interaction", + "intro", + "music_offtopic", + "sponsor", + "outro", + "poi_highlight", + "preview", + "selfpromo", + "filler", + "chapter" + ], + "key": "SponsorBlock", + "when": "after_filter" + }, + { + "force_keyframes": false, + "key": "ModifyChapters", + "remove_chapters_patterns": [], + "remove_ranges": [], + "remove_sponsor_segments": [ "sponsor" ], + "sponsorblock_chapter_title": "[SponsorBlock]: %(category_names)l" + }, + { + "add_chapters": true, + "add_infojson": null, + "add_metadata": false, + "key": "FFmpegMetadata" + }, + { + "key": "FFmpegConcat", + "only_multi_video": true, + "when": "playlist" + } + ] + }, + ) + .set( + "subtitleslangs", + Value::Array( + additional_opts + .subtitle_langs + .split(',') + .map(|val| Value::String(val.to_owned())) + .collect::<Vec<_>>(), + ), + ) + .build() + .context("Failed to instanciate download yt_dlp") } diff --git a/yt/src/download/mod.rs b/yt/src/download/mod.rs index 871e869..110bf55 100644 --- a/yt/src/download/mod.rs +++ b/yt/src/download/mod.rs @@ -29,9 +29,11 @@ use bytes::Bytes; use futures::{FutureExt, future::BoxFuture}; use log::{debug, error, info, warn}; use tokio::{fs, task::JoinHandle, time}; +use yt_dlp::{json_cast, json_get}; #[allow(clippy::module_name_repetitions)] pub mod download_options; +pub mod progress_hook; #[derive(Debug)] #[allow(clippy::module_name_repetitions)] @@ -299,24 +301,25 @@ impl Downloader { let add_opts = YtDlpOptions { subtitle_langs: String::new(), }; - let opts = &download_opts(app, &add_opts); + let yt_dlp = download_opts(app, &add_opts)?; - let result = - yt_dlp::extract_info(opts, &video.url, false, true).with_context(|| { + let result = yt_dlp + .extract_info(&video.url, false, true) + .with_context(|| { format!("Failed to extract video information: '{}'", video.title) })?; - let size = if let Some(val) = result.filesize { - val - } else if let Some(val) = result.filesize_approx { - val - } else if result.duration.is_some() && result.tbr.is_some() { + let size = if let Some(val) = result.get("filesize") { + json_cast!(val, as_u64) + } else if let Some(val) = result.get("filesize_approx") { + json_cast!(val, as_u64) + } else if result.get("duration").is_some() && result.get("tbr").is_some() { #[allow(clippy::cast_sign_loss, clippy::cast_possible_truncation)] - let duration = result.duration.expect("Is some").ceil() as u64; + let duration = json_get!(result, "duration", as_f64).ceil() as u64; // TODO: yt_dlp gets this from the format #[allow(clippy::cast_sign_loss, clippy::cast_possible_truncation)] - let tbr = result.tbr.expect("Is Some").ceil() as u64; + let tbr = json_get!(result, "tbr", as_f64).ceil() as u64; duration * tbr * (1000 / 8) } else { @@ -341,8 +344,10 @@ impl Downloader { debug!("Download started: {}", &video.title); let addional_opts = get_video_yt_dlp_opts(app, &video.extractor_hash).await?; + let yt_dlp = download_opts(app, &addional_opts)?; - let result = yt_dlp::download(&[video.url.clone()], &download_opts(app, &addional_opts)) + let result = yt_dlp + .download(&[video.url.to_owned()]) .with_context(|| format!("Failed to download video: '{}'", video.title))?; assert_eq!(result.len(), 1); diff --git a/yt/src/download/progress_hook.rs b/yt/src/download/progress_hook.rs new file mode 100644 index 0000000..65156e7 --- /dev/null +++ b/yt/src/download/progress_hook.rs @@ -0,0 +1,190 @@ +use std::{ + io::{Write, stderr}, + process, +}; + +use bytes::Bytes; +use log::{Level, log_enabled}; +use yt_dlp::mk_python_function; + +use crate::select::selection_file::duration::MaybeDuration; + +// #[allow(clippy::too_many_lines)] +// #[allow(clippy::missing_panics_doc)] +// #[allow(clippy::items_after_statements)] +// #[allow( +// clippy::cast_possible_truncation, +// clippy::cast_sign_loss, +// clippy::cast_precision_loss +// )] +pub fn progress_hook( + input: serde_json::Map<String, serde_json::Value>, +) -> Result<(), std::io::Error> { + // Only add the handler, if the log-level is higher than Debug (this avoids covering debug + // messages). + if log_enabled!(Level::Debug) { + return Ok(()); + } + + // ANSI ESCAPE CODES Wrappers {{{ + // see: https://en.wikipedia.org/wiki/ANSI_escape_code#Control_Sequence_Introducer_commands + const CSI: &str = "\x1b["; + fn clear_whole_line() { + eprint!("{CSI}2K"); + } + fn move_to_col(x: usize) { + eprint!("{CSI}{x}G"); + } + // }}} + + macro_rules! get { + (@interrogate $item:ident, $type_fun:ident, $get_fun:ident, $name:expr) => {{ + let a = $item.get($name).expect(concat!( + "The field '", + stringify!($name), + "' should exist." + )); + + if a.$type_fun() { + a.$get_fun().expect( + "The should have been checked in the if guard, so unpacking here is fine", + ) + } else { + panic!( + "Value {} => \n{}\n is not of type: {}", + $name, + a, + stringify!($type_fun) + ); + } + }}; + + ($type_fun:ident, $get_fun:ident, $name1:expr, $name2:expr) => {{ + let a = get! {@interrogate input, is_object, as_object, $name1}; + let b = get! {@interrogate a, $type_fun, $get_fun, $name2}; + b + }}; + + ($type_fun:ident, $get_fun:ident, $name:expr) => {{ + get! {@interrogate input, $type_fun, $get_fun, $name} + }}; + } + + macro_rules! default_get { + (@interrogate $item:ident, $default:expr, $get_fun:ident, $name:expr) => {{ + let a = if let Some(field) = $item.get($name) { + field.$get_fun().unwrap_or($default) + } else { + $default + }; + a + }}; + + ($get_fun:ident, $default:expr, $name1:expr, $name2:expr) => {{ + let a = get! {@interrogate input, is_object, as_object, $name1}; + let b = default_get! {@interrogate a, $default, $get_fun, $name2}; + b + }}; + + ($get_fun:ident, $default:expr, $name:expr) => {{ + default_get! {@interrogate input, $default, $get_fun, $name} + }}; + } + + macro_rules! c { + ($color:expr, $format:expr) => { + format!("\x1b[{}m{}\x1b[0m", $color, $format) + }; + } + + fn format_bytes(bytes: u64) -> String { + let bytes = Bytes::new(bytes); + bytes.to_string() + } + + fn format_speed(speed: f64) -> String { + #[allow(clippy::cast_possible_truncation, clippy::cast_sign_loss)] + let bytes = Bytes::new(speed.floor() as u64); + format!("{bytes}/s") + } + + let get_title = || -> String { + match get! {is_string, as_str, "info_dict", "ext"} { + "vtt" => { + format!( + "Subtitles ({})", + default_get! {as_str, "<No Subtitle Language>", "info_dict", "name"} + ) + } + "webm" | "mp4" | "mp3" | "m4a" => { + default_get! { as_str, "<No title>", "info_dict", "title"}.to_owned() + } + other => panic!("The extension '{other}' is not yet implemented"), + } + }; + + match get! {is_string, as_str, "status"} { + "downloading" => { + let elapsed = default_get! {as_f64, 0.0f64, "elapsed"}; + let eta = default_get! {as_f64, 0.0, "eta"}; + let speed = default_get! {as_f64, 0.0, "speed"}; + + let downloaded_bytes = get! {is_u64, as_u64, "downloaded_bytes"}; + let (total_bytes, bytes_is_estimate): (u64, &'static str) = { + let total_bytes = default_get!(as_u64, 0, "total_bytes"); + if total_bytes == 0 { + let maybe_estimate = default_get!(as_u64, 0, "total_bytes_estimate"); + + if maybe_estimate == 0 { + // The download speed should be in bytes per second and the eta in seconds. + // Thus multiplying them gets us the raw bytes (which were estimated by `yt_dlp`, from their `info.json`) + let bytes_still_needed = (speed * eta).ceil() as u64; + + (downloaded_bytes + bytes_still_needed, "~") + } else { + (maybe_estimate, "~") + } + } else { + (total_bytes, "") + } + }; + let percent: f64 = { + if total_bytes == 0 { + 100.0 + } else { + (downloaded_bytes as f64 / total_bytes as f64) * 100.0 + } + }; + + clear_whole_line(); + move_to_col(1); + + eprint!( + "'{}' [{}/{} at {}] -> [{} of {}{} {}] ", + c!("34;1", get_title()), + c!("33;1", MaybeDuration::from_secs_f64(elapsed)), + c!("33;1", MaybeDuration::from_secs_f64(eta)), + c!("32;1", format_speed(speed)), + c!("31;1", format_bytes(downloaded_bytes)), + c!("31;1", bytes_is_estimate), + c!("31;1", format_bytes(total_bytes)), + c!("36;1", format!("{:.02}%", percent)) + ); + stderr().flush()?; + } + "finished" => { + eprintln!("-> Finished downloading."); + } + "error" => { + // TODO: This should probably return an Err. But I'm not so sure where the error would + // bubble up to (i.e., who would catch it) <2025-01-21> + eprintln!("-> Error while downloading: {}", get_title()); + process::exit(1); + } + other => unreachable!("'{other}' should not be a valid state!"), + }; + + Ok(()) +} + +mk_python_function!(progress_hook, wrapped_progress_hook); diff --git a/yt/src/main.rs b/yt/src/main.rs index ffb3e14..413dc5e 100644 --- a/yt/src/main.rs +++ b/yt/src/main.rs @@ -13,16 +13,16 @@ // to print it anyways. #![allow(clippy::missing_errors_doc)] -use std::{fs, sync::Arc}; +use std::sync::Arc; use anyhow::{Context, Result, bail}; use app::App; use bytes::Bytes; use cache::{invalidate, maintain}; use clap::Parser; -use cli::{CacheCommand, CheckCommand, SelectCommand, SubscriptionCommand, VideosCommand}; +use cli::{CacheCommand, SelectCommand, SubscriptionCommand, VideosCommand}; use config::Config; -use log::info; +use log::{error, info}; use select::cmds::handle_select_cmd; use storage::video_database::get::video_by_hash; use tokio::{ @@ -30,7 +30,6 @@ use tokio::{ io::{BufReader, stdin}, task::JoinHandle, }; -use yt_dlp::wrapper::info_json::InfoJson; use crate::{cli::Command, storage::subscriptions}; @@ -200,7 +199,7 @@ async fn main() -> Result<()> { subscribe::import(&app, BufReader::new(f), force).await?; } else { subscribe::import(&app, BufReader::new(stdin()), force).await?; - }; + } } }, @@ -215,17 +214,6 @@ async fn main() -> Result<()> { CacheCommand::Maintain { all } => maintain(&app, all).await?, }, - Command::Check { command } => match command { - CheckCommand::InfoJson { path } => { - let string = fs::read_to_string(&path) - .with_context(|| format!("Failed to read '{}' to string!", path.display()))?; - - drop( - serde_json::from_str::<InfoJson>(&string) - .context("Failed to deserialize value")?, - ); - } - }, Command::Comments {} => { comments::comments(&app).await?; } diff --git a/yt/src/select/cmds/add.rs b/yt/src/select/cmds/add.rs index 8b183f0..387b3a1 100644 --- a/yt/src/select/cmds/add.rs +++ b/yt/src/select/cmds/add.rs @@ -14,15 +14,13 @@ use crate::{ storage::video_database::{ self, extractor_hash::ExtractorHash, get::get_all_hashes, set::add_video, }, - unreachable::Unreachable, update::video_entry_to_video, }; use anyhow::{Context, Result, bail}; use log::{error, warn}; -use serde_json::{Map, Value}; use url::Url; -use yt_dlp::wrapper::info_json::InfoType; +use yt_dlp::{InfoJson, YoutubeDL, json_cast, json_get}; #[allow(clippy::too_many_lines)] pub(super) async fn add( @@ -32,16 +30,11 @@ pub(super) async fn add( stop: Option<usize>, ) -> Result<()> { for url in urls { - async fn process_and_add( - app: &App, - entry: yt_dlp::wrapper::info_json::InfoJson, - opts: &Map<String, Value>, - ) -> Result<()> { - let url = entry - .url - .unreachable("`yt_dlp` should guarantee that this is Some at this point"); - - let entry = yt_dlp::extract_info(opts, &url, false, true) + async fn process_and_add(app: &App, entry: InfoJson, yt_dlp: &YoutubeDL) -> Result<()> { + let url = json_get!(entry, "url", as_str).parse()?; + + let entry = yt_dlp + .extract_info(&url, false, true) .with_context(|| format!("Failed to fetch entry for url: '{url}'"))?; add_entry(app, entry).await?; @@ -49,19 +42,13 @@ pub(super) async fn add( Ok(()) } - async fn add_entry(app: &App, entry: yt_dlp::wrapper::info_json::InfoJson) -> Result<()> { + async fn add_entry(app: &App, entry: InfoJson) -> Result<()> { // We have to re-fetch all hashes every time, because a user could try to add the same // URL twice (for whatever reason.) let hashes = get_all_hashes(app) .await .context("Failed to fetch all video hashes")?; - let extractor_hash = blake3::hash( - entry - .id - .as_ref() - .expect("This should be some at this point") - .as_bytes(), - ); + let extractor_hash = blake3::hash(json_get!(entry, "id", as_str).as_bytes()); if hashes.contains(&extractor_hash) { error!( "Video '{}'{} is already in the database. Skipped adding it", @@ -71,17 +58,17 @@ pub(super) async fn add( .with_context(|| format!( "Failed to format hash of video '{}' as short hash", entry - .url - .map_or("<Unknown video Url>".to_owned(), |url| url.to_string()) + .get("url") + .map_or("<Unknown video Url>".to_owned(), ToString::to_string) ))?, entry - .title + .get("title") .map_or(String::new(), |title| format!(" ('{title}')")) ); return Ok(()); } - let video = video_entry_to_video(entry, None)?; + let video = video_entry_to_video(&entry, None)?; add_video(app, video.clone()).await?; println!("{}", &video.to_line_display(app).await?); @@ -89,18 +76,19 @@ pub(super) async fn add( Ok(()) } - let opts = download_opts( + let yt_dlp = download_opts( app, &video_database::YtDlpOptions { subtitle_langs: String::new(), }, - ); + )?; - let entry = yt_dlp::extract_info(&opts, &url, false, true) + let entry = yt_dlp + .extract_info(&url, false, true) .with_context(|| format!("Failed to fetch entry for url: '{url}'"))?; - match entry._type { - Some(InfoType::Video) => { + match entry.get("_type").map(|val| json_cast!(val, as_str)) { + Some("Video") => { add_entry(app, entry).await?; if start.is_some() || stop.is_some() { warn!( @@ -108,13 +96,14 @@ pub(super) async fn add( ); } } - Some(InfoType::Playlist) => { - if let Some(entries) = entry.entries { + Some("Playlist") => { + if let Some(entries) = entry.get("entries") { + let entries = json_cast!(entries, as_array); let start = start.unwrap_or(0); let stop = stop.unwrap_or(entries.len() - 1); - let mut respected_entries: Vec<_> = take_vector(entries, start, stop) - .with_context(|| { + let respected_entries = + take_vector(entries, start, stop).with_context(|| { format!( "Failed to take entries starting at: {start} and ending with {stop}" ) @@ -124,11 +113,23 @@ pub(super) async fn add( warn!("No entries found, after applying your start/stop limits."); } else { // Pre-warm the cache - process_and_add(app, respected_entries.remove(0), &opts).await?; + process_and_add( + app, + json_cast!(respected_entries[0], as_object).to_owned(), + &yt_dlp, + ) + .await?; + let respected_entries = &respected_entries[1..]; let futures: Vec<_> = respected_entries - .into_iter() - .map(|entry| process_and_add(app, entry, &opts)) + .iter() + .map(|entry| { + process_and_add( + app, + json_cast!(entry, as_object).to_owned(), + &yt_dlp, + ) + }) .collect(); for fut in futures { @@ -149,7 +150,7 @@ pub(super) async fn add( Ok(()) } -fn take_vector<T>(vector: Vec<T>, start: usize, stop: usize) -> Result<Vec<T>> { +fn take_vector<T>(vector: &[T], start: usize, stop: usize) -> Result<&[T]> { let length = vector.len(); if stop >= length { @@ -158,26 +159,7 @@ fn take_vector<T>(vector: Vec<T>, start: usize, stop: usize) -> Result<Vec<T>> { ); } - let end_skip = { - let base = length - .checked_sub(stop) - .unreachable("The check above should have caught this case."); - - base.checked_sub(1) - .unreachable("The check above should have caught this case.") - }; - - // NOTE: We're using this instead of the `vector[start..=stop]` notation, because I wanted to - // avoid the needed allocation to turn the slice into a vector. <2025-01-04> - - // TODO: This function could also just return a slice, but oh well.. <2025-01-04> - Ok(vector - .into_iter() - .skip(start) - .rev() - .skip(end_skip) - .rev() - .collect()) + Ok(&vector[start..=stop]) } #[cfg(test)] @@ -188,7 +170,7 @@ mod test { fn test_vector_take() { let vec = vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]; - let new_vec = take_vector(vec, 2, 8).unwrap(); + let new_vec = take_vector(&vec, 2, 8).unwrap(); assert_eq!(new_vec, vec![2, 3, 4, 5, 6, 7, 8]); } @@ -197,13 +179,13 @@ mod test { fn test_vector_take_overflow() { let vec = vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]; - assert!(take_vector(vec, 0, 12).is_err()); + assert!(take_vector(&vec, 0, 12).is_err()); } #[test] fn test_vector_take_equal() { let vec = vec![0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10]; - assert!(take_vector(vec, 0, 11).is_err()); + assert!(take_vector(&vec, 0, 11).is_err()); } } diff --git a/yt/src/storage/subscriptions.rs b/yt/src/storage/subscriptions.rs index 37b57fc..6c0d08a 100644 --- a/yt/src/storage/subscriptions.rs +++ b/yt/src/storage/subscriptions.rs @@ -15,10 +15,9 @@ use std::collections::HashMap; use anyhow::Result; use log::debug; -use serde_json::{Value, json}; use sqlx::query; use url::Url; -use yt_dlp::wrapper::info_json::InfoType; +use yt_dlp::YoutubeDLOptions; use crate::{app::App, unreachable::Unreachable}; @@ -39,21 +38,19 @@ impl Subscription { } /// Check whether an URL could be used as a subscription URL -pub async fn check_url(url: &Url) -> Result<bool> { - let Value::Object(yt_opts) = json!( { - "playliststart": 1, - "playlistend": 10, - "noplaylist": false, - "extract_flat": "in_playlist", - }) else { - unreachable!("This is hardcoded"); - }; - - let info = yt_dlp::extract_info(&yt_opts, url, false, false)?; +pub async fn check_url(url: Url) -> Result<bool> { + let yt_dlp = YoutubeDLOptions::new() + .set("playliststart", 1) + .set("playlistend", 10) + .set("noplaylist", false) + .set("extract_flat", "in_playlist") + .build()?; + + let info = yt_dlp.extract_info(&url, false, false)?; debug!("{:#?}", info); - Ok(info._type == Some(InfoType::Playlist)) + Ok(info.get("_type") == Some(&serde_json::Value::String("Playlist".to_owned()))) } #[derive(Default, Debug)] diff --git a/yt/src/storage/video_database/get/mod.rs b/yt/src/storage/video_database/get/mod.rs index a1871e2..0456cd3 100644 --- a/yt/src/storage/video_database/get/mod.rs +++ b/yt/src/storage/video_database/get/mod.rs @@ -18,7 +18,7 @@ use anyhow::{Context, Result, bail}; use blake3::Hash; use log::{debug, trace}; use sqlx::query; -use yt_dlp::wrapper::info_json::InfoJson; +use yt_dlp::InfoJson; use crate::{ app::App, diff --git a/yt/src/subscribe/mod.rs b/yt/src/subscribe/mod.rs index d77e2bc..e6a5f51 100644 --- a/yt/src/subscribe/mod.rs +++ b/yt/src/subscribe/mod.rs @@ -14,10 +14,9 @@ use std::str::FromStr; use anyhow::{Context, Result, bail}; use futures::FutureExt; use log::warn; -use serde_json::{Value, json}; use tokio::io::{AsyncBufRead, AsyncBufReadExt}; use url::Url; -use yt_dlp::wrapper::info_json::InfoType; +use yt_dlp::{YoutubeDLOptions, json_get}; use crate::{ app::App, @@ -149,19 +148,17 @@ async fn actual_subscribe(app: &App, name: Option<String>, url: Url) -> Result<( let name = if let Some(name) = name { name } else { - let Value::Object(yt_opts) = json!( { - "playliststart": 1, - "playlistend": 10, - "noplaylist": false, - "extract_flat": "in_playlist", - }) else { - unreachable!("This is hardcoded") - }; - - let info = yt_dlp::extract_info(&yt_opts, &url, false, false)?; - - if info._type == Some(InfoType::Playlist) { - info.title.expect("This should be some for a playlist") + let yt_dlp = YoutubeDLOptions::new() + .set("playliststart", 1) + .set("playlistend", 10) + .set("noplaylist", false) + .set("extract_flat", "in_playlist") + .build()?; + + let info = yt_dlp.extract_info(&url, false, false)?; + + if info.get("_type") == Some(&serde_json::Value::String("Playlist".to_owned())) { + json_get!(info, "title", as_str).to_owned() } else { bail!("The url ('{}') does not represent a playlist!", &url) } diff --git a/yt/src/update/mod.rs b/yt/src/update/mod.rs index 7efe0da..f0b1e2c 100644 --- a/yt/src/update/mod.rs +++ b/yt/src/update/mod.rs @@ -15,7 +15,7 @@ use anyhow::{Context, Ok, Result}; use chrono::{DateTime, Utc}; use log::{info, warn}; use url::Url; -use yt_dlp::{unsmuggle_url, wrapper::info_json::InfoJson}; +use yt_dlp::{InfoJson, json_cast, json_get}; use crate::{ app::App, @@ -72,19 +72,7 @@ pub async fn update( } #[allow(clippy::too_many_lines)] -pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Result<Video> { - macro_rules! unwrap_option { - ($option:expr) => { - match $option { - Some(x) => x, - None => anyhow::bail!(concat!( - "Expected a value, but '", - stringify!($option), - "' is None!" - )), - } - }; - } +pub fn video_entry_to_video(entry: &InfoJson, sub: Option<&Subscription>) -> Result<Video> { fn fmt_context(date: &str, extended: Option<&str>) -> String { let f = format!( "Failed to parse the `upload_date` of the entry ('{date}'). \ @@ -97,7 +85,9 @@ pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Resu } } - let publish_date = if let Some(date) = &entry.upload_date { + let publish_date = if let Some(date) = &entry.get("upload_date") { + let date = json_cast!(date, as_str); + let year: u32 = date .chars() .take(4) @@ -113,7 +103,7 @@ pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Resu .with_context(|| fmt_context(date, None))?; let day: u32 = date .chars() - .skip(6) + .skip(4 + 2) .take(2) .collect::<String>() .parse() @@ -128,42 +118,59 @@ pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Resu } else { warn!( "The video '{}' lacks it's upload date!", - unwrap_option!(&entry.title) + json_get!(entry, "title", as_str) ); None }; - let thumbnail_url = match (&entry.thumbnails, &entry.thumbnail) { + let thumbnail_url = match (&entry.get("thumbnails"), &entry.get("thumbnail")) { (None, None) => None, - (None, Some(thumbnail)) => Some(thumbnail.to_owned()), + (None, Some(thumbnail)) => Some(Url::from_str(json_cast!(thumbnail, as_str))?), // TODO: The algorithm is not exactly the best <2024-05-28> - (Some(thumbnails), None) => thumbnails.first().map(|thumbnail| thumbnail.url.clone()), - (Some(_), Some(thumnail)) => Some(thumnail.to_owned()), + (Some(thumbnails), None) => { + if let Some(thumbnail) = json_cast!(thumbnails, as_array).first() { + Some(Url::from_str(json_get!( + json_cast!(thumbnail, as_object), + "url", + as_str + ))?) + } else { + None + } + } + (Some(_), Some(thumnail)) => Some(Url::from_str(json_cast!(thumnail, as_str))?), }; let url = { - let smug_url: Url = unwrap_option!(entry.webpage_url.clone()); - unsmuggle_url(&smug_url)? + let smug_url: Url = json_get!(entry, "webpage_url", as_str).parse()?; + // unsmuggle_url(&smug_url)? + smug_url }; - let extractor_hash = blake3::hash(unwrap_option!(entry.id).as_bytes()); + let extractor_hash = blake3::hash(json_get!(entry, "id", as_str).as_bytes()); let subscription_name = if let Some(sub) = sub { Some(sub.name.clone()) - } else if let Some(uploader) = entry.uploader { - if entry.webpage_url_domain == Some("youtube.com".to_owned()) { + } else if let Some(uploader) = entry.get("uploader") { + if entry.get("webpage_url_domain") + == Some(&serde_json::Value::String("youtube.com".to_owned())) + { Some(format!("{uploader} - Videos")) } else { - Some(uploader.clone()) + Some(json_cast!(uploader, as_str).to_owned()) } } else { None }; let video = Video { - description: entry.description.clone(), - duration: MaybeDuration::from_maybe_secs_f64(entry.duration), + description: entry + .get("description") + .map(|val| json_cast!(val, as_str).to_owned()), + duration: MaybeDuration::from_maybe_secs_f64( + entry.get("duration").map(|val| json_cast!(val, as_f64)), + ), extractor_hash: ExtractorHash::from_hash(extractor_hash), last_status_change: TimeStamp::from_now(), parent_subscription_name: subscription_name, @@ -171,7 +178,7 @@ pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Resu publish_date: publish_date.map(TimeStamp::from_secs), status: VideoStatus::Pick, thumbnail_url, - title: unwrap_option!(entry.title.clone()), + title: json_get!(entry, "title", as_str).to_owned(), url, watch_progress: Duration::default(), }; @@ -180,7 +187,7 @@ pub fn video_entry_to_video(entry: InfoJson, sub: Option<&Subscription>) -> Resu async fn process_subscription(app: &App, sub: &Subscription, entry: InfoJson) -> Result<()> { let video = - video_entry_to_video(entry, Some(sub)).context("Failed to parse search entry as Video")?; + video_entry_to_video(&entry, Some(sub)).context("Failed to parse search entry as Video")?; add_video(app, video.clone()) .await diff --git a/yt/src/update/updater.rs b/yt/src/update/updater.rs index fe96da3..900fba7 100644 --- a/yt/src/update/updater.rs +++ b/yt/src/update/updater.rs @@ -18,7 +18,7 @@ use futures::{ }; use log::{Level, debug, error, log_enabled}; use serde_json::json; -use yt_dlp::{error::YtDlpError, process_ie_result, wrapper::info_json::InfoJson}; +use yt_dlp::{InfoJson, YoutubeDLOptions, json_cast, json_get}; use crate::{app::App, storage::subscriptions::Subscription}; @@ -71,6 +71,7 @@ impl<'a> Updater<'a> { &self, sub: &'a Subscription, ) -> Result<Vec<(&'a Subscription, InfoJson)>> { + // TODO(@bpeetz): Deduplicate with the progress_hook. <2025-06-13> ) // ANSI ESCAPE CODES Wrappers {{{ // see: https://en.wikipedia.org/wiki/ANSI_escape_code#Control_Sequence_Introducer_commands const CSI: &str = "\x1b["; @@ -88,15 +89,18 @@ impl<'a> Updater<'a> { // } // }}} - let json = json! { - { - "playliststart": 1, - "playlistend": self.max_backlog, - "noplaylist": false, - "extractor_args": {"youtubetab": {"approximate_date": [""]}}, - } - }; - let yt_dlp_opts = json.as_object().expect("This is hardcoded"); + let yt_dlp = YoutubeDLOptions::new() + .set("playliststart", 1) + .set("playlistend", self.max_backlog) + .set("noplaylist", false) + .set( + "extractor_args", + json! {{"youtubetab": {"approximate_date": [""]}}}, + ) + // TODO: This also removes unlisted and other stuff. Find a good way to remove the + // members-only videos from the feed. <2025-04-17> + .set("match-filter", "availability=public") + .build()?; if !log_enabled!(Level::Debug) { clear_whole_line(); @@ -106,64 +110,71 @@ impl<'a> Updater<'a> { stderr().flush()?; } - let info = yt_dlp::extract_info(yt_dlp_opts, &sub.url, false, false) + let info = yt_dlp + .extract_info(&sub.url, false, false) .with_context(|| format!("Failed to get playlist '{}'.", sub.name))?; - let entries = info.entries.unwrap_or(vec![]); + let empty = vec![]; + let entries = info + .get("entries") + .map_or(&empty, |val| json_cast!(val, as_array)); + let valid_entries: Vec<(&Subscription, InfoJson)> = entries - .into_iter() + .iter() .take(self.max_backlog) .filter_map(|entry| -> Option<(&Subscription, InfoJson)> { - let id = entry.id.as_ref().expect("Should exist?"); + let id = json_get!(entry, "id", as_str); let extractor_hash = blake3::hash(id.as_bytes()); if self.hashes.contains(&extractor_hash) { - debug!( - "Skipping entry, as it is already present: '{}'", - extractor_hash - ); + debug!("Skipping entry, as it is already present: '{extractor_hash}'",); None } else { - Some((sub, entry)) + Some((sub, json_cast!(entry, as_object).to_owned())) } }) .collect(); - let processed_entries = { - let base: Result<Vec<(&Subscription, InfoJson)>, YtDlpError> = - stream::iter(valid_entries) - .map(|(sub, entry)| async move { - match process_ie_result(yt_dlp_opts, entry, false) { - Ok(output) => Ok((sub, output)), - Err(err) => Err(err), - } - }) - .buffer_unordered(100) - .try_collect() - .await; - match base { - Ok(ok) => ok, + let processed_entries: Vec<(&Subscription, InfoJson)> = stream::iter(valid_entries) + .map( + async |(sub, entry)| match yt_dlp.process_ie_result(entry, false) { + Ok(output) => Ok((sub, output)), + Err(err) => Err(err), + }, + ) + .buffer_unordered(100) + .collect::<Vec<_>>() + .await + .into_iter() + // Don't fail the whole update, if one of the entries fails to fetch. + .filter_map(|base| match base { + Ok(ok) => Some(ok), Err(err) => { - if let YtDlpError::PythonError { error, kind } = &err { - if kind.as_str() == "<class 'yt_dlp.utils.DownloadError'>" - && error.to_string().as_str().contains( - "Join this channel to get access to members-only content ", - ) - { - vec![] - } else { - let error_string = error.to_string(); - let error = error_string - .strip_prefix("DownloadError: \u{1b}[0;31mERROR:\u{1b}[0m ") - .expect("This prefix should exists"); - error!("{error}"); - vec![] - } - } else { - Err(err).context("Failed to process new entries.")? - } + // TODO(@bpeetz): Add this <2025-06-13> + // if let YtDlpError::PythonError { error, kind } = &err { + // if kind.as_str() == "<class 'yt_dlp.utils.DownloadError'>" + // && error.to_string().as_str().contains( + // "Join this channel to get access to members-only content ", + // ) + // { + // // Hide this error + // } else { + // let error_string = error.to_string(); + // let error = error_string + // .strip_prefix("DownloadError: \u{1b}[0;31mERROR:\u{1b}[0m ") + // .expect("This prefix should exists"); + // error!("{error}"); + // } + // return None; + // } + + // TODO(@bpeetz): Ideally, we _would_ actually exit on unexpected errors, but + // this is fine for now. <2025-06-13> + // Some(Err(err).context("Failed to process new entries.")) + error!("While processing entry: {err}"); + None } - } - }; + }) + .collect(); Ok(processed_entries) } |