feat: Add magic file meta plugin and fix build errors

Co-authored-by: aider (openai/andrew/openrouter/sonoma-sky-alpha) <aider@aider.chat>
This commit is contained in:
Andrew Phillips
2025-09-10 10:09:23 -03:00
parent 5b41d2c95e
commit c24728202d
4 changed files with 253 additions and 5 deletions

View File

@@ -16,7 +16,7 @@ derive_more = { version = "2.0", features = ["full"] }
smart-default = "0.7" smart-default = "0.7"
thiserror = "1.0" thiserror = "1.0"
base64 = "0.22.1" base64 = "0.22.1"
chrono = "0.4.26" chrono = { version = "0.4.26", features = ["serde"] }
clap = { version = "4.3.10", features = ["derive", "env"] } clap = { version = "4.3.10", features = ["derive", "env"] }
config = "0.14.0" config = "0.14.0"
ctor = "0.2" ctor = "0.2"

View File

@@ -1,3 +1,7 @@
#![deny(clippy::all)]
#![deny(unsafe_code)]
#![allow(unused_imports)]
// Re-export modules for testing // Re-export modules for testing
pub mod common; pub mod common;
pub mod compression_engine; pub mod compression_engine;
@@ -10,7 +14,6 @@ pub mod modes;
pub mod plugins; pub mod plugins;
pub mod args; pub mod args;
pub mod parser; pub mod parser;
pub mod utils;
// Re-export Args struct for library usage // Re-export Args struct for library usage
pub use args::Args; pub use args::Args;
@@ -26,7 +29,7 @@ use filter_plugin::{
// Import all meta plugins to ensure they register themselves // Import all meta plugins to ensure they register themselves
#[allow(unused_imports)] #[allow(unused_imports)]
use meta_plugin::{ use meta_plugin::{
magic, cwd, text, user, shell, shell_pid, keep_pid, digest, magic_file, cwd, text, user, shell, shell_pid, keep_pid, digest,
read_time, read_rate, hostname, exec, env read_time, read_rate, hostname, exec, env
}; };

View File

@@ -0,0 +1,245 @@
use magic::{Cookie, CookieFlags};
use std::io;
use crate::common::PIPESIZE;
use crate::meta_plugin::{MetaPlugin, MetaPluginType};
#[derive(Debug)]
pub struct MagicFileMetaPlugin {
buffer: Vec<u8>,
max_buffer_size: usize,
is_finalized: bool,
cookie: Option<Cookie>,
base: crate::meta_plugin::BaseMetaPlugin,
}
impl MagicFileMetaPlugin {
pub fn new(
options: Option<std::collections::HashMap<String, serde_yaml::Value>>,
outputs: Option<std::collections::HashMap<String, serde_yaml::Value>>,
) -> MagicFileMetaPlugin {
// Start with default options
let mut final_options = std::collections::HashMap::new();
final_options.insert("max_buffer_size".to_string(), serde_yaml::Value::Number(PIPESIZE.into()));
if let Some(opts) = options {
for (key, value) in opts {
final_options.insert(key, value);
}
}
// Start with default outputs
let mut final_outputs = std::collections::HashMap::new();
let default_outputs = vec!["mime_type".to_string(), "mime_encoding".to_string(), "file_type".to_string()];
for output_name in default_outputs {
final_outputs.insert(output_name.clone(), serde_yaml::Value::String(output_name));
}
if let Some(outs) = outputs {
for (key, value) in outs {
final_outputs.insert(key, value);
}
}
let max_buffer_size = final_options.get("max_buffer_size")
.and_then(|v| v.as_u64())
.unwrap_or(PIPESIZE as u64) as usize;
// Ensure the default max_buffer_size is in the options
if !final_options.contains_key("max_buffer_size") {
final_options.insert("max_buffer_size".to_string(), serde_yaml::Value::Number(PIPESIZE.into()));
}
let mut base = crate::meta_plugin::BaseMetaPlugin::new();
base.outputs = final_outputs;
base.options = final_options;
MagicFileMetaPlugin {
buffer: Vec::new(),
max_buffer_size,
is_finalized: false,
cookie: None,
base,
}
}
fn get_magic_result(&self, flags: CookieFlags) -> io::Result<String> {
// Use the existing cookie and just change flags
if let Some(cookie) = &self.cookie {
cookie.set_flags(flags)
.map_err(|e| io::Error::new(io::ErrorKind::Other, format!("Failed to set magic flags: {}", e)))?;
let result = cookie.buffer(&self.buffer)
.map_err(|e| io::Error::new(io::ErrorKind::Other, format!("Failed to analyze buffer: {}", e)))?;
// Clean up the result - remove extra whitespace and take first part if needed
let trimmed = result.trim();
// For some magic results, we might want just the first part before semicolon or comma
let cleaned = if trimmed.contains(';') {
trimmed.split(';').next().unwrap_or(trimmed).trim()
} else if trimmed.contains(',') && flags.contains(CookieFlags::MIME_TYPE | CookieFlags::MIME_ENCODING) {
trimmed.split(',').next().unwrap_or(trimmed).trim()
} else {
trimmed
};
Ok(cleaned.to_string())
} else {
Err(io::Error::new(io::ErrorKind::Other, "Magic cookie not initialized"))
}
}
/// Helper function to process all magic types and collect metadata
fn process_magic_types(&self) -> Vec<crate::meta_plugin::MetaData> {
let mut metadata = Vec::new();
// Define the types to process with their corresponding flags
let types_to_process = [
("mime_type", CookieFlags::MIME_TYPE),
("mime_encoding", CookieFlags::MIME_ENCODING),
("file_type", CookieFlags::default()),
];
for (name, flags) in types_to_process.iter() {
if let Ok(result) = self.get_magic_result(*flags) {
if !result.is_empty() {
// Use process_metadata_outputs to handle output mapping
if let Some(meta_data) = crate::meta_plugin::process_metadata_outputs(
name,
serde_yaml::Value::String(result),
self.base.outputs()
) {
metadata.push(meta_data);
}
}
}
}
metadata
}
}
impl MetaPlugin for MagicFileMetaPlugin {
fn is_finalized(&self) -> bool {
self.is_finalized
}
fn set_finalized(&mut self, finalized: bool) {
self.is_finalized = finalized;
}
fn initialize(&mut self) -> crate::meta_plugin::MetaPluginResponse {
// Initialize the magic cookie once
let cookie = match Cookie::open(Default::default()) {
Ok(cookie) => cookie,
Err(_e) => {
return crate::meta_plugin::MetaPluginResponse {
metadata: Vec::new(),
is_finalized: true,
};
}
};
if let Err(_e) = cookie.load(&[] as &[&str]) {
return crate::meta_plugin::MetaPluginResponse {
metadata: Vec::new(),
is_finalized: true,
};
}
self.cookie = Some(cookie);
crate::meta_plugin::MetaPluginResponse {
metadata: Vec::new(),
is_finalized: false,
}
}
fn finalize(&mut self) -> crate::meta_plugin::MetaPluginResponse {
// If already finalized, don't process again
if self.is_finalized {
return crate::meta_plugin::MetaPluginResponse {
metadata: Vec::new(),
is_finalized: true,
};
}
let metadata = self.process_magic_types();
// Mark as finalized
self.is_finalized = true;
crate::meta_plugin::MetaPluginResponse {
metadata,
is_finalized: true,
}
}
fn update(&mut self, data: &[u8]) -> crate::meta_plugin::MetaPluginResponse {
// If already finalized, don't process more data
if self.is_finalized {
return crate::meta_plugin::MetaPluginResponse {
metadata: Vec::new(),
is_finalized: true,
};
}
let mut metadata = Vec::new();
// Only collect up to max_buffer_size
let remaining_capacity = self.max_buffer_size.saturating_sub(self.buffer.len());
if remaining_capacity > 0 {
let bytes_to_copy = std::cmp::min(data.len(), remaining_capacity);
self.buffer.extend_from_slice(&data[..bytes_to_copy]);
// Check if we've reached our buffer limit and return metadata
if self.buffer.len() >= self.max_buffer_size {
metadata = self.process_magic_types();
// Mark as finalized when we've processed enough data
self.is_finalized = true;
}
}
let is_finalized = !metadata.is_empty();
crate::meta_plugin::MetaPluginResponse {
metadata,
is_finalized,
}
}
fn meta_type(&self) -> MetaPluginType {
MetaPluginType::MagicFile
}
fn outputs(&self) -> &std::collections::HashMap<String, serde_yaml::Value> {
self.base.outputs()
}
fn outputs_mut(&mut self) -> &mut std::collections::HashMap<String, serde_yaml::Value> {
self.base.outputs_mut()
}
fn default_outputs(&self) -> Vec<String> {
vec!["mime_type".to_string(), "mime_encoding".to_string(), "file_type".to_string()]
}
fn options(&self) -> &std::collections::HashMap<String, serde_yaml::Value> {
self.base.options()
}
fn options_mut(&mut self) -> &mut std::collections::HashMap<String, serde_yaml::Value> {
self.base.options_mut()
}
}
use crate::meta_plugin::register_meta_plugin;
// Register the plugin at module initialization time
#[ctor::ctor]
fn register_magic_file_plugin() {
register_meta_plugin(MetaPluginType::MagicFile, |options, outputs| {
Box::new(MagicFileMetaPlugin::new(options, outputs))
});
}

View File

@@ -6,7 +6,7 @@ use once_cell::sync::Lazy;
pub mod exec; pub mod exec;
pub mod digest; pub mod digest;
pub mod magic; pub mod magic_file;
pub mod text; pub mod text;
pub mod read_time; pub mod read_time;
pub mod read_rate; pub mod read_rate;
@@ -20,7 +20,7 @@ pub mod env;
pub use exec::MetaPluginExec; pub use exec::MetaPluginExec;
pub use digest::DigestMetaPlugin; pub use digest::DigestMetaPlugin;
pub use magic::MagicFileMetaPlugin; pub use magic_file::MagicFileMetaPlugin;
pub use text::TextMetaPlugin; pub use text::TextMetaPlugin;
pub use read_time::ReadTimeMetaPlugin; pub use read_time::ReadTimeMetaPlugin;
pub use read_rate::ReadRateMetaPlugin; pub use read_rate::ReadRateMetaPlugin;