refactor: move binary plugin to its own module

Co-authored-by: aider (openai/andrew/openrouter/qwen/qwen3-coder) <aider@aider.chat>
This commit is contained in:
Andrew Phillips
2025-08-18 20:30:33 -03:00
parent 40637c8881
commit 4ba8ce74cb
3 changed files with 95 additions and 89 deletions

View File

@@ -6,11 +6,13 @@ pub mod program;
pub mod digest;
pub mod system;
pub mod magic;
pub mod binary;
use crate::meta_plugin::program::MetaPluginProgram;
use crate::meta_plugin::digest::{DigestSha256MetaPlugin, ReadTimeMetaPlugin, ReadRateMetaPlugin};
use crate::meta_plugin::system::{CwdMetaPlugin, BinaryMetaPlugin, UidMetaPlugin, UserMetaPlugin, GidMetaPlugin, GroupMetaPlugin, ShellMetaPlugin, ShellPidMetaPlugin, KeepPidMetaPlugin, HostnameMetaPlugin, FullHostnameMetaPlugin};
use crate::meta_plugin::system::{CwdMetaPlugin, UidMetaPlugin, UserMetaPlugin, GidMetaPlugin, GroupMetaPlugin, ShellMetaPlugin, ShellPidMetaPlugin, KeepPidMetaPlugin, HostnameMetaPlugin, FullHostnameMetaPlugin};
use crate::meta_plugin::magic::MagicFileMetaPlugin;
use crate::meta_plugin::binary::BinaryMetaPlugin;
#[derive(Debug, Eq, PartialEq, Clone, strum::EnumIter, strum::Display, strum::EnumString)]
#[strum(ascii_case_insensitive)]

View File

@@ -0,0 +1,92 @@
use anyhow::Result;
use rusqlite::Connection;
use crate::common::is_binary::is_binary;
use crate::meta_plugin::{MetaPlugin, output_metadata};
#[derive(Debug, Clone, Default)]
pub struct BinaryMetaPlugin {
meta_name: String,
buffer: Vec<u8>,
max_buffer_size: usize,
is_saved: bool,
item_id: Option<i64>,
output_names: std::collections::HashMap<String, String>,
}
impl BinaryMetaPlugin {
pub fn new() -> BinaryMetaPlugin {
BinaryMetaPlugin {
meta_name: "binary".to_string(),
buffer: Vec::new(),
max_buffer_size: 4096, // 4KB
is_saved: false,
item_id: None,
output_names: std::collections::HashMap::new(),
}
}
}
impl MetaPlugin for BinaryMetaPlugin {
fn is_internal(&self) -> bool {
true
}
fn finalize(&mut self, conn: &Connection) -> Result<()> {
// Save the binary detection result when finalizing, after all data has been collected
if !self.is_saved {
if let Some(item_id) = self.item_id {
let is_binary = is_binary(&self.buffer);
let value = if is_binary { "true".to_string() } else { "false".to_string() };
// Save to database immediately using central output handler
let _ = output_metadata(conn, item_id, &self.meta_name, value, &self.output_names);
self.is_saved = true;
}
}
Ok(())
}
fn update(&mut self, data: &[u8], _conn: &Connection) {
// 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]);
}
}
fn meta_name(&mut self) -> String {
self.meta_name.clone()
}
fn initialize(&mut self, _conn: &Connection, item_id: i64) -> Result<()> {
self.item_id = Some(item_id);
Ok(())
}
fn configure(&mut self, options: &std::collections::HashMap<String, serde_yaml::Value>) -> Result<()> {
if let Some(max_buffer_size) = options.get("max_buffer_size") {
if let Some(size) = max_buffer_size.as_u64() {
self.max_buffer_size = size as usize;
}
}
if let Some(outputs) = options.get("outputs") {
if let Some(outputs_map) = outputs.as_mapping() {
for (key, value) in outputs_map {
if let (Some(key_str), Some(value_str)) = (key.as_str(), value.as_str()) {
self.output_names.insert(key_str.to_string(), value_str.to_string());
}
}
}
}
Ok(())
}
fn get_output_name(&self, default_name: &str) -> String {
self.output_names.get(default_name).cloned().unwrap_or_else(|| default_name.to_string())
}
}

View File

@@ -7,7 +7,6 @@ use std::process;
use uzers::{get_current_uid, get_current_gid, get_current_username, get_current_groupname};
use rusqlite::Connection;
use crate::common::is_binary::is_binary;
use crate::meta_plugin::{MetaPlugin, output_metadata};
#[derive(Debug, Clone, Default)]
@@ -17,93 +16,6 @@ pub struct CwdMetaPlugin {
output_names: std::collections::HashMap<String, String>,
}
#[derive(Debug, Clone, Default)]
pub struct BinaryMetaPlugin {
meta_name: String,
buffer: Vec<u8>,
max_buffer_size: usize,
is_saved: bool,
item_id: Option<i64>,
output_names: std::collections::HashMap<String, String>,
}
impl BinaryMetaPlugin {
pub fn new() -> BinaryMetaPlugin {
BinaryMetaPlugin {
meta_name: "binary".to_string(),
buffer: Vec::new(),
max_buffer_size: 4096, // 4KB
is_saved: false,
item_id: None,
output_names: std::collections::HashMap::new(),
}
}
}
impl MetaPlugin for BinaryMetaPlugin {
fn is_internal(&self) -> bool {
true
}
fn finalize(&mut self, conn: &Connection) -> Result<()> {
// Save the binary detection result when finalizing, after all data has been collected
if !self.is_saved {
if let Some(item_id) = self.item_id {
let is_binary = is_binary(&self.buffer);
let value = if is_binary { "true".to_string() } else { "false".to_string() };
// Save to database immediately using central output handler
let _ = output_metadata(conn, item_id, &self.meta_name, value, &self.output_names);
self.is_saved = true;
}
}
Ok(())
}
fn update(&mut self, data: &[u8], _conn: &Connection) {
// 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]);
}
}
fn meta_name(&mut self) -> String {
self.meta_name.clone()
}
fn initialize(&mut self, _conn: &Connection, item_id: i64) -> Result<()> {
self.item_id = Some(item_id);
Ok(())
}
fn configure(&mut self, options: &std::collections::HashMap<String, serde_yaml::Value>) -> Result<()> {
if let Some(max_buffer_size) = options.get("max_buffer_size") {
if let Some(size) = max_buffer_size.as_u64() {
self.max_buffer_size = size as usize;
}
}
if let Some(outputs) = options.get("outputs") {
if let Some(outputs_map) = outputs.as_mapping() {
for (key, value) in outputs_map {
if let (Some(key_str), Some(value_str)) = (key.as_str(), value.as_str()) {
self.output_names.insert(key_str.to_string(), value_str.to_string());
}
}
}
}
Ok(())
}
fn get_output_name(&self, default_name: &str) -> String {
self.output_names.get(default_name).cloned().unwrap_or_else(|| default_name.to_string())
}
}
impl CwdMetaPlugin {
pub fn new() -> CwdMetaPlugin {
CwdMetaPlugin {