rsconstruct 0.9.78

use anyhow::{Context, Result};
use std::collections::BTreeMap;
use std::path::{Path, PathBuf};
use std::process::Command;

use serde::{Deserialize, Serialize};

use crate::config::{StandardConfig, output_config_hash, resolve_extra_inputs};
use crate::file_index::FileIndex;
use crate::graph::{BuildGraph, Product};
use crate::processors::{Processor, run_command, check_command_output};

fn default_pdfunite_source_dir() -> String {
    "marp/courses".into()
}

fn default_pdfunite_source_ext() -> String {
    ".md".into()
}

fn default_pdfunite_source_output_dir() -> String {
    "out/marp".into()
}

#[derive(Debug, Deserialize, Serialize, Clone)]
pub struct PdfuniteConfig {
    #[serde(default = "default_pdfunite_source_dir")]
    pub source_dir: String,
    #[serde(default = "default_pdfunite_source_ext")]
    pub source_ext: String,
    #[serde(default = "default_pdfunite_source_output_dir")]
    pub source_output_dir: String,
    #[serde(flatten)]
    pub standard: StandardConfig,
}

impl Default for PdfuniteConfig {
    fn default() -> Self {
        Self {
            source_dir: "marp/courses".into(),
            source_ext: ".md".into(),
            source_output_dir: "out/marp".into(),
            standard: StandardConfig::default(),
        }
    }
}

pub struct PdfuniteProcessor {
    config: PdfuniteConfig,
}

impl PdfuniteProcessor {
    pub const fn new(config: PdfuniteConfig) -> Self {
        Self {
            config,
        }
    }

    /// Source files under `source_dir` with the configured extension, grouped
    /// by their containing directory (sorted, so both the directory order and
    /// the merge order within a directory are deterministic).
    ///
    /// Goes through `FileIndex` rather than walking the filesystem: that is
    /// what makes `.rsconstructignore` apply and lets virtual files produced
    /// by an upstream processor be merged. Walking `fs::read_dir` here saw
    /// neither, and re-ran the IO on every fixed-point discovery pass.
    fn source_dirs(&self, file_index: &FileIndex) -> Vec<(PathBuf, Vec<PathBuf>)> {
        let ext = if self.config.source_ext.starts_with('.') {
            self.config.source_ext.clone()
        } else {
            format!(".{}", self.config.source_ext)
        };
        let base = Path::new(&self.config.source_dir);
        let files = file_index.query(base, &[&ext], &[], &[], &[], &[]);

        let mut by_dir: BTreeMap<PathBuf, Vec<PathBuf>> = BTreeMap::new();
        for file in files {
            let dir = crate::processors::parent_dir_or_empty(&file).to_path_buf();
            by_dir.entry(dir).or_default().push(file);
        }
        for files in by_dir.values_mut() {
            files.sort();
        }
        by_dir.into_iter().collect()
    }
}

impl Processor for PdfuniteProcessor {
    fn scan_config(&self) -> &crate::config::StandardConfig {
        &self.config.standard
    }

    // Serialize the FULL config (the trait default covers StandardConfig
    // only), so the extra fields reach config-change detection.
    fn config_json(&self) -> Option<String> {
        crate::processors::ProcessorBase::config_json(&self.config)
    }

    fn clean(&self, product: &crate::graph::Product, verbose: bool) -> anyhow::Result<usize> {
        crate::processors::ProcessorBase::clean(product, &product.processor, verbose)
    }

    fn auto_detect(&self, file_index: &FileIndex) -> bool {
        !self.source_dirs(file_index).is_empty()
    }

    fn required_tools(&self) -> Vec<String> {
        vec![self.config.standard.command.clone()]
    }

    fn discover(&self, graph: &mut BuildGraph, file_index: &FileIndex, instance_name: &str) -> Result<()> {
        let base = Path::new(&self.config.source_dir);

        let hash = Some(output_config_hash(&self.config, &crate::config::checksum_fields_of(instance_name)));
        let extra = resolve_extra_inputs(&self.config.standard.dep_inputs)?;

        // Compute upstream scan dir once
        let upstream_scan_dir = Path::new(&self.config.source_dir)
            .components()
            .next()
            .map(|c| c.as_os_str().to_string_lossy().into_owned())
            .context("source_dir is empty")?;
        let upstream_scan_dirs = [upstream_scan_dir];

        for (dir_path, source_files) in self.source_dirs(file_index) {
            let inputs: Vec<PathBuf> = source_files.iter().map(|src| {
                super::output_path(src, &upstream_scan_dirs, &self.config.source_output_dir, "pdf")
            }).chain(extra.iter().cloned()).collect();

            // Mirror the directory structure from source_dir into output_dir,
            // naming each merged PDF after its leaf directory.
            let relative = dir_path.strip_prefix(base).unwrap_or(&dir_path);
            let parent = crate::processors::parent_dir_or_empty(relative);
            let leaf = relative.file_name()
                .with_context(|| format!("Cannot extract leaf directory name from {}", dir_path.display()))?;
            let outputs = vec![
                Path::new(&self.config.standard.output_dir).join(parent).join(format!("{}.pdf", leaf.to_string_lossy())),
            ];

            graph.add_product(inputs, outputs, instance_name, hash.clone())?;
        }

        Ok(())
    }

    fn execute(&self, ctx: &crate::build_context::BuildContext, product: &Product) -> Result<()> {
        let output = product.primary_output();

        crate::processors::ensure_output_dir(output)?;

        // Inputs also carry dep_inputs (extra rebuild triggers); only actual
        // PDFs may be passed to pdfunite as pages.
        let pdf_inputs: Vec<&PathBuf> = product.inputs.iter()
            .filter(|p| p.extension().is_some_and(|e| e == "pdf"))
            .collect();
        if pdf_inputs.is_empty() {
            anyhow::bail!("No PDF inputs found for {}", output.display());
        }

        let mut cmd = Command::new(&self.config.standard.command);
        for arg in &self.config.standard.args {
            cmd.arg(arg);
        }
        // pdfunite takes input files followed by output file
        for input in pdf_inputs {
            cmd.arg(input);
        }
        cmd.arg(output);

        let out = run_command(ctx, &cmd)?;
        check_command_output(&out, format_args!("pdfunite {}", output.display()))
    }

}

fn plugin_create(toml: &toml::Value) -> anyhow::Result<Box<dyn crate::processors::Processor>> {
    crate::registries::deserialize_and_create(toml, |cfg| Box::new(PdfuniteProcessor::new(cfg)))
}
inventory::submit! {
    crate::registries::ProcessorPlugin {
        version: 1,
        name: "pdfunite",
        processor_type: crate::processors::ProcessorType::Generator,
        create: plugin_create,
        fields: &[
            crate::config::FieldSpec { name: "source_dir", ty: crate::config::FieldType::String,
                affects_output: true, required: false,
                doc: "Directory containing course YAML files listing PDFs to merge" },
            crate::config::FieldSpec { name: "source_ext", ty: crate::config::FieldType::String,
                affects_output: true, required: false,
                doc: "Extension of source files used to find PDFs" },
            crate::config::FieldSpec { name: "source_output_dir", ty: crate::config::FieldType::String,
                affects_output: true, required: false,
                doc: "Directory where source PDFs (to be merged) are located" },
        ],
        omit_standard_fields: &["formats"],
        scan_defaults: Some(crate::config::ScanDefaultsData { src_dirs: &[], src_extensions: &["course.yaml"], src_exclude_dirs: &[] }),
        defaults: Some(crate::config::ProcessorDefaults { command: "pdfunite", output_dir: "out/pdfunite", ..crate::config::ProcessorDefaults::EMPTY }),
        defconfig_json: crate::registries::default_config_json::<PdfuniteConfig>,
        keywords: &["pdf", "merger", "generator"],
        description: "Merge PDFs from subdirectories into course bundles",
        is_native: false,
        can_fix: false,
        supports_batch: false,
        max_jobs_cap: Some(1),
    }
}