package_info: refactor into distro_info and package_info split, yaml data

src/distro_info.rs

@@ -0,0 +1,185 @@
+use chrono::NaiveDate;
+use lazy_static::lazy_static;
+use serde::Deserialize;
+use std::error::Error;
+use std::path::Path;
+
+#[derive(Debug, Deserialize)]
+struct SeriesInfo {
+    local: String,
+    network: String,
+}
+
+#[derive(Debug, Deserialize)]
+struct DistData {
+    base_url: String,
+    pockets: Vec<String>,
+    series: SeriesInfo,
+}
+
+#[derive(Debug, Deserialize)]
+struct Data {
+    dist: std::collections::HashMap<String, DistData>,
+}
+
+const DATA_YAML: &str = include_str!("../distro_info.yml");
+lazy_static! {
+    static ref DATA: Data = serde_yaml::from_str(DATA_YAML).unwrap();
+}
+
+fn parse_series_csv(content: &str) -> Result<Vec<String>, Box<dyn Error>> {
+    let mut rdr = csv::ReaderBuilder::new()
+        .flexible(true)
+        .from_reader(content.as_bytes());
+
+    let headers = rdr.headers()?.clone();
+    let series_idx = headers
+        .iter()
+        .position(|h| h == "series")
+        .ok_or("Column 'series' not found")?;
+    let created_idx = headers
+        .iter()
+        .position(|h| h == "created")
+        .ok_or("Column 'created' not found")?;
+
+    let mut entries = Vec::new();
+    for result in rdr.records() {
+        let record = result?;
+        if let (Some(s), Some(c)) = (record.get(series_idx), record.get(created_idx))
+            && let Ok(date) = NaiveDate::parse_from_str(c, "%Y-%m-%d")
+        {
+            entries.push((s.to_string(), date));
+        }
+    }
+
+    // Sort by date descending (newest first)
+    entries.sort_by(|a, b| b.1.cmp(&a.1));
+
+    Ok(entries.into_iter().map(|(s, _)| s).collect())
+}
+
+/// Get time-ordered list of series for a distribution, development series first
+pub async fn get_ordered_series(dist: &str) -> Result<Vec<String>, Box<dyn Error>> {
+    let series_info = &DATA.dist.get(dist).unwrap().series;
+    let content = if Path::new(series_info.local.as_str()).exists() {
+        std::fs::read_to_string(format!("/usr/share/distro-info/{dist}.csv"))?
+    } else {
+        reqwest::get(series_info.network.as_str())
+            .await?
+            .text()
+            .await?
+    };
+
+    let mut series = parse_series_csv(&content)?;
+
+    // For Debian, ensure 'sid' is first if it's not
+    // We want to try 'sid' (unstable) first for Debian.
+    if dist == "debian" {
+        series.retain(|s| s != "sid");
+        series.insert(0, "sid".to_string());
+    }
+
+    Ok(series)
+}
+
+/// Obtain the distribution (eg. debian, ubuntu) from a distribution series (eg. noble, bookworm)
+pub async fn get_dist_from_series(series: &str) -> Result<String, Box<dyn Error>> {
+    for dist in DATA.dist.keys() {
+        if get_ordered_series(dist)
+            .await?
+            .contains(&series.to_string())
+        {
+            return Ok(dist.to_string());
+        }
+    }
+    Err(format!("Unknown series: {}", series).into())
+}
+
+/// Get the package pockets available for a given distribution
+///
+/// Example: get_dist_pockets(ubuntu) => ["proposed", "updates", ""]
+pub fn get_dist_pockets(dist: &str) -> Vec<String> {
+    let mut pockets = DATA.dist.get(dist).unwrap().pockets.clone();
+
+    // Explicitely add 'main' pocket, which is just the empty string
+    pockets.push("".to_string());
+
+    pockets
+}
+
+/// Get the sources URL for a distribution, series, pocket, and component
+pub fn get_sources_url(base_url: &str, series: &str, pocket: &str, component: &str) -> String {
+    let pocket_full = if pocket.is_empty() {
+        String::new()
+    } else {
+        format!("-{}", pocket)
+    };
+    format!("{base_url}/dists/{series}{pocket_full}/{component}/source/Sources.gz")
+}
+
+/// Get the archive base URL for a distribution
+///
+/// Example: ubuntu => http://archive.ubuntu.com/ubuntu
+pub fn get_base_url(dist: &str) -> String {
+    DATA.dist.get(dist).unwrap().base_url.clone()
+}
+
+/// Obtain the URL for the 'Release' file of a distribution series
+fn get_release_url(base_url: &str, series: &str, pocket: &str) -> String {
+    let pocket_full = if pocket.is_empty() {
+        String::new()
+    } else {
+        format!("-{}", pocket)
+    };
+    format!("{base_url}/dists/{series}{pocket_full}/Release")
+}
+
+/// Obtain the components of a distribution series by parsing the 'Release' file
+pub async fn get_components(
+    base_url: &str,
+    series: &str,
+    pocket: &str,
+) -> Result<Vec<String>, Box<dyn Error>> {
+    let url = get_release_url(base_url, series, pocket);
+    log::debug!("Fetching Release file from: {}", url);
+
+    let content = reqwest::get(&url).await?.text().await?;
+
+    for line in content.lines() {
+        if line.starts_with("Components:")
+            && let Some((_, components)) = line.split_once(':')
+        {
+            return Ok(components
+                .split_whitespace()
+                .map(|s| s.to_string())
+                .collect());
+        }
+    }
+
+    Err("Components not found.".into())
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[tokio::test]
+    async fn test_get_debian_series() {
+        let series = get_ordered_series("debian").await.unwrap();
+        assert!(series.contains(&"sid".to_string()));
+        assert!(series.contains(&"bookworm".to_string()));
+    }
+
+    #[tokio::test]
+    async fn test_get_ubuntu_series() {
+        let series = get_ordered_series("ubuntu").await.unwrap();
+        assert!(series.contains(&"noble".to_string()));
+        assert!(series.contains(&"jammy".to_string()));
+    }
+
+    #[tokio::test]
+    async fn test_get_dist_from_series() {
+        assert_eq!(get_dist_from_series("sid").await.unwrap(), "debian");
+        assert_eq!(get_dist_from_series("noble").await.unwrap(), "ubuntu");
+    }
+}