cuddle-please/crates/cuddle-please-misc/src/cliff/mod.rs

use anyhow::Context;
use chrono::{DateTime, NaiveDate, Utc};
use git_cliff_core::{
    changelog::Changelog,
    commit::Commit,
    config::{ChangelogConfig, CommitParser, Config, GitConfig},
    release::Release,
};
use regex::Regex;

pub struct ChangeLogBuilder {
    commits: Vec<String>,
    version: String,
    config: Option<Config>,
    release_date: Option<NaiveDate>,
    release_link: Option<String>,
}

impl ChangeLogBuilder {
    pub fn new<C>(commits: C, version: impl Into<String>) -> Self
    where
        C: IntoIterator,
        C::Item: AsRef<str>,
    {
        Self {
            commits: commits
                .into_iter()
                .map(|s| s.as_ref().to_string())
                .collect(),
            version: version.into(),
            config: None,
            release_date: None,
            release_link: None,
        }
    }

    pub fn with_release_date(self, release_date: NaiveDate) -> Self {
        Self {
            release_date: Some(release_date),
            ..self
        }
    }

    pub fn with_release_link(self, release_link: impl Into<String>) -> Self {
        Self {
            release_link: Some(release_link.into()),
            ..self
        }
    }

    pub fn with_config(self, config: Config) -> Self {
        Self {
            config: Some(config),
            ..self
        }
    }

    pub fn build<'a>(self) -> ChangeLog<'a> {
        let git_config = self
            .config
            .clone()
            .map(|c| c.git)
            .unwrap_or_else(default_git_config);
        let timestamp = self.release_timestamp();
        let commits = self
            .commits
            .clone()
            .into_iter()
            .map(|c| Commit::new("id".into(), c))
            .filter_map(|c| c.process(&git_config).ok())
            .collect();

        ChangeLog {
            release: Release {
                version: Some(self.version),
                commits,
                commit_id: None,
                timestamp,
                previous: None,
            },
            config: self.config,
            release_link: self.release_link,
        }
    }

    fn release_timestamp(&self) -> i64 {
        self.release_date
            .and_then(|date| date.and_hms_opt(0, 0, 0))
            .map(|d| DateTime::<Utc>::from_utc(d, Utc))
            .unwrap_or_else(Utc::now)
            .timestamp()
    }
}

pub struct ChangeLog<'a> {
    release: Release<'a>,
    config: Option<Config>,
    release_link: Option<String>,
}

impl ChangeLog<'_> {
    pub fn generate(&self) -> anyhow::Result<String> {
        let config = self.config.clone().unwrap_or_else(|| self.default_config());
        let changelog = Changelog::new(vec![self.release.clone()], &config)?;
        let mut buffer = Vec::new();
        changelog
            .generate(&mut buffer)
            .context("failed to generate changelog")?;
        String::from_utf8(buffer)
            .context("cannot convert bytes to string (contains non utf-8 char indices)")
    }

    pub fn prepend(self, old_changelog: impl Into<String>) -> anyhow::Result<String> {
        let old_changelog = old_changelog.into();
        if let Ok(Some(last_version)) = changelog_parser::last_version_from_str(&old_changelog) {
            let next_version = self
                .release
                .version
                .as_ref()
                .context("current release contains no version")?;
            if next_version == &last_version {
                return Ok(old_changelog);
            }
        }

        let old_header = changelog_parser::parse_header(&old_changelog);
        let config = self
            .config
            .clone()
            .unwrap_or_else(|| self.default_config_with_header(old_header));
        let changelog = Changelog::new(vec![self.release], &config)?;
        let mut out = Vec::new();
        changelog.prepend(old_changelog, &mut out)?;
        String::from_utf8(out)
            .context("cannot convert bytes to string (contains non utf-8 char indices)")
    }

    fn default_config(&self) -> Config {
        let config = Config {
            changelog: default_changelog_config(None, self.release_link.as_deref()),
            git: default_git_config(),
        };

        config
    }

    fn default_config_with_header(&self, header: Option<String>) -> Config {
        let config = Config {
            changelog: default_changelog_config(header, self.release_link.as_deref()),
            git: default_git_config(),
        };

        config
    }
}

fn default_git_config() -> GitConfig {
    GitConfig {
        conventional_commits: Some(true),
        filter_unconventional: Some(false),
        filter_commits: Some(true),
        commit_parsers: Some(default_commit_parsers()),
        ..Default::default()
    }
}

fn default_commit_parsers() -> Vec<CommitParser> {
    fn create_commit_parser(message: &str, group: &str) -> CommitParser {
        CommitParser {
            message: Regex::new(&format!("^{message}")).ok(),
            body: None,
            group: Some(group.into()),
            default_scope: None,
            scope: None,
            skip: None,
        }
    }

    vec![
        create_commit_parser("feat", "added"),
        create_commit_parser("changed", "changed"),
        create_commit_parser("deprecated", "deprecated"),
        create_commit_parser("removed", "removed"),
        create_commit_parser("fix", "fixed"),
        create_commit_parser("security", "security"),
        create_commit_parser("docs", "docs"),
        CommitParser {
            message: Regex::new(".*").ok(),
            group: Some(String::from("other")),
            body: None,
            default_scope: None,
            skip: None,
            scope: None,
        },
    ]
}

const CHANGELOG_HEADER: &str = r#"# Changelog
All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]
"#;

fn default_changelog_config(header: Option<String>, release_link: Option<&str>) -> ChangelogConfig {
    ChangelogConfig {
        header: Some(header.unwrap_or(String::from(CHANGELOG_HEADER))),
        body: Some(default_changelog_body_config(release_link)),
        footer: None,
        trim: Some(true),
    }
}

fn default_changelog_body_config(release_link: Option<&str>) -> String {
    const PRE: &str = r#"
    ## [{{ version | trim_start_matches(pat="v") }}]"#;
    const POST: &str = r#" - {{ timestamp | date(format="%Y-%m-%d") }}
{% for group, commits in commits | group_by(attribute="group") %}
### {{ group | upper_first }}
{% for commit in commits %}
{%- if commit.scope -%}
- *({{commit.scope}})* {% if commit.breaking %}[**breaking**] {% endif %}{{ commit.message }}{%- if commit.links %} ({% for link in commit.links %}[{{link.text}}]({{link.href}}) {% endfor -%}){% endif %}
{% else -%}
- {% if commit.breaking %}[**breaking**] {% endif %}{{ commit.message }}
{% endif -%}
{% endfor -%}
{% endfor %}"#;

    match release_link {
        Some(link) => format!("{}{}{}", PRE, link, POST),
        None => format!("{}{}", PRE, POST),
    }
}

pub mod changelog_parser {

    use anyhow::Context;
    use regex::Regex;

    /// Parse the header from a changelog.
    /// The changelog header is a string at the begin of the changelog that:
    /// - Starts with `# Changelog`, `# CHANGELOG`, or `# changelog`
    /// - ends with `## Unreleased`, `## [Unreleased]` or `## ..anything..`
    ///   (in the ..anything.. case, `## ..anything..` is not included in the header)
    pub fn parse_header(changelog: &str) -> Option<String> {
        lazy_static::lazy_static! {
            static ref FIRST_RE: Regex = Regex::new(r"(?s)^(# Changelog|# CHANGELOG|# changelog)(.*)(## Unreleased|## \[Unreleased\])").unwrap();

            static ref SECOND_RE: Regex = Regex::new(r"(?s)^(# Changelog|# CHANGELOG|# changelog)(.*)(\n## )").unwrap();
        }
        if let Some(captures) = FIRST_RE.captures(changelog) {
            return Some(format!("{}\n", &captures[0]));
        }

        if let Some(captures) = SECOND_RE.captures(changelog) {
            return Some(format!("{}{}", &captures[1], &captures[2]));
        }

        None
    }

    pub fn last_changes(changelog: &str) -> anyhow::Result<Option<String>> {
        last_changes_from_str(changelog)
    }

    pub fn last_changes_from_str(changelog: &str) -> anyhow::Result<Option<String>> {
        let parser = ChangelogParser::new(changelog)?;
        let last_release = parser.last_release().map(|r| r.notes.to_string());
        Ok(last_release)
    }

    pub fn last_version_from_str(changelog: &str) -> anyhow::Result<Option<String>> {
        let parser = ChangelogParser::new(changelog)?;
        let last_release = parser.last_release().map(|r| r.version.to_string());
        Ok(last_release)
    }

    pub fn last_release_from_str(changelog: &str) -> anyhow::Result<Option<ChangelogRelease>> {
        let parser = ChangelogParser::new(changelog)?;
        let last_release = parser.last_release().map(ChangelogRelease::from_release);
        Ok(last_release)
    }

    pub struct ChangelogRelease {
        title: String,
        notes: String,
    }

    impl ChangelogRelease {
        fn from_release(release: &parse_changelog::Release) -> Self {
            Self {
                title: release.title.to_string(),
                notes: release.notes.to_string(),
            }
        }

        pub fn title(&self) -> &str {
            &self.title
        }

        pub fn notes(&self) -> &str {
            &self.notes
        }
    }

    pub struct ChangelogParser<'a> {
        changelog: parse_changelog::Changelog<'a>,
    }

    impl<'a> ChangelogParser<'a> {
        pub fn new(changelog_text: &'a str) -> anyhow::Result<Self> {
            let changelog =
                parse_changelog::parse(changelog_text).context("can't parse changelog")?;
            Ok(Self { changelog })
        }

        fn last_release(&self) -> Option<&parse_changelog::Release> {
            let last_release = release_at(&self.changelog, 0)?;
            let last_release = if last_release.version.to_lowercase().contains("unreleased") {
                release_at(&self.changelog, 1)?
            } else {
                last_release
            };
            Some(last_release)
        }
    }

    fn release_at<'a>(
        changelog: &'a parse_changelog::Changelog,
        index: usize,
    ) -> Option<&'a parse_changelog::Release<'a>> {
        let release = changelog.get_index(index)?.1;
        Some(release)
    }

    #[cfg(test)]
    mod tests {
        use super::*;

        fn last_changes_from_str_test(changelog: &str) -> String {
            last_changes_from_str(changelog).unwrap().unwrap()
        }

        #[test]
        fn changelog_header_is_parsed() {
            let changelog = "\
# Changelog

My custom changelog header

## [Unreleased]
";
            let header = parse_header(changelog).unwrap();
            let expected_header = "\
# Changelog

My custom changelog header

## [Unreleased]
";
            assert_eq!(header, expected_header);
        }

        #[test]
        fn changelog_header_without_unreleased_is_parsed() {
            let changelog = "\
# Changelog

My custom changelog header

## [0.2.5] - 2022-12-16
";
            let header = parse_header(changelog).unwrap();
            let expected_header = "\
# Changelog

My custom changelog header
";
            assert_eq!(header, expected_header);
        }

        #[test]
        fn changelog_header_with_versions_is_parsed() {
            let changelog = "\
# Changelog

My custom changelog header

## [Unreleased]

## [0.2.5] - 2022-12-16
";
            let header = parse_header(changelog).unwrap();
            let expected_header = "\
# Changelog

My custom changelog header

## [Unreleased]
";
            assert_eq!(header, expected_header);
        }

        #[test]
        fn changelog_header_isnt_recognized() {
            // A two-level header similar to `## [Unreleased]` is missing
            let changelog = "\
# Changelog

My custom changelog header
";
            let header = parse_header(changelog);
            assert_eq!(header, None);
        }

        #[test]
        fn changelog_with_unreleased_section_is_parsed() {
            let changelog = "\
# Changelog
All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]

## [0.2.5] - 2022-12-16

### Added
- Add function to retrieve default branch (#372)

## [0.2.4] - 2022-12-12

### Changed
- improved error message
";
            let changes = last_changes_from_str_test(changelog);
            let expected_changes = "\
### Added
- Add function to retrieve default branch (#372)";
            assert_eq!(changes, expected_changes);
        }

        #[test]
        fn changelog_without_unreleased_section_is_parsed() {
            let changelog = "\
# Changelog
All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [0.2.5](https://github.com/MarcoIeni/release-plz/compare/git_cmd-v0.2.4...git_cmd-v0.2.5) - 2022-12-16

### Added
- Add function to retrieve default branch (#372)

## [0.2.4] - 2022-12-12

### Changed
- improved error message
";
            let changes = last_changes_from_str_test(changelog);
            let expected_changes = "\
### Added
- Add function to retrieve default branch (#372)";
            assert_eq!(changes, expected_changes);
        }
    }
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn bare_release() {
        let commits: Vec<&str> = Vec::new();
        let changelog = ChangeLogBuilder::new(commits, "0.0.0")
            .with_release_date(NaiveDate::from_ymd_opt(1995, 5, 15).unwrap())
            .build();

        let expected = r######"# Changelog
All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]
"######;

        pretty_assertions::assert_eq!(expected, &changelog.generate().unwrap())
    }

    #[test]
    fn generates_changelog() {
        let commits: Vec<&str> = vec![
            "feat: some feature",
            "some random commit",
            "fix: some fix",
            "chore(scope): some chore",
        ];
        let changelog = ChangeLogBuilder::new(commits, "1.0.0")
            .with_release_date(NaiveDate::from_ymd_opt(1995, 5, 15).unwrap())
            .build();

        let expected = r######"# Changelog
All notable changes to this project will be documented in this file.

The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).

## [Unreleased]

## [1.0.0] - 1995-05-15

### Added
- some feature

### Fixed
- some fix

### Other
- some random commit
- *(scope)* some chore
"######;

        pretty_assertions::assert_eq!(expected, &changelog.generate().unwrap())
    }
}