v6/src.bak/generator/posts/parse.rs

use crate::generator;
use crate::generator::markdown;
use crate::generator::util::slugify::slugify;
use crate::generator::util::{from_frontmatter, word_count};
use anyhow::anyhow;
use chrono::{DateTime, Datelike, FixedOffset};
use once_cell::sync::Lazy;
use regex::Regex;
use serde::de::{SeqAccess, Visitor};
use serde::Deserialize;
use std::fs;
use std::hash::Hash;
use std::io::Read;
use std::path::PathBuf;

pub async fn parse() -> anyhow::Result<Vec<Post<HtmlContent>>> {
    let posts_path = generator::content_path("posts/");
    let mut posts = vec![];
    for ent in fs::read_dir(posts_path)? {
        let ent = ent.unwrap();
        let path: PathBuf;
        if ent.file_type().unwrap().is_dir() {
            path = find_index(ent.path()).expect("folder posts must have index file");
        } else {
            path = ent.path();
        }
        let post = parse_post(path)?;
        let prepared = prepare_post(post);
        posts.push(prepared);
    }
    posts.sort_by(|a, b| b.metadata.date.cmp(&a.metadata.date));
    Ok(posts)
}

fn find_index(path: PathBuf) -> Option<PathBuf> {
    let dir = std::fs::read_dir(path).ok()?;
    dir.map(|e| e.unwrap())
        .find(|e| e.path().file_stem().unwrap().eq_ignore_ascii_case("index"))
        .map(|e| e.path())
}

pub fn parse_post(path: PathBuf) -> anyhow::Result<Post<AnyContent>> {
    let mut f = fs::File::open(&path)?;
    let mut buffer = String::new();
    f.read_to_string(&mut buffer)?;
    Post::new(path, &buffer)
}

static EXCERPT_REGEX: Lazy<Regex> = Lazy::new(|| Regex::new("<!--\\s*excerpt-end\\s*-->").unwrap());

pub fn prepare_post(post: Post<AnyContent>) -> Post<HtmlContent> {
    let wc = post.content.word_count();
    let mut new_post = post.to_html();
    new_post.word_count = Some(wc);
    if let Some(m) = EXCERPT_REGEX.find(&new_post.content.0) {
        new_post.excerpt = Some(new_post.content.0[0..m.start()].trim().into());
    }
    new_post
}

#[derive(Debug, PartialEq)]
pub struct Post<Content: PostContent> {
    pub path: PathBuf,
    pub metadata: PostMetadata,
    pub slug: String,
    pub word_count: Option<u32>,
    pub excerpt: Option<String>,
    pub content: Content,
}

impl Post<AnyContent> {
    pub fn new(path: PathBuf, contents: &str) -> anyhow::Result<Self> {
        let (metadata, rest_contents) = match from_frontmatter::<'_, PostMetadata>(contents) {
            Ok(res) => res,
            Err(e) => return Err(e),
        };

        let slug = metadata
            .slug
            .clone()
            .unwrap_or_else(|| slugify(&metadata.title));

        let ext = path.extension().unwrap().to_str().unwrap();
        let content = content_from(ext, rest_contents)?;

        Ok(Post {
            path,
            metadata,
            slug: slug.to_owned(),
            word_count: None,
            excerpt: None,
            content,
        })
    }
}

impl<C: PostContent> Post<C> {
    fn to_html(self) -> Post<HtmlContent> {
        Post {
            path: self.path,
            metadata: self.metadata,
            slug: self.slug,
            word_count: self.word_count,
            excerpt: self.excerpt,
            content: self.content.to_html(),
        }
    }

    pub fn permalink(&self) -> String {
        format!("/{}/{}/", self.metadata.date.year(), self.slug)
    }

    pub fn comments_permalink(&self) -> String {
        match self.metadata.use_old_permalink_for_comments {
            Some(true) => {
                let old = self.metadata.old_permalink.as_ref().unwrap();
                assert_eq!(old.len(), 1);
                return old[0].clone();
            }
            _ => self.permalink(),
        }
    }
}

#[derive(Debug, Deserialize, PartialEq, Clone)]
#[serde(deny_unknown_fields)]
pub struct PostMetadata {
    pub title: String,
    pub html_title: Option<String>,
    pub tags: Option<Vec<Tag>>,
    pub date: DateTime<FixedOffset>,
    pub short_desc: Option<String>,
    slug: Option<String>,
    pub preamble: Option<String>,
    #[serde(deserialize_with = "deserialize_old_permalink", default)]
    pub old_permalink: Option<Vec<String>>,
    pub use_old_permalink_for_comments: Option<bool>,
    pub card_image_path: Option<String>,
}

fn deserialize_old_permalink<'de, D>(deserializer: D) -> Result<Option<Vec<String>>, D::Error>
where
    D: serde::Deserializer<'de>,
{
    struct StringOrVec;
    impl<'de> Visitor<'de> for StringOrVec {
        type Value = Option<Vec<String>>;

        fn expecting(&self, formatter: &mut std::fmt::Formatter) -> std::fmt::Result {
            formatter.write_str("string or vec of strings")
        }

        fn visit_str<E>(self, value: &str) -> Result<Self::Value, E>
        where
            E: serde::de::Error,
        {
            Ok(Some(vec![value.to_owned()]))
        }

        fn visit_seq<A>(self, mut seq: A) -> Result<Self::Value, A::Error>
        where
            A: SeqAccess<'de>,
        {
            let mut vec = Vec::with_capacity(seq.size_hint().unwrap_or(1));
            loop {
                match seq.next_element::<String>() {
                    Ok(Some(s)) => vec.push(s),
                    Ok(None) => break,
                    Err(e) => return Err(e),
                }
            }
            Ok(Some(vec))
        }

        fn visit_none<E>(self) -> Result<Self::Value, E>
        where
            E: serde::de::Error,
        {
            println!("visit_none");
            Ok(None)
        }
    }

    deserializer.deserialize_any(StringOrVec)
}

#[derive(Debug, Eq, Clone)]
pub struct Tag {
    pub name: String,
    pub slug: String,
}

impl<'de> Deserialize<'de> for Tag {
    fn deserialize<D>(deserializer: D) -> Result<Self, D::Error>
    where
        D: serde::Deserializer<'de>,
    {
        let name = String::deserialize(deserializer)?;
        let slug = slugify(&name);
        Ok(Tag { name, slug })
    }
}

impl PartialEq for Tag {
    fn eq(&self, other: &Self) -> bool {
        self.slug == other.slug
    }
}

impl Hash for Tag {
    fn hash<H: std::hash::Hasher>(&self, state: &mut H) {
        self.slug.hash(state);
    }
}

pub trait PostContent: std::fmt::Debug + PartialEq {
    fn to_html(self) -> HtmlContent;
    fn word_count(&self) -> u32;
}

#[derive(Debug, PartialEq)]
pub enum AnyContent {
    Markdown(MarkdownContent),
    Html(HtmlContent),
}

impl PostContent for AnyContent {
    fn to_html(self) -> HtmlContent {
        match self {
            AnyContent::Markdown(inner) => inner.to_html(),
            AnyContent::Html(inner) => inner,
        }
    }

    fn word_count(&self) -> u32 {
        match self {
            AnyContent::Markdown(inner) => inner.word_count(),
            AnyContent::Html(inner) => inner.word_count(),
        }
    }
}

#[derive(Debug, PartialEq)]
pub struct MarkdownContent(String);

impl PostContent for MarkdownContent {
    fn to_html(self) -> HtmlContent {
        let mut buf = vec![];
        markdown::render(&self.0, &mut buf);
        HtmlContent(String::from_utf8(buf).unwrap())
    }

    fn word_count(&self) -> u32 {
        word_count::markdown(&self.0)
    }
}

#[derive(Debug, Clone, PartialEq)]
pub struct HtmlContent(String);

impl PostContent for HtmlContent {
    fn to_html(self) -> HtmlContent {
        self
    }

    fn word_count(&self) -> u32 {
        word_count::html(&self.0)
    }
}

impl HtmlContent {
    pub fn html(&self) -> &str {
        &self.0
    }
}

fn content_from(extension: &str, content: &str) -> anyhow::Result<AnyContent> {
    match extension {
        "md" => Ok(AnyContent::Markdown(MarkdownContent(content.to_owned()))),
        "html" => Ok(AnyContent::Html(HtmlContent(content.to_owned()))),
        _ => Err(anyhow!("unknown extension {}", extension)),
    }
}

#[cfg(test)]
mod tests {
    use chrono::{DateTime, FixedOffset, TimeZone};
    use serde::Deserialize;

    #[derive(Deserialize)]
    struct DeserializedDate {
        d: DateTime<FixedOffset>,
    }

    #[test]
    fn test_deserialize_date() {
        let deserialized: DeserializedDate =
            toml::from_str(r#"d = "2017-02-17 14:30:42 -0400""#).unwrap();
        let expected = FixedOffset::west(4 * 3600)
            .ymd(2017, 2, 17)
            .and_hms(14, 30, 42);
        assert_eq!(deserialized.d, expected);
    }

    #[test]
    fn test_deserialize_old_permalink() {
        let none: super::PostMetadata = toml::from_str(
            r#"
title = "Mocking HTTP Requests for iOS App UI Tests"
date = "2019-12-22 19:12:42 -0400"
            "#,
        )
        .unwrap();
        assert_eq!(none.old_permalink, None);
        let single: super::PostMetadata = toml::from_str(
            r#"
title = "Mocking HTTP Requests for iOS App UI Tests"
date = "2019-12-22 19:12:42 -0400"
old_permalink = "/ios/2019/mock-http-ios-ui-testing/"
            "#,
        )
        .unwrap();
        assert_eq!(
            single.old_permalink,
            Some(vec!["/ios/2019/mock-http-ios-ui-testing/".into()])
        );
        let multi: super::PostMetadata = toml::from_str(
            r#"
title = "Mocking HTTP Requests for iOS App UI Tests"
date = "2019-12-22 19:12:42 -0400"
old_permalink = ["/ios/2019/mock-http-ios-ui-testing/", "something else"]
            "#,
        )
        .unwrap();
        assert_eq!(
            multi.old_permalink,
            Some(vec![
                "/ios/2019/mock-http-ios-ui-testing/".into(),
                "something else".into()
            ])
        );
    }
}