Прочитать на русском

A lightweight blog platform, written in Haskell using Yesod. Using Redis for caching data and sqlite for providing sorted data.

This branch does not includes any CSS for web pages. If you're looking for already designed site, check main-bulma branch.

• Fully SSR! 🚀
• Posts stored in MD and YAML files
• Sitemap.xml autogeneration
• Internal Redis caching
• Headers ID autogeneration
• OpenGraph and JSON-LD generation for pages

• Generation of table of contents
• Option to disable header's id autogeneration
• React on 0 seconds cache time as its disabling

• How it works
• Feature details
• Usage
• Configuration
• Deploy
• License

The main idea of project is storing posts of blog in Markdown files in templates directory, so the name of files and folders will be URL of page. Each post consists of two files:

• Markdown file, which includes page content (sample file)
• YML file, which inclues information about post like title, image, description etc. (sample file)

For example, post https://YOURSITE.HERE/post/haskell/how-to-start-yesod will be stored in files ./templates/haskell/how-to-start-yesod.md and ./templates/haskell/how-to-start-yesod.yml.

Parsing of markdown file written in Attoparsec in Parser module and theoretically can be extended. Now, its implemented basic elements of Markdown like headers, lists, code blocks, quotes, images, links, bold and italic texts.

Result of parsing both files is stored in Redis for one minute. This timing can be changed in Redis module.

Also, blog-hs stores data about posts in SQLite database, which is used for showing latest posts on index page and generating sitemap.xml file. Database is creating manually using create-db command and stores data from YML files.

Currently it is not available for configuration. Each header gets id from its content - images transformed into its alternative text, text format is ignored. All text is converting into lowercase and gets joined using dash. Punctuation marks is ignored.

Next to each header will be paragraph symbol, which allows to go to this header.

All details specified in Parser/Inline.hs and Parser/Html.hs modules.

# This header will be transformed into this HTML code:

<h1 id="this-header-will-be-transformed-into-this-html-code"> This header will be transformed into HTML code: </h1>

Some symbols (like russian alphabet by default) gets translated into english symbols. All symbols except this by default will be ignored. For correcting this behavior see transliterateCharacter function.

Post info always stored in .yml (not .yaml) file with same name and have following fields:

Sample file is example.yml.

Each post can be attached to different categories using categories section. It must contain list of strings, which is category names. Using categories, you can group different posts and display then on post category screen.

Posts of current categories can be accessed by URL /category/<category name>. It displays 10 posts per page by default.

• -f <PATH>: sets path for config file

• -p <PORT>: sets port for server to listen. Default is 3000.

• -c / --create-db: create SQLite database before starting the server. Flag for run command.

Blog-hs currently have several commands:

• check: validates existence of YML files for every MD file in ./templates/ directory (note that its relative path and depends on your working directory).

• create-db: creates (or replaces) SQLite database. Also validates files first.

• run: launches the server.

Configuration file can be passed using -f flag. In other case, server will take important parameters from environment variables or will use default values.

If nececcary, you can disable some pages, like index or category page. In this case, you should use disabledPages option in configuration file. It accepts list of strings, where:

• index is index page option
• category/categories is category page option. Also, it disables link generation on categories on post page.
• robots/robots.txt is robots.txt route.
• sitemap/sitemap.xml is sitemap.xml route.

In other case, you can use ENABLE_INDEX and ENABLE_CATEGORIES environment variables, which accept 1 as True or any other value as False.

Configuration file also contains categories section. Its consists of objects of following scheme:

name: "category-name-used-in-URL"
displayName: "Category name, which will be displayed on website"
description: "Category description"

Currently, there is no way to customize categories using environment.

By default, web server renders at the top of the page post title, date and categories. This behavior can be changed using renderSettings section of server config.

This section has fields renderTitle, renderDate and renderCategories, which are controlling the behavior. If field is not presented, value is True by default.

You can disable rendering OpenGraph and JSON-LD information using field renderMeta.

renderSettings:
  renderDate: false
  renderTitle: true
  renderMeta: false
# renderCategories: true by default

Instead of launching built binary there is way for launch server in Docker container using Docker Compose. If you have make, you make it using:

make build-image
make deploy

For destroying containers, use make destroy command.

This method by default suggests that you will configure server using environment variables. Add volume for config file if its nececcary.

If you want only to build image, use make build-image command.

❗ Dont forget to correct robots.txt file! It's not getting generated automatically.

Code is under BSD 3 Clause license. Please have a look at the LICENSE for details.