A blogging platform written in Rust for Cloudflare Workers, integrated with Standard Notes
Find a file
2020-04-14 16:52:31 +08:00
common implement RSS feed 2020-04-12 14:28:26 +08:00
src add support for hard-coded extra proxied resources 2020-04-13 18:52:16 +08:00
theme/default theme: animate toc item "current" state too 2020-04-14 16:52:31 +08:00
.cargo-ok initial commit 2020-04-07 16:10:01 +08:00
.gitignore add template and implement rendering 2020-04-10 20:38:53 +08:00
build.rs build.rs: re-run after any file inside directories have changed 2020-04-14 16:00:07 +08:00
Cargo.lock drop rbtag and use our own hand-rolled implementation 2020-04-12 17:25:55 +08:00
Cargo.toml drop rbtag and use our own hand-rolled implementation 2020-04-12 17:25:55 +08:00
package-lock.json integrate hljs into backend to provide code highlighting 2020-04-12 16:53:51 +08:00
package.json integrate hljs into backend to provide code highlighting 2020-04-12 16:53:51 +08:00
README.md add support for hard-coded extra proxied resources 2020-04-13 18:52:16 +08:00
webpack.config.js project restruct: use webpack instead of wrangler's wasm-pack support 2020-04-09 12:40:15 +08:00
worker.js project restruct: use webpack instead of wrangler's wasm-pack support 2020-04-09 12:40:15 +08:00

Paprika

Yet another blog system that runs on Cloudflare Workers, and integrates with Standard Notes (a self-hosted note-taking software) for a great editing experience, while giving you more freedom than the Listed service provided by Standard Notes. Publish directly from your Standard Notes notebook with Paprika!

This requires Cloudflare Workers KV for storage and thus needs the paid Unlimited plan to work. However, it is possible to swap out the storage, if you would like to fiddle with the code, and use something like S3 to use it 100% free on Workers (barring S3 costs).

As a practice, Paprika was written in Rust and compiled to WebAssembly for execution on Cloudflare Workers, using wasm-bindgen to interact with the JS environment. One single JS dependency, highlight.js, was used because there's simply no good alternative from the Rust ecosystem. webpack was used for an automatic, cached loading experience of the WebAssembly module (the official template for WebAssembly by Cloudflare is terrible because it tries to re-instantiate the module every time a request comes in; using webpack fixed the issue because it's much smarter), along with the ability to load highlight.js modularly.

WARNING: I haven't even converted my own blog to Paprika yet.

WARNING: This project is neither complete nor tested. Use at your own risk. Always keep backups.

Prerequisites

  1. The wrangler cli tool from Cloudflare
  2. node.js + npm
  3. rustc + cargo, with the latest nightly toolchain
  4. wasm-pack and wasm-opt

Deployment

  1. Complete all configuration files according to the sections below (otherwise the project won't build)
  2. Run npm install --development to install webpack and other node build dependencies.
  3. Run wrangler publish to upload to Cloudflare Workers
  4. Set up correct routes in Cloudflare control panel
  5. Add your own instance of Paprika to your Standard Notes as a plugin (instructions available below)
  6. Publish!

Note that all configuration and themes will be included statically in the final binary. To modify any of them, you will need to re-run wrangler publish to rebuild the entire program.

Configuration: wrangler.toml

The wrangler.toml is what instructs the wrangler tool to configure Cloudflare Workers correctly. An example for Paprika is listed below:

name = "paprika"
type = "webpack"
webpack_config = "webpack.config.js"
account_id = "<account_id>"
workers_dev = true
route = ""
zone_id = "<zone_id>"

kv-namespaces = [ 
         { binding = "PAPRIKA", id = "<kv_namespace_id>" } 
]

You need to replace everything within <>. The KV namespace can be created manually or via wrangler, but it must be binded in wrangler.toml with the name PAPRIKA like shown in the above example. Using any other name will not work.

Configuration: config.json

This is the main configuration file. The file will be compiled statically into the executable, and will be read in build.rs. Make sure you have the format correct and everything declared.

{
  "secret": "<generate_some_random_secret>",
  "theme": "default",
  "title": "<title>",
  "lang": "en",
  "description": "<description>",
  "plugin_identifier": "com.example.change.this.to.whatever.you.like",
  "posts_per_page": 5,
  "cache_maxage": 86400,
  "redirects": {
    "/foo": "/bar",
    ...
  },
  "extra_remote_proxy_whitelist": [
    "<url>",
    ...
  ],
  "hljs": [
    "rust",
    "javascript",
    "bash",
    ...
  ]
}

secret: This will be the sole credential you can use to access the publishing endpoints (i.e. Standard Notes plugin endpoints). Generate something via openssl rand or whatever you think is secure.

theme: The name of the theme to use for your blog. Must be a subdirectory in ./theme, and the default one is default. The selected theme will be compiled statically into the final .wasm binary. For more information on themes, continue reading this documentation.

plugin_identifier: Used in Standard Notes to distinguish plugins.

redirects: OPTIONAL. A map of URLs where the key will be mapped to the value by Paprika using 301 redirects. This is mainly useful for migration from another blogging platform.

cache_maxage: OPTIONAL. A value in seconds determining how long the browser should cache static resources from the blog. If omitted, the default value is a week.

extra_remote_proxy_whitelist: OPTIONAL. See the Remote Resource Proxy section below for details.

hljs: An array of language support from highlight.js to be included in the final binary. The full highlight.js is notoriously huge and there's really no reason to include a bazillion languages you will never actually use in your blog posts. This will be read by build.rs to generate a JS shim that will load all languages in the array to the final binary via webpack support for require.

Configuration: theme_config.json

theme_config.json will be passed to Handlebars templates in the theme as blog.theme_config (See BlogRootContext and other Context structures in src/render.rs; serde_json::Value means arbitrary JSON object). A theme can thus use arbitrary extra information available via this configuration file. The default theme currently supports the following options:

{
  "avatar_url": "<url_of_your_avatar>",
  "nav_links": [
    {
      "name": "<nav_name>",
      "url": "<nav_url>",
      "target": "_blank"
    },
    ...
  ]
}

nav_links: a set of navigation links to be displayed in the sidebar (or at the top on mobile). You can set "target": "_blank" to make the link open in new tabs, while omitting this attribute will make the link behave as a normal link, that is, open in the current page.

Installation in Standard Notes

After your blog is up and running, you can import the plugin to your Standard Notes account using the following URL:

https://<your_domain.com>/actions?secret=<your_secret>

The secret should be replaced with what you generated when creating config.json.

After the plugin is imported successfully, you can begin posting any of your notes from teh Actions menu of Standard Notes.

Post Format

You can use all valid Markdown syntax and some GitHub Flavoured Markdown supported by the pulldown_cmark crate. Read their documentation for what is actually supported.

In the post, you can insert <!-- More --> as a standalone paragraph to indicate that everything before this marker should be considered the summary, and should be displayed in place of the full text when viewing from the home page (post list). This does not affect the single post page and will not be displayed whatsoever.

By default, the timestamp and the URL of any new post will be generated automatically. You can override this behavior by inserting a fenced JSON code block at the very beginning of the post, followed by an empty line:

```json
{
    "url": "some-awesome-url",
    "timestamp": "YYYY-mm-dd",
    "unlist": true
}
```

url: Customize the URL of the post. Only the pathname part, and should not include the starting /.

timestamp: Customize the displayed date of the post. This does not affect the order of posts on home page -- posts that were created later always take precedence, regardless of their timestamp. This is mainly useful when migrating old articles.

unlist / unlisted: when set to true, the post won't appear in home page, while still being accessible via its URL.

Normally, if such a customization header is not present, a post's metadata (URL and timestamp) will not be updated when you update a post. However, when this header is present, then the metadata will always be updated.

When a post's url is changed, the old one will become an alias, 302-redirected to the new one.

Themes

A theme is a set of Handlebars templates and static assets located in a direct subdirectory of the theme folder. The structure should be as follows:

- theme/
  - <theme_name>/
    - static/
      - ....
    - home.hbs
    - post.hbs
    - ...

Everything inside the static folder will be available at https://<your_domain>/static/. These resources are subject to browser caching, so to ensure the resources are always re-loaded by the browser when Paprika or the theme updates, you should add a query string to links to your static resources using the build_num helper available via Handlebars, e.g. ?ver={{ build_num }} (see the default theme for details). This ensures that each time Paprika gets rebuilt, the links will be different due to the build_num and all clients should fetch the updated resources.

home.hbs will be used to render the home page (post list), while post.hbs will be used for single-post pages (i.e. the detail page). These templates can import other templates located in the same directory via the {{> some_other_template.hbs }} syntax.

The execution context of each template is defined in src/render.rs, as those *Context structs. Extra helpers are also defined in that file with the handlebars_helper! macros. Code there is pretty self-explanatory, please refer to the structs and the default theme for details on how to use the execution contexts.

The theme directory selected via config.json will be included into the final binary. Therefore, please make sure your assets are not too huge to fit in the 1MB binary limit of Cloudflare Worker.

Remote Resource Proxy

Paprika replaces all external images inserted into your posts by Markdown with a proxied version hosted on the same URL of your blog under the /imgcache/ path. This ensures that the source websites cannot see your visitors' IP addresses and that the Cloudflare CDN policy can be applied to them to ensure faster loading time.

The cached URL is formatted like below:

https://<your_domain>/imgcache/<origin_url_urlencoded>

where origin_url_urlencoded is the URL-encoded version (as per JavaScript encodeURIComponent function) of the URL to the original resource. A whitelist of origin URLs is maintained in Workers KV so that this URL cannot be used on arbitrary content -- only those that are present in your published posts will be reverse-proxied. The whitelist is updated every time a post is re-rendered -- that is, when you create / update a post or update the Paprika program or its other resources.

You can hard-code more whitelisted URLs (non-URL-encoded version) in the extra_remote_proxy_whitelist array of config.json. This may be useful if your theme supports things like avatars, where custom URLs need to be provided (which is the case with the default theme).

The reverse-proxy only forwards the Content-Type header and the actual body of the response (of course, after the body is decoded properly and cached by Cloudflare's Fetch API). It also follows 30x redirects by default. Other fields will be re-calculated by the runtime before returning to the client.

FAQs

  • Whats the point of using Cloudflare Workers when you need to self-host a Standard Notes instance anyway?

    Well, mainly, for me, I just wanted to try something new. Since blogs are not exactly private data, I am pretty comfortable hosting it anywhere -- and because I like Standard Notes, I wanted to have something that integrates directly with it just like the official Listed service, but with more fine-grained control over both how it looks and how it behaves. Also, the blog is something that I'd like to always keep online, even during maintanence of my server, or even after hardware failures, and, well, at least Workers can be online more than my server can do. In addition, Standard Notes does not necessarily need to be self-hosted, since the protocol basically makes the server blind of most of information about your notes. In that case, with Paprika, you can basically have a maintanence-free blogging experience right out of your own notebook.