Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension


Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
28 changes: 28 additions & 0 deletions .github/workflows/docs.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,28 @@
name: build-docs
on:
push:
branches:
- main
permissions:
contents: write
jobs:
deploy:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- name: Configure Git Credentials
run: |
git config user.name github-actions[bot]
git config user.email 41898282+github-actions[bot]@users.noreply.github.com
- uses: actions/setup-python@v5
with:
python-version: 3.x
- run: echo "cache_id=$(date --utc '+%V')" >> $GITHUB_ENV
- uses: actions/cache@v4
with:
key: mkdocs-material-${{ env.cache_id }}
path: .cache
restore-keys: |
mkdocs-material-
- run: pip install mkdocs-material
- run: mkdocs gh-deploy --force
25 changes: 25 additions & 0 deletions CHANGELOG.md
Original file line number Diff line number Diff line change
Expand Up @@ -34,6 +34,31 @@
* Remove safety check that now requires login


## Version 5.3.3

* Fixes bug with svg element dumping. (#48)[https://github.com/thomasborgen/hypermedia/issues/48]

## Version 5.3.2

* Fixes `style` attribute rendering.

## Version 5.3.1

* Adds `lang` as an attribute to the `Html` element.

## Version 5.3.4

### Features

* Add `capture` attribute to input element with the allowed values `"user"` and `"environment"`
* Add `dialog` as a posible value to a forms `method` attribute

### Internal

* Migrate dependency manager to uv.
* Remove safety check that now requires login


## Version 5.3.3

* Fixes bug with svg element dumping. (#48)[https://github.com/thomasborgen/hypermedia/issues/48]
Expand Down
188 changes: 30 additions & 158 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -4,6 +4,15 @@ Hypermedia is a pure python library for working with `HTML`. Hypermedia's killer

Hypermedia is made to work with `FastAPI` and `</> htmx`, but can be used by anything to create HTML.

___
![Python Version from PEP 621 TOML](https://img.shields.io/python/required-version-toml?tomlFilePath=https%3A%2F%2Fraw.githubusercontent.com%2Fthomasborgen%2Fhypermedia%2Fmain%2Fpyproject.toml)
___

**Documentation
[Stable](https://thomasborgen.github.io/hypermedia/) |
[Source Code](https://github.com/thomasborgen/hypermedia) |
[Task Tracker](https://github.com/thomasborgen/hypermedia/issues)**

## Features

* Build __HTML__ with python classes
Expand Down Expand Up @@ -44,20 +53,22 @@ use `.dump()` to dump your code to html.
from hypermedia import Bold, Div

Div("Hello ", Bold("world!")).dump()
```

output

# outputs
# '<div>Hello <b>world!</b></div>'
```html
<div>Hello <b>world!</b></div>
```

## Composability with slots

```python
from hypermedia import Html, Body, Div, Menu, Header, Div, Ul, Li
from hypermedia import Html, Body, Div, Menu, H1, Div, Ul, Li

base = Html(
Body(
Menu(slot="menu"),
Header("my header", slot="header"),
Div(slot="content"),
),
)
Expand All @@ -69,162 +80,23 @@ base.extend("menu", menu)
base.extend("content", content)

base.dump()

# outputs
# '<html><body><menu><ul><li>main</li></ul></menu><header>my header</header><div><div>Some content</div></div></body></html>'
```

## Attribute names with special characters

Most `html` and `</>htmx` attributes are typed and has Aliases where needed. That means that most of the time you won't have to think about this and it should _just work_.

The attribute name output rules are:

1. Any attribute that does not have an Alias will have any underscores (`_`) changed to hyphens (`-`).
2. Any attribute that is prefixed with `$` will be outputted as is without the first `$`.

ie

```python
# Is a specified attribute(typed) with an Alias:
Div(on_afterprint="test") # <div onafterprint='test'></div>

# Unspecified attribute without Alias:
Div(data_test="test") # <div data-test='test'></div>

# Spread without $ prefix gets its underscores changed to hyphens:
Div(**{"funky-format_test.value": True}) # <div funky-format-test.value></div>

# Spread with $ prefix
Div(**{"$funky-format_test.value": True}) # <div funky-format_test.value></div>
Div(**{"$funky-format_test.value": "name"}) # <div funky-format_test.value='name'></div>
```

Note: About the </> HTMX attributes. [The documentation](https://htmx.org/attributes/hx-on/) specifies that all hx attributes can be written with all dashes. Because of that Hypermedia lets users write hx attributes with underscores and Hypermedia changes them to dashes for you.

```python

Div(hx_on_click='alert("Making a request!")')
# <div hx-on-click='alert("Making a request!")'></div>
# Which is equivalent to:
# <div hx-on:click='alert("Making a request!"'></div>

Div(hx_on_htmx_before_request='alert("Making a request!")')
# <div hx-on-htmx-before-request='alert("Making a request!")'></div>

# shorthand version of above statement
Div(hx_on__before_request='alert("Making a request!")')
# <div hx-on--before-request='alert("Making a request!")'></div>
```

# HTMX

## The Concept

The core concept of HTMX is that the server responds with HTML, and that we can choose with a CSS selector which part of the page will be updated with the HTML response from the server.

This means that we want to return snippets of HTML, or `partials`, as they are also called.

## The Problem

The problem is that we need to differentiate if it's HTMX that called an endpoint for a `partial`, or if the user just navigated to it and needs the `whole page` back in the response.

## The Solution

HTMX provides an `HX-Request` header that is always true. We can check for this header to know if it's an HTMX request or not.

We've chosen to implement that check in a `@htmx` decorator. The decorator expects `partial` and optionally `full` arguments in the endpoint definition. These must be resolved by FastAPI's dependency injection system.

```python
from hypermedia.fastapi import htmx, full
```

The `partial` argument is a function that returns the partial HTML.
The `full` argument is a function that needs to return the whole HTML, for example on first navigation or a refresh.

> Note: The `full` argument needs to be wrapped in `Depends` so that the full function's dependencies are resolved! Hypermedia ships a `full` wrapper, which is basically just making the function lazily loaded. The `full` wrapper _must_ be used, and the `@htmx` decorator will call the lazily wrapped function to get the full HTML page when needed.

> Note: The following code is in FastAPI, but could have been anything. As long as you check for HX-Request and return partial/full depending on if it exists or not.

```python
def render_base():
"""Return base HTML, used by all full renderers."""
return ElementList(Doctype(), Body(slot="body"))


def render_fruits_partial():
"""Return partial HTML."""
return Div(Ul(Li("Apple"), Li("Banana"), Button("reload", hx_get="/fruits")))


def render_fruits():
"""Return base HTML extended with `render_fruits_partial`."""
return render_base().extend("body", render_fruits_partial())


@router.get("/fruits", response_class=HTMLResponse)
@htmx
async def fruits(
request: Request,
partial: Element = Depends(render_fruits_partial),
full: Element = Depends(full(render_fruits)),
) -> None:
"""Return the fruits page, partial or full."""
pass
```

That's it. Now we have separated the rendering from the endpoint definition and handled returning partials and full pages when needed. Doing a full refresh will render the whole page. Clicking the button will make a htmx request and only return the partial.

What is so cool about this is that it works so well with FastAPI's dependency injection.

## Really making use of dependency injection


```python
fruits = {1: "apple", 2: "orange"}

def get_fruit(fruit_id: int = Path(...)) -> str:
"""Get fruit ID from path and return the fruit."""
return fruits[fruit_id]

def render_fruit_partial(
fruit: str = Depends(get_fruit),
) -> Element:
"""Return partial HTML."""
return Div(fruit)

def render_fruit(
partial: Element = Depends(render_fruit_partial),
):
return render_base().extend("content", partial)

@router.get("/fruits/{fruit_id}", response_class=HTMLResponse)
@htmx
async def fruit(
request: Request,
partial: Element = Depends(render_fruit_partial),
full: Element = Depends(full(render_fruit)),
) -> None:
"""Return the fruit page, partial or full."""
pass
output

```html
<html>
<body>
<menu>
<ul><li>main</li></ul>
</menu>
<div>
<div>Some content</div>
</div>
</body>
</html>
```

Here we do basically the same as the previous example, except that we make use of FastAPI's great dependency injection system. Notice the path of our endpoint has `fruit_id`. This is not used in the definition. However, if we look at our partial renderer, it depends on `get_fruit`, which is a function that uses FastAPI's `Path resolver`. The DI then resolves (basically calls) the fruit function, passes the result into our partial function, and we can use it as a value!

__This pattern with DI, Partials, and full renderers is what makes using FastAPI with HTMX worth it.__

In addition to this, one thing many are concerned about with HTMX is that since we serve HTML, there will be no way for another app/consumer to get a fruit in JSON. But the solution is simple:

Because we already have a dependency that retrieves the fruit, we just need to add a new endpoint:

```python
@router.get("/api/fruit/{fruit_id}")
async def fruit(
request: Request,
fruit: str = Depends(get_fruit),
) -> str:
"""Return the fruit data."""
return fruit
```
# Documentation

Notice we added `/api/` and just used DI to resolve the fruit and just returned it. Nice!
Head over to our [documentation](https://thomasborgen.github.io/hypermedia)
Empty file added docs/advanced/caching.md
Empty file.
Empty file added docs/advanced/custom_models.md
Empty file.
18 changes: 18 additions & 0 deletions docs/contributing.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,18 @@
# Contributing

The best thing that can be contributed with is bug reports and documentation.

To help with documentation fork the repo, and run the mkdocs site with:

```sh
uv sync

uv run mkdocs serve
```

!!! tip

Add `--livereload` to enable autorefresh

This starts up a server at [localhost:8000](http://localhost:8000) navigate there and start adding or editing the doc files in `/docs`

Loading
Loading