rural-dict/src/main.py

import re
from contextlib import asynccontextmanager
from datetime import datetime
from json import JSONDecodeError

import aiohttp
from fastapi import FastAPI, Request
from fastapi.responses import HTMLResponse, RedirectResponse
from fastapi.staticfiles import StaticFiles
from fastapi.templating import Jinja2Templates
from selectolax.parser import HTMLParser, Node


@asynccontextmanager
async def lifespan(app: FastAPI):
    """Establishing an aiohttp ClientSession for the duration of the app's lifecycle."""
    global session
    session = aiohttp.ClientSession(timeout=aiohttp.ClientTimeout(10))
    yield
    await session.close()


app = FastAPI(lifespan=lifespan, docs_url=None, redoc_url=None)
app.mount("/static", StaticFiles(directory="static"), name="static")
templates = Jinja2Templates(directory="templates")
session: aiohttp.ClientSession = None  # pyright: ignore[reportAssignmentType]


def remove_classes(node: Node) -> Node:
    """Recursively remove all classes from all nodes."""
    if "class" in node.attributes:
        del node.attrs["class"]  # pyright: ignore [reportIndexIssue]
    for child in node.iter():
        remove_classes(child)
    return node


@app.get("/{path:path}", response_class=HTMLResponse)
async def catch_all(response: Request):
    """Handle all routes on Urban Dictionary and perform redirection if necessary."""
    path_without_host = (
        f"{response.url.path}{f'?{response.url.query}' if response.url.query else ''}"
    )
    url = f"https://www.urbandictionary.com{path_without_host}"
    term = response.query_params.get("term")

    async with session.get(url) as dict_response:
        if dict_response.history:
            return RedirectResponse(str(dict_response.url.relative()), status_code=301)
        html = await dict_response.text()
        parser = HTMLParser(html)
        if dict_response.status != 200:
            similar_words = None
            if (try_this := parser.css_first("div.try-these")) is not None:
                similar_words = [remove_classes(word).html for word in try_this.css("li a")]
            return templates.TemplateResponse(
                "404.html",
                {
                    "request": response,
                    "similar_words": similar_words,
                    "term": term,
                    "site_title": f"Rural Dictionary: {term}",
                    "site_description": (
                        "View on Rural Dictionary, an alternative private "
                        "frontend to Urban Dictionary."
                    ),
                },
                status_code=404,
            )

    results = []
    definitions = parser.css("div[data-defid]")
    try:
        thumbs_api_url = (
            f'https://api.urbandictionary.com/v0/uncacheable?ids='
            f'{",".join(d.attributes["data-defid"] or "-1" for d in definitions)}'
        )
        async with session.get(thumbs_api_url) as thumbs_response:
            thumbs_json = await thumbs_response.json()
            thumbs_data = {el["defid"]: el for el in thumbs_json["thumbs"]}
    except (KeyError, JSONDecodeError, TimeoutError):
        thumbs_data = {}

    site_description = None
    for definition in definitions:
        word = definition.css_first("a.word").text()
        meaning_node = remove_classes(definition.css_first("div.meaning"))
        if site_description is None:
            site_description = re.sub(r"\s+", " ", meaning_node.text(strip=True, separator=" "))
        meaning = meaning_node.html
        example = remove_classes(definition.css_first("div.example")).html
        contributor = remove_classes(definition.css_first("div.contributor")).html
        definition_id = int(definition.attributes["data-defid"] or "-1")
        definition_thumbs = thumbs_data.get(definition_id, {})
        thumbs_up = definition_thumbs.get("up")
        thumbs_down = definition_thumbs.get("down")
        results.append(
            [definition_id, word, meaning, example, contributor, thumbs_up, thumbs_down]
        )
    if (pagination := parser.css_first("div.pagination")) is not None:
        pagination = remove_classes(pagination)
        pagination.attrs["class"] = "pagination"  # pyright: ignore [reportIndexIssue]
        pagination = pagination.html

    term = term or results[0][1]
    site_title = "Rural Dictionary"
    match response.url.path:
        case "/":
            # add current date for page with words of the day
            site_title += f', {datetime.now().strftime("%d %B")}'
        case "/random.php":
            term = "Random words"
    site_title += f": {term}"

    return templates.TemplateResponse(
        "index.html",
        {
            "request": response,
            "results": results,
            "pagination": pagination,
            "term": term,
            "site_title": site_title,
            "site_description": site_description,
        },
    )