assets-only root README

README.md

@@ -1,5 +1,5 @@
 ---
-title: API Py
+title: Asset Extractor API
 emoji: 🐍
 colorFrom: blue
 colorTo: green
@@ -8,17 +8,16 @@ pinned: false
 app_port: 7860
 ---
 
-# API Py
+# Asset Extractor API
 
-Monolith-modular Python (FastAPI) API on Hugging Face Spaces.
-Utility endpoints for scraping, URL mapping, favicons — and **asset extraction**.
+FastAPI service that extracts assets from any URL, grouped by type:
+**image · video · audio · svg/icon · font**.
 
-## Endpoints
+## Usage
 
-### `POST /assets` — extract assets from a URL (new 🎯)
+### `POST /assets`
 
-Give it a URL, get back the page's assets grouped by type:
-**image · video · audio · svg_icon · font**.
+Give it a URL, get back the page's assets grouped by category.
 
 ```bash
 curl -X POST https://lljz66-api.hf.space/assets \
@@ -41,7 +40,7 @@ curl -X POST https://lljz66-api.hf.space/assets \
 }
 ```
 
-Optional request fields:
+### Request fields
 
 | field | type | default | meaning |
 |-------|------|---------|---------|
@@ -49,46 +48,30 @@ Optional request fields:
 | `include_data_uris` | bool | `false` | keep `data:` URIs |
 | `groups` | list[string] | all | subset: `image`, `video`, `audio`, `svg_icon`, `font` |
 
-> Static HTML only — JS-rendered assets are not captured.
-
-### `POST /scrape`
-Scrape a website's clean content via Firecrawl.
-
-```bash
-curl -X POST https://lljz66-api.hf.space/scrape \
-  -H 'Content-Type: application/json' \
-  -d '{"url": "https://example.com"}'
-```
-
-### `GET /map`
-Discover a website's URLs via sitemap + katana crawler.
+### Response fields
 
-```bash
-curl "https://lljz66-api.hf.space/map?url=https://example.com&limit=100"
-```
-
-### `GET /favicon`
-Get the favicon URL for a website via Google s2.
-
-```bash
-curl "https://lljz66-api.hf.space/favicon?url=https://example.com"
-```
+| field | type | meaning |
+|-------|------|---------|
+| `url` | string | final URL after redirects |
+| `title` | string \| null | page title |
+| `total` | int | total asset count |
+| `assets` | object | grouped asset URL lists |
 
-### `GET /health`
-Service status.
-
-### `GET /debug/katana`
-Diagnose katana crawler flags.
+> Static HTML only — JS-rendered assets are not captured.
 
 ## Structure (monolith-modular)
 
 ```
 app/
 ├── main.py                 # FastAPI app assembly
-├── config.py               # settings
+├── config.py               # settings (fetch limits, SSRF block-list)
 ├── schemas.py              # pydantic request/response models
-├── routers/                # HTTP layer (health, scrape, map, favicon, assets, debug)
-└── services/               # business logic (+ asset extractors/)
+├── routers/
+│   └── assets.py           # POST /assets (HTTP layer)
+└── services/
+    ├── fetcher.py          # safe async page fetch (SSRF/size/timeout guards)
+    ├── resolver.py         # absolutize asset references
+    ├── asset_service.py    # orchestrator: fetch -> parse -> extract -> group
     └── extractors/         # images, videos, audios, icons_svg, fonts
 ```