1 files changed, 113 insertions, 0 deletions
diff --git a/README.md b/README.md
new file mode 100644
index 0000000..df0525f
--- /dev/null
+++ b/README.md
@@ -0,0 +1,113 @@
+# oreilly-epub
+
+A small CLI tool that downloads a book from **O'Reilly Learning** (given a
+`bookid`) and repackages it into a **valid EPUB**. It mirrors the publisher's
+layout, fixes resource links (images, etc.) so they work offline, and zips
+everything into a ready‑to‑read `.epub`.
+
+> :warning: You must have a valid O'Reilly Learning subscription and your own
+> session cookies. This tool is intended for personal/offline use with content
+> you're authorized to access.
+
+## Features
+
+- **Deterministic by `bookid`**: no fuzzy search—just pass a known identifier.
+- **Authenticated requests** via `cookies.json` (flat key‑value).
+- **Exploded tree download**: mirrors O'Reilly's `full_path` hierarchy under a
+local `epub_root/`.
+- **Valid EPUB packaging**:
+  - Writes the `mimetype` entry first, **uncompressed** (EPUB requirement).
+  - Generates `META-INF/container.xml` pointing to the publisher's OPF.
+  - Zips all assets (HTML, images, CSS, fonts, OPF, NCX, …).
+- **HTML link rewriting during zip**:
+  - Converts absolute O'Reilly API URLs (e.g.,
+`/api/v2/epubs/.../files/images/foo.jpg`) to **relative paths** inside the
+EPUB so images/styles render correctly offline.
+
+## Requirements
+
+- **Rust** (stable, 1.75+ recommended) with Cargo.
+- A **`cookies.json`** file for `learning.oreilly.com` (see below).
+- Network access to O'Reilly's API while running the tool.
+
+## Installation
+
+```bash
+git clone <REPO_URL>
+cd oreilly-epub
+cargo build --release
+```
+
+The binary will be at `target/release/oreilly-epub`.
+
+## Cookies setup (`cookies.json`)
+
+Place a `cookies.json` file in the project root (or pass `--cookies <path>`). \
+The file is a **flat JSON object**: cookie name → cookie value.
+
+**Example:**
+
+```json
+{
+  "orm-session": "REDACTED",
+  "orm-cred": "REDACTED",
+  "another_cookie": "value"
+}
+```
+
+> Tip: You can obtain the cookies for `learning.oreilly.com` from your
+> browser's developer tools by visiting the website and running the command
+> below in the console. Write down the output to `cookies.json` and keep the
+> file private.
+
+```js
+JSON.stringify(document.cookie.split(";").map(c=>c.split("=")).reduce((r,[k,v])=>({...r,[k.trim()]:v?.trim()}),{}))
+```
+
+## Usage
+
+```bash
+# Basic:
+oreilly-epub <bookid>
+
+# With a custom cookies file:
+oreilly-epub <bookid> --cookies /path/to/cookies.json
+```
+
+**Example:**
+
+```bash
+target/release/oreilly-epub 9781787782204 --cookies ./cookies.json
+```
+
+## Notes & Limitations
+
+- This tool assumes the O'Reilly “files” API includes OPF/NCX and all
+referenced assets.
+- HTML rewriting currently uses simple string replacement guided by a URL→local
+mapping; it is **not** a full DOM parse (see Roadmap).
+- File streaming during zip currently buffers each file into memory (OK for
+most books; see Roadmap for planned streaming improvements).
+- Concurrency is not enabled yet; downloads are sequential.
+
+## Roadmap / TODO
+
+- [ ] **CONTRIBUTING.md**: add architecture notes & contributor guidelines.
+- [ ] **Robust HTML rewriting**: replace string replacement with real HTML
+parsing to update `src`, `href`, and other attributes precisely.
+- [ ] **Stylesheets completeness**: ensure all CSS referenced by chapters is
+included and linked properly (cross-check chapters endpoint vs files list).
+- [ ] **License**: add a LICENSE file.
+- [ ] **XDG directories**: use XDG‑compatible defaults for config and the
+download root.
+- [ ] **Concurrency**: implement parallel downloads with a configurable limit.
+- [ ] **Progress reporting**: display per‑file and overall progress (bytes
+and/or file counts).
+- [ ] **Richer metadata**: pull extended book metadata from O'Reilly's metadata
+endpoint and embed into the EPUB.
+- [ ] **XML generation**: build `container.xml` using an XML writer instead of
+raw strings.
+- [ ] **Low‑memory zip**: stream files to the archive in chunks to reduce peak
+memory.
+- [ ] **CI/CD**: add a basic pipeline (build, fmt, clippy, test, release
+artifact).