Metadata-Version: 2.2
Name: downie
Version: 1.0.0
Summary: A fast async download parser with concurrency control. Written for specific use case to crawl a certain site. Will only download urls under the base url. Remote links are not followed.
Author-email: retoor <retoor@molodetz.nl>
Requires-Python: >=3.7
Description-Content-Type: text/markdown
Requires-Dist: aiohttp
Requires-Dist: beautifulsoup4
Requires-Dist: aiofiles
Requires-Dist: app@ git+https://retoor.molodetz.nl/retoor/app.git
Provides-Extra: autocomplete
Requires-Dist: argcomplete>=1.12.3; extra == "autocomplete"

# Downie

**Downie** is a full **site downloader**. I tried several downloaders and they all were not capable to download big sites like **molodetz.nl** (containing **300.000+** links / resources). See **features** to see what makes this site downloader **better than most competition**.

If you encounter a site that can't be downloaded using this downloader, contact me @ *retoor@molodetz.nl*. Also, creating a bug ticket is possible after logging in. **I respond to both methods within 24h**.

## Features:
 - Fast as light **async** fetching of your pages allowing concurrency.
 - Limit fetching only to given domain. **Remote URLs are not followed**. Also redirects to remote URLs (yes, that can be an issue) are not followed.
 - Discriminates between file and folder for urls by adding `.html` to pages not having an extension. So `/pony` becomes /pony.html so that `/pony/photos.html` is possible.
 - **Absolute URLs will be converted to relative URLs** so that the site becomes **portable** and **will work isolated** on your webserver.

## Using Downie
After following the **installation instructions**, use `downie https://target-site.com` if **globally installed**. If not globally installed, execute `./downie https://target-site.com` from the current directory.

## Installation

### Debian (for development)
Only **requirement** is `make`. Install by executing `sudo apt install make -y` in the terminal.
Installation of all required Python dependencies are done by executing `make install`.
This will **install system packages** using `apt`:
 - `python`
 - `python3-venv`
 - `python3-pip`
 *There is a big chance you already have them.*
 After that:
 1. A Python environment will be created named `.venv`. *You can ignore this folder.*
 2. Required Python libraries will be installed. *(aiohttp, aiofiles, beautifullsoup4, app (molodetz)).*
 3. Downie will be installed. *(For the devs: in edit mode with -e!).*
 4. A executable file named `downie` will be placed in the current directory. This is the main application.
**Tip:** make downie globally accessable by copying `downie` to `/usr/local/bin`.

### Other operating systems; Windows / Mac; not for development.
You have to figure out how to install python3, python3-venv and python3-pip on your OS. Make sure you have these on your system. Chance is big that you already got those.
Run the following commands:
```bash
python3 -m venv .venv 
source .venv/bin/activate 
pip install git+https://retoor.molodetz.nl/retoor/downie.git
cp .venv/bin/downie .
```
**Tip:** make downie globally accessable by copying `downie` to `/usr/local/bin`.

## Default configuration / internals
Not defined by CLI parameters, but configurable as Downie class constructor parameters:
 - `concurrent_request_count` - Amount of concurrent requests. This defaults to 500. 
 - `concurrent_write_count` - Amount of concurrent downloads of files. This defaults to 10.
These values are based on my usage. There is a possibility that there are more optimial parameters.

## Running a downloaded site:

### Open site statically
Just open a page (.html) from the site using your browser.

### Serve the site
Simple, run:
```bash
python3 -m http.server [port]
```
If you don't know what port to use, use `7331`. Your website is accessable on `https://localhost:7331` now.

## Caution
URLs of downloaded content can be modified to make the site portable. So it is possible that your downloaded site is literally the content online. It is optimized. A literal copy of the site could be unusable.
