backpocket

README

Backpocket

Backpocket is a command line utility for storing a reading list of articles from the Internet to your local disk for the future.

It's an alternative to Pocket offering all the required features without handing over all your private reading materials to any 3rd party. And it's totally free, offering even full-text search via other tools.

Backpocket is based on the algorithm created by Mozilla used to create Firefox Readability.js functionality.

Features

Backpocket aims to be as simple as possible while offering a lot of features:

• command-line tool for very easy usage;
• stores articles on a local disk (see an example);
• images are stored as base64 data sources;
• supports all kind of formats - when not html, then will be stored AS IS;
• stored articles can be opened just with a browser or other tools depending on the format;
• since articles are stored in their native format, then can do whatever with them - perform full-text search, convert to text etc;
• articles which fail to be stored for some reason will still save a reference to them storing the original link (see an example).

Installation

Download latest binary from releases, extract it and add it to somewhere in your PATH. That's it.

Of course, you're free to compile your own version of binary to be 100% sure that it has not been tampered with, since this is an open-source project after all.

Usage

Using backpocket is simple too:

$ backpocket 'ARTICLE_URL'

To open archived article immediately from command-line:

$ open `backpocket 'ARTICLE_URL'`

To open any archived article later:

$ open examples/successful.html

Running backpocket will try to create a readable version of the article specified at ARTICLE_URL and stores it on a local disk. When readable version cannot be created (for example, ARTICLE_URL points to an image) then that URL will be downloaded and stored AS IS. You can configure storage dir by editing backpocket configuration file config.json, which is stored in a location specified by XDG Base Directory standard).

What about full text search?!

Easy, use ripgrep for that!

$ rg -ni 'trump' $(backpocket path)

For better search, convert all html's to text first and search over these! See below.

Aliases for reading, archiving and searching

Add all the following aliases/functions to your .bashrc, .zshrc or to another scripts which are executed on login.

• First create a function for the backpocket itself, which creates in addition to .html a .txt version of each article for faster searching (requires html2text to be installed):

function bp() { ARTICLE_PATH=$(backpocket $1) && echo $ARTICLE_PATH && html2text -style pretty -o $ARTICLE_PATH.txt $ARTICLE_PATH }

• Easiest way to read oldest article would be to create an alias:

alias bp-read='open `ls -Adp $(backpocket path)/* | grep -v "/$" | head -1`'

• Also function for archival makes sense:

function bp-archive() { mkdir -p `backpocket path`/archive && ( if [[ -z $1 ]]; then ARTICLE_PATH=`ls $(backpocket path)/* | grep -v "/$" | head -1` && mv -v "$ARTICLE_PATH" `backpocket path`/archive && test -f "$ARTICLE_PATH.txt" && mv "$ARTICLE_PATH.txt" `backpocket path`/archive; else mv -v $1 `backpocket path`/archive && test -f $1.txt && mv $1.txt `backpocket path`/archive; fi ) }

And a function for search:

function bp-search() { rg -ni "$1" `backpocket path`/**/*.txt }

And now just use these commands:

$ bp 'ARTICLE_URL'
$ bp-read
$ bp-archive
$ bp-search 'TERM'

Importing articles from Pocket

It is pretty easy and straightforward to import articles from Pocket too:

1. Install backpocket
2. Install pup and jq;
3. Export from Pocket to HTML file;
4. Run the following command to import all archived articles where ril_export.html is the file Pocket exported:

cat ril_export.html | pup 'ul:last-of-type a json{}' | jq -r '.[] | "\(.href) \(.time_added)"' | tail -r | xargs -P4 -L1 backpocket
mkdir -p `backpocket path`/archive && mv $(ls -Adp `backpocket path`/* | grep -v "/$") `backpocket path`/archive

1. Run the following command to import all non-archived articles from Pocket:

cat ril_export.html | pup 'ul:first-of-type a json{}' | jq -r '.[] | "\(.href) \(.time_added)"' | tail -r | xargs -P4 -L1 backpocket

1. (Optional) Run the following command to create .txt versions for each .html article for better searching functionality (see Alias for searching) using html2text:

find $(backpocket path) -name "*.html" | xargs -I % html2text -o %.txt -style pretty %

This all might take some time depending on the count of articles, size of articles, speed of your internet connection and so on. Also, please note that there's a high probability that many of the articles saved to the Pocket in the past do not exist in the Internet anymore and will be saved as a failed articles - that's the problem with Internet and exactly the reason why archiving interesting/important things on your local disk makes sense in case you ever want to return to any of that content.

Backup, syncing, mobile support etc.

Backpocket does not offer anything else except storing articles for future use. Use any 3rd party tools for backup, syncing or showing articles on mobile (for example, serve articles using a static web server).

Windows support

Since backpocket is written using Go and it has a wonderful cross-compile support then it also works without any problems on Windows. However, if you need to use any UNIX tools described in this README then use WSL or some other Linux subsystem on Windows.

There is one special trick under WSL Ubuntu so that function bp-read defined above would work correctly since path conversion from WSL to Windows needs to be performed:

function bp-read() { if [[ -z $1 ]]; then cmd.exe /c start $(wslpath -aw `ls -Adp $(backpocket path)/* | grep -v "/$" | head -1`); else cmd.exe /c start $(wslpath -aw $1); fi; }