bobolink

README

Bobolink

Bobolink is a small tool that allows you to save links and search for them later. More specifically, Bobolink provides full text search on the body of HTML documents that you've added to your index. Bobolink is written in go. The frontend to the web service is written with go html templates and vanilla js

Installation

go get -u github.com/jtanza/bobolink

Bobolink requires an index to manage stored text. The path to this index can either be passed explicitly with each invocation of bobolink e.g.--index-path /opt/bobolink or can be set once via the env variable BOBOLINK_DIR=/opt/bobolink

This directory can of course be anywhere you'd like, but I personally use /opt/bobolink.

Usage

Usage:
  bobolink [flags]
  bobolink [command]

Available Commands:
  add         Adds urls to store
  find        Searches for urls
  help        Help about any command
  remove      Deletes urls from store
  server      Starts the bobolink web server

Flags:
  -h, --help                help for bobolink
  -i, --index-path string   path to bobolink index

Use "bobolink [command] --help" for more information about a command.

Some Examples

$ # adding new links is easy
$ bobolink --index-path /opt/bobolink add https://web.stanford.edu/class/cs101/bits-gigabytes.html
$ successfully added 1 resource\s to index.

$ # search your saved bookmarks. matching text is highlighted in the terminal (can't be seen here)
$ bobolink --index-path /opt/bobolink find tera*

URL: https://web.stanford.edu/class/cs101/bits-gigabytes.html
Match: …ytes Gigabytes Terabytes Kilobytes Megabytes Gigabytes Terabytes The size of information in the computer is measured in kilobytes, megabytes, gigabytes, and terabytes. In this section, we'll look at c…

$ # set env var to avoid having to pass --index-path with each invocation
$ export BOBOLINK_DIR=/opt/bobolink

$ # backup all your bookmarked urls
$ bobolink add https://en.wikipedia.org/wiki/Protection_ring
$ bobolink find --all >> bookmarks.txt && cat bookmarks.txt
https://en.wikipedia.org/wiki/Protection_ring
https://web.stanford.edu/class/cs101/bits-gigabytes.html

$ bobolink remove https://web.stanford.edu/class/cs101/bits-gigabytes.html
$ bobolink find --all
https://en.wikipedia.org/wiki/Protection_ring

Web App

If users do not wish to use bobolink via the CLI or would like to offload the requisite disk space needed for the document index to a remote server, a web interface is also provided. One can simply run:

user@remote $ bobolink server -p 25000
Listening on port :2500...
user@local $ ssh -L 8080:localhost:25000 user@remote -N

and access the web app in their browser.

Please note that because the static files used on the frontend are not packaged with the bobolink executable, in order to run the web app it is necessary to install bobolink manually, i.e. clone the repo into $GOPATH/src and go install.

Web App Screenshots

API

All the commands offered in the CLI are likewise offered over HTTP, "Restfully":

$ bobolink find hash
URL: https://xlinux.nist.gov/dads/HTML/secondaryClustering.html
Match: …to create long run of filled slots away from a key hash position, e.g., along the probe sequence. 
See also primary clustering, clustering free, hash table, open addressing, clustering, linear probing,…

$ curl -d '{"query": "hash"}' localhost:8080/links/find
[{"Body":"...to create long run of filled slots away from a key hash position, e.g., along the probe sequence. 
See also primary clustering, clustering free, hash table, open addressing, clustering, linear probing...,",
"URL":"https://xlinux.nist.gov/dads/HTML/secondaryClustering.html"}]

All endpoints return data as JSON. However, if HTML is desired, once can simply request it.

$ curl -H "Accept: text/html" -d '{"query": "hash"}' localhost:8080/links/find
<ul>
  <li><a href="https://xlinux.nist.gov/dads/HTML/secondaryClustering.html">xlinux.nist.gov</a></li>
  <li>…to create long run of filled slots away from a key <mark>hash</mark>; position, e.g., along the probe sequence. See also primary clustering, clustering free, <mark>hash</mark>; table, open addressing, clustering, linear probing,…</li>
</ul>

Motivation

I save lots of links without the use a bookmark manager. When I find a new article I'd like to reference or read at another time, I simply append it to a long-running text file I keep locally. A downside to this approach is that URLs alone often lack adequate detail of the contents into the web page, so grep'ing for an article by some keyword is not always possible.

Bobolink addresses that: one can simply add URLs, search by some keyword and have returned all articles that mention these terms.

N.B. Bobolink is not a "read-it-later" tool and does not offer any such functionality; searches return snippets of matching text only to highlight where an article matched a user's search query.