kgourgou / arxivsearch

Search arXiv from the command line. Easily open as URL, PDF or output a generated BibTex entry. Integration with BibDesk to automatically insert references. Naive keyword scoring system to filter results to those likely to be relevant.

Tested

Only tested on OSX 10.8 with the following dependent packages:

• requests v0.14.1
• feedparser v5.1.3
• docopt v0.6.1
• nltk v2.0.4

Dependencies

pip install requests
pip install feedparser
pip install docopt
pip install nltk

Installation

• Update the options at the top of config.py to suit your preferences (See: Options)

Usage

arXiv Search

Usage:
arxiv.py id <arxiv_id> [--abstract] [--bib | --url | --pdf]
arxiv.py search <query> [--author=<author>] [--category=<category>] [--period=<period>] [--limit=<limit>] [--score=<score>] [--abstract] [--bib | --url | --pdf]
arxiv.py new [--author=<author>] [--category=<category>] [--period=<period>] [--limit=<limit>] [--score=<score>] [--abstract] [--bib | --url | --pdf]
arxiv.py bib <arxiv_id>
arxiv.py url <arxiv_id>
arxiv.py pdf <arxiv_id>

Options:
arxiv_id                      The ID of an arXiv paper 
query                         A query string to search arXiv
url, --url                    Opens the following arXiv id as a URL
pdf, --pdf                    Opens the following arXiv id as a PDF
bib, --bib                    Will print out bibtex entry for each matching result
--category=<category>         The category to restrict searches to
--author=<author>             The author to restrict searches to
--limit=<limit>               The maximum number of results to retrieve
--score=<score>               The score cutoff for the display of papers [default: 0]
--period=<period>             The days into the past to retrieve results for 
--abstract                    Will print out the abstract for each matching result

Options (config.py)

# Whether or not you use BibDesk to handle your references (will automatically
# export BibTex references into the currently open BibDesk bilbiography)
USE_BIBDESK = True

# The default categories to search if no other categories are specified
DEFAULT_CATEGORIES = ['quant-ph']

# Whether or not to include the abstract in exported BibTex references
INC_ABSTRACT = True

# The number of results to return if no number is specified
DEFAULT_LIMIT = 100

# The software to use to 'open' a URL. Try "open" for OSX, "gnome-open" for
# Linux or "cmd /c start" for Windows. Commands for your specific setup may
# need to be determined.
OPEN_SOFTWARE = "open"

# A Python .format() string representing how results will be formatted
OUTPUT_FORMAT = "{i:<6} {date} ({arxiv_id}) | {score} | {cate} | {mark}{title}{mark}"

# Keywords for scoring purposes. Keyword/Phrase: Score.
KEYWORDS = {
    'topological': 10, 
    'qubit': 5, 
    'qec': 20,
    'fault tolerant': 25,
    'error correction': 20,
    'error': 5,
}

Future Prospects

• Automatically generate scoring keywords and weights based on a provided .bib file of references.
• Improve and de-hackify the arXiv journal_ref parsing