arxiv-scan

Scan abstract listings on ArXiV for keywords and favorite authors in your categories to distill a list of papers most relevant for you. Keywords can be typed in manually or be found by ranking word occurrences in a provided file (e.g., a .bib file).

arxiv-scan was created by Robert Glas, Simeon Doetsch, and Martin Schlecker.

Installation

Requirements: Python >=3.5

Using pip

We recommend to install the latest stable version of arxiv-scan using pip (or pipx for an isolated environment):

pip install arxiv-scan

ℹ️ Depending on your Python installation, you might instead need pip3 install arxiv-scan or python3 -m pip install arxiv-scan.

From source

arxiv-scan is being developed on github. If you feel like hacking, feel free to install the latest version from there:

pip install --upgrade git+https://github.com/matiscke/arxiv-scan.git

Usage

Query today's arXiV listing for relevant papers

First setup your keywords and authors (see configuration section), then just run arxiv-scan (or python -m arxiv_scan) to get the relevant listings.

Command line reference

usage: arxiv-scan [-h] [--config /path/to/config] [--default-config [/path/to/config]]
                  [--config-convert [/path/to/config]] [--edit] [-d DATE] [-l LENGTH]
                  [-v RATING] [-c CATEGORIES] [--reverse] [--only-resubmissions]
                  [--ignore-cross-lists] [--ignore-abstract] [--log {info,debug}]
                  [--version]

optional arguments:
  -h, --help            show this help message and exit
  --config /path/to/config
                        Path to configuration file (check README for defaults)
  --default-config [/path/to/config]
                        Write default config to default location (or specified path)
  --config-convert [/path/to/config]
                        Convert authors and keywords config from legacy format
  --edit                Edit config in default text editor
  -d DATE, --date DATE  "new", or "recent", number of days in the past, "YYYY-MM" or
                        "YYYY-MM-DD". Defaults to "new"
  -l LENGTH, --len LENGTH
                        length of result list, all is -1
  -v RATING, --rating RATING
                        minimum rating for result list
  -c CATEGORIES, --categories CATEGORIES
                        arXiv subjects to scan, comma separated list
  --reverse             reverse list (lowest ranked paper on top)
  --only-resubmissions  Show only resubmissions
  --ignore-cross-lists  Ignore cross-lists
  --ignore-abstract     Ignore abstract in rating
  --log {info,debug}    Set loglevel
  --version             show program's version number and exit

Configuration

In the configuration file all the keywords and authors have to be set, as well as other optional configuration.

The easiest way to get started is to run arxiv-scan --edit, this will open the configuration file in the default text editor.

Alternatively create a default configfile with arxiv-scan --default-config, and edit it manually.

arXiv topics can be selected with the categories option, it accepts a comma-separated list of topics. List of topics

Configuration format:

[authors]
# author = rating
Alpher = 1
Bethe = 2
Gamov = 3

[keywords]
# keyword = rating
star = 1
planet = 2
habitable = 3

[options]
# other options (can also be set on CLI)
# default is used if omitted
categories = astro-ph.EP
date = new
length = -1
minimum_rating = 10 
reverse_list = False
only_resubmissions = False
show_cross_lists = True
ignore_abstract = False

Automatically extract keywords from a file (e.g. one with bibtex entries):

• Run arxiv-scan.wordcounter file_to_scan (or python -m scan_astroph.wordcounter file_to_scan). It scans the text file and extracts words with 4-12 characters from it, sorted by occurrence in the file.
• You will be asked to rank these suggested keywords. For each word shown, press 'Enter' to reject it or provide an integer rating, e.g., from 1 to 5 (higher=more relevant). Conclude by pressing C.
• Manually insert particularly important authors into the config file (e.g. with arxiv-scan --edit)

Configuration locations:

arxiv-scan searches the these paths for the config file, and loads the first found:

• from environment variable: $ARXIV_SCAN_CONF
• from home directory: ~/.arxiv-scan.conf
• default path (platform dependent):
  • on Linux / Unix (except MacOS): $XDG_CONFIG_HOME/arxiv-scan/arxiv-scan.conf (XDG_CONFIG_HOME defaults to ~/.config)
  • on MacOS: ~/Library/Application Support/arxiv-scan/arxiv-scan.conf
  • on Windows: $HOME/Documents/arxiv-scan/arxiv-scan.conf

Feedback

All feedback, including bug reports, feature requests, pull requests, etc., is welcome. arxiv-scan is being actively developed in an open repository; if you have any trouble please raise an issue.

License: MIT License

We thank arXiv for its open access interoperability. This package was not reviewed or approved by, nor does it necessarily express or reflect the policies or opinions of, arXiv.