Academic Tracker was created to automate the process of making sure that federally funded publications get listed on PubMed and that the grant funding source for them is cited.
Academic Tracker is a command line tool to search PubMed, ORCID, Google Scholar, and Crossref for publications. The program can either be given a list of authors to look for publications for, or references/citations to publications themselves. The program will then will look for publications on the aforementioned sources and return what relevant information is available from those sources.
The primary use case is to give the program a list of authors to find publications for. From this list of publications it can then be determined which ones need further action to be in compliance.
A secondary use case for finding author's publications is to create a report of the collaborators they have worked with, and can be done by specifying the creation of that report in the configuration file. Details on reports are in the documentation.
The other primary use case is to give the program a list of publication references to find information for.
The Academic Tracker package runs under Python 3.7+. Use pip to install. Starting with Python 3.4, pip is included by default. Be sure to use the latest version of pip as older versions are known to have issues grabbing all dependencies. Academic Tracker relies on sendmail to send emails, so if you need to use that feature be sure sendmail is installed and in PATH.
Install on Linux, Mac OS X
python3 -m pip install academic_tracker
Install on Windows
py -3 -m pip install academic_tracker
Upgrade on Linux, Mac OS X
python3 -m pip install academic_tracker --upgrade
Upgrade on Windows
py -3 -m pip install academic_tracker --upgrade
Academic Tracker has several commands and options. The simplest most common use case is simply:
academic_tracker author_search config_file.json
Creating The Configuration JSON
A configuration JSON file is required to run Academic Tracker, but initially creating it the first time can be burdensome. Unfortunately, there is no easy solution for this. It is encouraged to read the configuration JSON section in jsonschema and use the example there to create it initially. The add_authors command can help with building the Authors section if you already have a csv file with author information. A good tool to help track down pesky JSON syntax errors is here.
Registering With ORCID
In order to have this program search ORCID you must register with ORCID and obtain a key and secret. Details on how to do that are here. If you do not want to do that then the --no_ORCID option can be used to skip searching ORCID.
Mac OS Note
When you try to run the program on Mac OS you may get an SSL error.
certificate verify failed: unable to get local issuer certificate
This is due to a change in Mac OS and Python. To fix it go to to your Python folder in Applications and run the Install Certificates.command shell command in the /Applications/Python 3.x folder. This should fix the issue.
Email Sending Note
Academic Tracker uses sendmail to send emails, so any system it is going to be used on needs to have sendmail installed and the path in PATH. If you try to send emails without this the program will display a warning. This can be avoided by using the --test option though. The --test option blocks email sending. Email sending can also be avoided by leaving the from_email attribute out of the report sections of the configuration JSON file.
How Authors Are Identified
When searching by authors it is necessary to confirm that the author given to Academic Tracker matches the author returned in the query. In general this matching is done by matching the first and last names and at least one affiliation given for the author in the configuration JSON file. Note that affiliations can change over time as authors move, so they may need many affiliations to accurately match them to their publications depending on how far back you want to search in time.
How Publications Are Matched
When searching by publications it is necessary to confirm that the publication in the given reference matches the publication returned in the query. This is done by either matching the DOIs, PMIDs, or the title and at least one author. Titles are fuzzy matched using fuzzywuzzy which is why at least one author must also be matched. Author's are matched using last name and at least one affiliation.
This package is distributed under the BSD license.