Find dataset DOIs from the Scholix API based on a list of target publication DOIs (e.g. from a CRIS system). The dataset DOIs can be looked up in DataCite to retrieve the dataset metadata (where it exists).
Please note that the datacite script uses the DataCite v1 API rather than the newer DataCite v2 API, so this code could likely be simplified.
These scripts require Python 3 with the requests, requests_cache, pandas, xmltodict and openpyxl modules. You can Install these modules using pip.
DataCite API results are cached in a file called datacite_api_cache.sqlite. Where DataCite does not have any information about a DOI (e.g. because it may not actually be a dataset) you will see 404 errors.
See the example input data input_dois.xlsx. This was a list of DOIs generated from a Scopus query.
A new file output.xlsx will be generated with the Dataset DOIs and Dataset Titles returned by the Scholix API. Note that this file can contain duplicate DOIs (where Scholix returned multiple links).
A new file output_datasets.xlsx will be generated with a summary of the DOIs which could be found in the DataCite v1 works API. An HTML export of each DOI (key fields only) will also be generated in the datasets_html folder.
$ python3 scholix_report.py
Loaded 500 unique DOIs (500 total)
Starting 8 threads
Downloading [********************] 100%
Processed 500 requests in 6 seconds (79 requests/second)
500 DOIs found, 0 DOIs had no data, 0 DOIs had errors
Found 644 dataset DOIs
Writing output file "output.xlsx"
python3 datacite.py
952 unique dataset DOIs (952 total)
...
https://api.datacite.org/works/10.1136/bmj.327.7417.708
title: Rapid tranquillisation for agitated patients in emergency psychiatric rooms: a randomised trial of midazolam versus haloperidol plus promethazine
published: 2003
authors: 1
...
Writing output file "output_datasets.xlsx"