The Research Organization Registry provides access to a growing collection of unique and persistent identifiers (RORs) for research organizations and a web interface and API for searching the collection with an organization name or a affiliation string. The web interface and API work well if you have a small number of affiliation strings that need identifiers, but what if you have several hundred organization names or thousands of affiliation strings? This is the daunting problem that repositories face when they are trying to populate their existing records with identifiers. ROR Retriever is a tool developed to help address these repositories and others that are ready to add identifiers to large collections with many affiliations.
Use python RORRetriever.py -h to see this usage description.
usage: RORRetriever [-h] [-a [AFFILIATIONLIST [AFFILIATIONLIST ...]]] [-af AFFILIATIONFILENAME] [-ad AFFILIATIONDATA] [--noacronyms] [--max] [--details] [-o OUTPUTINTERVAL] [--loglevel {debug,info,warning}]
[--logto FILE]
search organization names and affiliations for RORs using the ROR Affiliation Strategy
optional arguments:
-h, --help show this help message and exit
-a [AFFILIATIONLIST [AFFILIATIONLIST ...]], --affiliationList [AFFILIATIONLIST [AFFILIATIONLIST ...]]
a list of "affiliations in quotes"
-af AFFILIATIONFILENAME, --affiliationFilename AFFILIATIONFILENAME
a file with one affiliation per line in current working directory
-ad AFFILIATIONDATA, --affiliationData AFFILIATIONDATA
datafile (tsv, csv) with affiliations in cwd)
--noacronyms Exclude Acronym matches (default=False)
--max Accept max score if no result chosen by ROR algorithm (more results and noise)
--details Show detailed response data (automatic for one ROR)
-o OUTPUTINTERVAL, --outputInterval OUTPUTINTERVAL
For batch processing output results update interval default:20
--loglevel {debug,info,warning}
Logging level
--logto FILE Log file (will overwrite if exists)
ROR Retriever imports the following python modules: requests, json, pandas, urllib.parse, sys, datetime, argparse, datetime, logging, os, re. The file RORRetriever.yml can be used to create the environment using the command conda env create -f RORRetriever.yml
The results of the search are provided in a tab-delimited file called AffiliationAPI_RORData__TIMESTAMP.tsv. This file has at least one row for every input affiliation but can have more if more than one ROR is identified for a single affiliation. In many cases when multiple RORs are found, some are incorrect and curation is required to select the correct one.
The columns in this file are:
| Name | Definition |
|---|---|
| affiliation | the complete affiliation being searched |
| searchString_Affiliation | Complete affiliations are split into substrings during the search. This is the substring that foind the match. |
| ROR_Affiliation | The ROR identifier found using the substring. |
| organizationLookupName_Affiliation | The name of the organization with the found ROR. |
| country_Affiliation | The country the found organization is in. |
| match_Affiliation | The kind of match that was found (provided by algorithm) |
| chosen_Affiliation | TRUE if the algorithm chose this organization. |
| score | The algorithm score for this organization (0 - 1). If chosen is TRUE, score is 1. |
| numberOfResults_Affiliation | The number of results returned from the search. |
| valid | A column for recording validity during curation (always TRUE prior to curation). |
Some details of the ROR algorithm response can be displayed using the --details flag. This is automatic if only one affiliation is being searched. if --details is set, a table of the search results is shown with the following fields:
| Name | Definition |
|---|---|
| substring | Search string (can be substring of complete affiliation) |
| score | Match score between 0 and 1 (1 is the best match chosen by algorithm) |
| matchingType | Method that found the match (provided by algorithm) |
| chosen | True for chosen ROR False for others |
| organization | Name of organization for ROR (should match substring) |
| country | Country of organization |
As an example of some of the capabilities of this tool consider this list of affiliations: The University of Alabama Libraries Arizona State University Library UC Berkeley Library MIT Libraries
These strings are affiliations because they contain the names of research organizations along with other text. These are simple affiliations as they include university names with the words library or libraries. Despite this simplicity, they demonstrate some interesting characteristics of the ROR Affiliation search. These affiliations are in a test file in this repository (OrganizationNamesTest.txt).
The command RORRetriever -af OrganizationNamesTest.txt uses the ROR affiliation API to search for RORs for these affiliations and gives this output to the terminal:
RORRetriever -af OrganizationNamesTest.txt
2022-06-27 11:25:00:INFO:RORRetriever: Searching OrganizationNamesTest.txt for affiliations with Affiliation Strategy
2022-06-27 11:25:00:INFO:RORRetriever: 4 Input Affiliations
2022-06-27 11:25:03:INFO:RORRetriever: 3 new RORs written to AffiliationAPI_RORData__20220627_11.tsv
2022-06-27 11:25:03:INFO:RORRetriever: AffiliationAPI_RORData__20220627_11.tsv 3 RORs Found
More details are written to a tab-separated file (converted to a table here):
| affiliation | searchString_Affiliation | ROR_Affiliation | organizationLookupName_Affiliation | country_Affiliation | match_Affiliation | chosen_Affiliation | score | numberOfResults_Affiliation | valid |
|---|---|---|---|---|---|---|---|---|---|
| The University of Alabama Libraries | University of Alabama | https://ror.org/03xrrjk67 | University of Alabama | United States | HEURISTICS | True | 1.0 | 9 | True |
| Arizona State University Library | No Match | 4 | False | ||||||
| UC Berkeley Library | UC Berkeley Library | https://ror.org/01bndk311 | Berkeley Public Library | United States | COMMON TERMS | True | 0.9 | 10 | True |
| MIT Libraries | MIT | https://ror.org/047m6hg73 | Management Intelligenter Technologien (Germany) | Germany | ACRONYM | True | 0.9 | 18 | True |
The output data shows that between 4 and 18 results were found for each of these affiliations and that RORs were identified for three. The organizationLookupName_Affiliation column gives the name associated with the ROR. In the first row, this name clearly matches the organization name in the input affiliation (The University of Alabama Libraries). In the other cases, the names suggest that either no match was chosen or the incorrect match was chosen.
There are two ways to find out what actually happened. First, the --details option can be used to show details for all searches at once, i.e. RORRetriever -af OrganizationNamesTest.txt --details or we can run each affiliation one at a time which automatically shows details.
The command RORRetriever -a "The University of Alabama Libraries" shows what happened for the affiliation "The University of Alabama Libraries". There were 9 results returned and University of Alabama was selected as the best result (score is 1 and chosen is True). Even though everything went well in this case, it demonstrates the kind of challenges that occur when over 100,000 organization names are being searched. First, many Universities have multiple campuses with very similar names. There are also many cases where states have two Universities with names like "University of Alabama" and "Alabama State University". In this case the algorithm overcame these challenges to find the correct organization. Great job!
2022-06-27 12:40:48:INFO:RORRetriever: 1 Input Affiliations
substring score matching_type chosen ror organization country
University of Alabama 1.00 HEURISTICS True https://ror.org/03xrrjk67 University of Alabama United States
University of Alabama 0.89 HEURISTICS False https://ror.org/010acrp16 University of West Alabama United States
University of Alabama 0.88 HEURISTICS False https://ror.org/01s7b5y08 University of South Alabama United States
University of Alabama 0.88 HEURISTICS False https://ror.org/0584fj407 University of North Alabama United States
Alabama University 0.86 HEURISTICS False https://ror.org/01eedy375 Alabama State University United States
University of Alabama 0.86 HEURISTICS False https://ror.org/051fvmk98 University of Alabama System United States
Alabama University 0.82 HEURISTICS False https://ror.org/05hz8m414 Alabama Agricultural and Mechanical University United States
The University of Alabama Libraries 0.76 COMMON TERMS False https://ror.org/008s83205 University of Alabama at Birmingham United States
University of Alabama 0.75 HEURISTICS False https://ror.org/02zsxwr40 University of Alabama in Huntsville United States
2022-06-27 12:40:49:INFO:RORRetriever: 1 new RORs written to AffiliationAPI_RORData__20220627_12.tsv
2022-06-27 12:40:49:INFO:RORRetriever: AffiliationAPI_RORData__20220627_12.tsv 1 RORs Found
The command RORRetriever -a "Arizona State University Library" gives the following results:
2022-06-27 11:39:19:INFO:RORRetriever: 1 Input Affiliations
substring score matching_type chosen ror organization country
Arizona State University Library 0.86 COMMON TERMS False https://ror.org/03efmqc40 Arizona State University United States
Arizona State University Library 0.65 COMMON TERMS False https://ror.org/03y1zyv86 Oklahoma State University Oklahoma City United States
Arizona State University Library 0.54 COMMON TERMS False https://ror.org/00g3q3z36 Arizona State Museum United States
Arizona State University Library 0.54 COMMON TERMS False https://ror.org/01wwqrh75 Arizona State Library, Archives and Public Records United States
2022-06-27 11:39:20:INFO:RORRetriever: 0 new RORs written to AffiliationAPI_RORData__20220627_11.tsv
2022-06-27 11:39:20:INFO:RORRetriever: AffiliationAPI_RORData__20220627_11.tsv 0 RORs Found
The table shows the 4 results that were found and that the chosen column is False for all four, even though the right answer has the highest (0.86). The algorithm considers a number of factors in making a choice and, in this case, the correct answer did not pass the test.
The --max option can be used to choose the best answer even if the algorithm does not, so the command RORRetriever -a "Arizona State University Library" --max gives the results:
2022-06-27 11:45:15:INFO:RORRetriever: 1 Input Affiliations
2022-06-27 11:45:15:INFO:RORRetriever: ************** Best match is being found (may not be chosen by algorithm, score < 1.0)
substring score matching_type chosen ror organization country
Arizona State University Library 0.86 COMMON TERMS False https://ror.org/03efmqc40 Arizona State University United States
Arizona State University Library 0.65 COMMON TERMS False https://ror.org/03y1zyv86 Oklahoma State University Oklahoma City United States
Arizona State University Library 0.54 COMMON TERMS False https://ror.org/00g3q3z36 Arizona State Museum United States
Arizona State University Library 0.54 COMMON TERMS False https://ror.org/01wwqrh75 Arizona State Library, Archives and Public Records United States
2022-06-27 11:45:16:INFO:RORRetriever: 1 new RORs written to AffiliationAPI_RORData__20220627_11.tsv
2022-06-27 11:45:16:INFO:RORRetriever: AffiliationAPI_RORData__20220627_11.tsv 1 RORs Found
and the correct ROR is found. Note that the terminal output warns that this flag is set to make sure the user is aware that, while some results may improve, as in this example, there is also more noise introduced into the results. Also note that the results file contains a column named chosen which will be True for RORs selected by the algorithm and False for those not selected by the algorithm.
The details for the affiliation "UC Berkelely Library" can be seen using the command RORRetriever -a "UC Berkelely Library" with the results:
2022-06-27 11:54:23:INFO:RORRetriever: 1 Input Affiliations
substring score matching_type chosen ror organization country
UC Berkeley Library 0.90 COMMON TERMS True https://ror.org/01bndk311 Berkeley Public Library United States
UC Berkeley Library 0.79 COMMON TERMS False https://ror.org/02jbv0t02 Lawrence Berkeley National Laboratory United States
UC Berkeley Library 0.58 COMMON TERMS False https://ror.org/01an7q238 University of California, Berkeley United States
UC Berkeley Library 0.57 COMMON TERMS False https://ror.org/02xewxa75 Berkeley College United States
UC Berkeley Library 0.52 COMMON TERMS False https://ror.org/01n7gem85 Deutsche Nationalbibliothek Germany
UC Berkeley Library 0.50 COMMON TERMS False https://ror.org/00b6wzg36 Acupuncture & Integrative Medicine College United States
UC Berkeley Library 0.46 COMMON TERMS False https://ror.org/03fgher32 UC Irvine Health United States
UC Berkeley Library 0.44 COMMON TERMS False https://ror.org/048vdhs48 Verbundzentrale des GBV Germany
UC Berkeley Library 0.34 COMMON TERMS False https://ror.org/05ehe8t08 UC Davis Children's Hospital United States
UC Berkeley Library 0.32 COMMON TERMS False https://ror.org/04wvygn49 Fundación Copec UC Chile
In this case, the algorithm choses the ROR for "Berkeley Public Library" which has the maximum score (0.90) but is incorrect. The correct answer, University of California, Berkeley is in the results with a score of 0.58. This example is included here to remind us that there is an algorithm involved in these searches and we should help it as much as possible when we are providing affiliations to journals or libraries. In this case the acronym UC is a problem. The algorithm cannot guess that it stands for University of California. Help it out by writing the affiliation as University of California Berkeley Library and the algorithm gets the right answer (RORRetriever -a "University of California Berkeley Library").
The final example (MIT Libraries) illustrated another problem case, using an acronym in an affiliation rather than spelling it out. In this case, the command RORRetriever -a "MIT Libraries" shows the 7 organizations with the acronym MIT:
RORRetriever % RORRetriever -a "MIT Libraries"
2022-06-27 12:03:48:INFO:RORRetriever: 1 Input Affiliations
substring score matching_type chosen ror organization country
MIT 0.90 ACRONYM False https://ror.org/002r67t24 Manukau Institute of Technology New Zealand
MIT 0.90 ACRONYM False https://ror.org/02mb6z761 Myanmar Institute of Theology Myanmar
MIT 0.90 ACRONYM False https://ror.org/02qk9nj07 Ministry of Infrastructures and Transport Italy
MIT 0.90 ACRONYM False https://ror.org/02yyma482 International Tourism Institute Slovenia
MIT 0.90 ACRONYM False https://ror.org/042nb2s44 Massachusetts Institute of Technology United States
MIT 0.90 ACRONYM True https://ror.org/047m6hg73 Management Intelligenter Technologien (Germany) Germany
MIT 0.90 ACRONYM False https://ror.org/04mtcj695 University of Southern Mindanao Philippines
MIT Libraries 0.77 COMMON TERMS False https://ror.org/00n3xk635 Libraries Tasmania Australia
MIT Libraries 0.76 COMMON TERMS False https://ror.org/05mgwy297 Smithsonian Libraries United States
MIT Libraries 0.59 FUZZY False https://ror.org/03qa42z85 Librairie Droz Switzerland
MIT Libraries 0.58 COMMON TERMS False https://ror.org/00w2b1v71 Hesburgh Libraries United States
MIT Libraries 0.57 COMMON TERMS False https://ror.org/022z6jk58 MIT Lincoln Laboratory United States
MIT Libraries 0.56 COMMON TERMS False https://ror.org/0239rpj17 ImagineIF Libraries United States
MIT Libraries 0.55 COMMON TERMS False https://ror.org/00k8bhh21 Polar Libraries Colloquy Canada
MIT Libraries 0.50 COMMON TERMS False https://ror.org/00rj4dg52 Cambridge–MIT Institute United Kingdom
MIT Libraries 0.46 COMMON TERMS False https://ror.org/03fg5ns40 MIT Sea Grant United States
MIT Libraries 0.44 COMMON TERMS False https://ror.org/00v140q16 MIT University North Macedonia
MIT Libraries 0.22 FUZZY False https://ror.org/05nfbnp91 Academy of Cryptography Techniques Vietnam
The flag --noacronyms can prevent RORs based on acronyms from being selected. In this case, that misses the correct ROR but, in cases where RORs are selected on the basis of abbreviations in many affiliations, this can be helpful. Try RORRetriever -a "Metadata LLC USA" for an interesting example.
Keep in mind that ROR uses an algorithm to try to find RORs for affiliation strings and, as is always true, the algorithm is designed to work well in most cases. RORRetriever can help find RORs in batch mode if you have many affiliations or organization names to search. I find it very helpful, but always treat the results with care. See the ROR API Documentation for more tips and tricks.
Most important - write clear affilaitions when submitting data and publications!
DataCiteFacets by Ted Habermann is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
