Improved handling of empty/short affiliations

source-data · Oct 4, 2023 · d5580b3 · d5580b3
1 parent 9f3ac7d
commit d5580b3
Show file tree

Hide file tree

Showing 18 changed files with 762 additions and 58 deletions.
diff --git a/mecadoi/crossref/peer_review.py b/mecadoi/crossref/peer_review.py
@@ -27,7 +27,7 @@
     DoiBatch,
     DoiData,
     Head,
-    Institution,
+    Institution as CrossrefInstitution,
     InterWorkRelation,
     Orcid,
     PeerReview,
@@ -37,7 +37,7 @@
     ReviewDate,
     Titles,
 )
-from mecadoi.model import Author
+from mecadoi.model import Author, Institution as MecadoiInstitution
 
 
 def generate_peer_review_deposition(articles: List[Article]) -> str:
@@ -119,7 +119,7 @@ def generate_reviews(article: Article) -> Generator[PeerReview, None, None]:
                     month=review.publication_date.month,
                     day=review.publication_date.day,
                 ),
-                institution=Institution(institution_name=INSTITUTION_NAME),
+                institution=CrossrefInstitution(institution_name=INSTITUTION_NAME),
                 running_number=str(running_number),
                 program=Program(related_item=[is_review_of_relation]),
                 doi_data=DoiData(
@@ -149,7 +149,7 @@ def generate_reviews(article: Article) -> Generator[PeerReview, None, None]:
                     month=review.publication_date.month,
                     day=review.publication_date.day,
                 ),
-                institution=Institution(institution_name=INSTITUTION_NAME),
+                institution=CrossrefInstitution(institution_name=INSTITUTION_NAME),
                 running_number="Author Reply",
                 program=Program(
                     related_item=[is_review_of_relation]
@@ -171,6 +171,30 @@ def generate_reviews(article: Article) -> Generator[PeerReview, None, None]:
             )
 
 
+def create_institution(institution: MecadoiInstitution) -> CrossrefInstitution:
+    city = (
+        institution.city
+        if institution.city and len(institution.city) > 1
+        else None
+    )
+    country = (
+        institution.country
+        if institution.country and len(institution.country) > 1
+        else None
+    )
+    place = (
+        f"{city}, {country}" if city and country
+        else city if city
+        else country if country
+        else None
+    )
+    return CrossrefInstitution(
+        institution_name=institution.name,
+        institution_department=institution.department,
+        institution_place=place,
+    )
+
+
 def create_contributors(authors: List[Author]) -> Contributors:
     contributors = Contributors(
         person_name=[
@@ -182,19 +206,7 @@ def create_contributors(authors: List[Author]) -> Contributors:
                 affiliations=(
                     Affiliations(
                         institution=[
-                            Institution(
-                                institution_name=institution.name,
-                                institution_department=institution.department,
-                                institution_place=(
-                                    f"{institution.city}, {institution.country}"
-                                    if institution.city and institution.country
-                                    else institution.city
-                                    if institution.city
-                                    else institution.country
-                                    if institution.country
-                                    else None
-                                ),
-                            )
+                            create_institution(institution)
                             for institution in author.institutions
                         ]
                     )

diff --git a/scripts/unittest.sh b/scripts/unittest.sh
@@ -3,4 +3,4 @@
 # If any command fails, exit immediately with that command's exit status
 set -eo pipefail
 
-ENV_FILE=.env.ci python -m unittest
+ENV_FILE=.env.ci python -m unittest $@
diff --git a/tests/resources/expected/no-institution.xml b/tests/resources/expected/no-institution.xml
@@ -0,0 +1,115 @@
+<doi_batch xmlns="http://www.crossref.org/schema/5.3.1" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" version="5.3.1" xsi:schemaLocation="http://www.crossref.org/schema/5.3.1 http://www.crossref.org/schemas/crossref5.3.1.xsd">
+  <head>
+    <doi_batch_id>0000</doi_batch_id>
+    <timestamp>0</timestamp>
+    <depositor>
+      <depositor_name>EMBO</depositor_name>
+      <email_address>[email protected]</email_address>
+    </depositor>
+    <registrant>EMBO</registrant>
+  </head>
+  <body>
+    <peer_review language="en" revision-round="0" stage="pre-publication" type="referee-report">
+      <contributors>
+        <anonymous sequence="first" contributor_role="author"></anonymous>
+      </contributors>
+      <titles>
+        <title>Peer Review #1 of An article with a single revision round.</title>
+      </titles>
+      <review_date>
+        <month>10</month>
+        <day>20</day>
+        <year>2020</year>
+      </review_date>
+      <institution>
+        <institution_name>Review Commons</institution_name>
+      </institution>
+      <running_number>1</running_number>
+      <rel:program xmlns:rel="http://www.crossref.org/relations.xsd">
+        <rel:related_item>
+          <rel:inter_work_relation relationship-type="isReviewOf" identifier-type="doi">10.1101/single-revision-round.123.456.7890</rel:inter_work_relation>
+        </rel:related_item>
+      </rel:program>
+      <doi_data>
+        <doi>10.15252/rc.2020123456</doi>
+        <resource>https://eeb.embo.org/doi/10.1101/single-revision-round.123.456.7890#rev0-rr1</resource>
+      </doi_data>
+    </peer_review>
+    <peer_review language="en" revision-round="0" stage="pre-publication" type="referee-report">
+      <contributors>
+        <anonymous sequence="first" contributor_role="author"></anonymous>
+      </contributors>
+      <titles>
+        <title>Peer Review #2 of An article with a single revision round.</title>
+      </titles>
+      <review_date>
+        <month>10</month>
+        <day>20</day>
+        <year>2020</year>
+      </review_date>
+      <institution>
+        <institution_name>Review Commons</institution_name>
+      </institution>
+      <running_number>2</running_number>
+      <rel:program xmlns:rel="http://www.crossref.org/relations.xsd">
+        <rel:related_item>
+          <rel:inter_work_relation relationship-type="isReviewOf" identifier-type="doi">10.1101/single-revision-round.123.456.7890</rel:inter_work_relation>
+        </rel:related_item>
+      </rel:program>
+      <doi_data>
+        <doi>10.15252/rc.2020123456</doi>
+        <resource>https://eeb.embo.org/doi/10.1101/single-revision-round.123.456.7890#rev0-rr2</resource>
+      </doi_data>
+    </peer_review>
+    <peer_review language="en" revision-round="0" stage="pre-publication" type="author-comment">
+      <contributors>
+        <person_name contributor_role="author" sequence="first">
+          <given_name>Jane</given_name>
+          <surname>Doe</surname>
+          <affiliations>
+            <institution>
+              <institution_name>/</institution_name>
+            </institution>
+          </affiliations>
+          <ORCID authenticated="true">https://orcid.org/0000-0012-3456-7890</ORCID>
+        </person_name>
+        <person_name contributor_role="author" sequence="additional">
+          <given_name>John</given_name>
+          <surname>Doe</surname>
+          <affiliations>
+            <institution>
+              <institution_name>/</institution_name>
+            </institution>
+          </affiliations>
+        </person_name>
+      </contributors>
+      <titles>
+        <title>Author Reply to Peer Reviews of An article with a single revision round.</title>
+      </titles>
+      <review_date>
+        <month>10</month>
+        <day>20</day>
+        <year>2020</year>
+      </review_date>
+      <institution>
+        <institution_name>Review Commons</institution_name>
+      </institution>
+      <running_number>Author Reply</running_number>
+      <rel:program xmlns:rel="http://www.crossref.org/relations.xsd">
+        <rel:related_item>
+          <rel:inter_work_relation identifier-type="doi" relationship-type="isReviewOf">10.1101/single-revision-round.123.456.7890</rel:inter_work_relation>
+        </rel:related_item>
+        <rel:related_item>
+          <rel:inter_work_relation identifier-type="doi" relationship-type="isReplyTo">10.15252/rc.2020123456</rel:inter_work_relation>
+        </rel:related_item>
+        <rel:related_item>
+          <rel:inter_work_relation identifier-type="doi" relationship-type="isReplyTo">10.15252/rc.2020123456</rel:inter_work_relation>
+        </rel:related_item>
+      </rel:program>
+      <doi_data>
+        <doi>10.15252/rc.2020123456</doi>
+        <resource>https://eeb.embo.org/doi/10.1101/single-revision-round.123.456.7890#rev0-ar</resource>
+      </doi_data>
+    </peer_review>
+  </body>
+</doi_batch>
diff --git a/tests/resources/expected/no-institution.yml b/tests/resources/expected/no-institution.yml
@@ -0,0 +1,34 @@
+doi: 10.1101/single-revision-round.123.456.7890
+review_process:
+- author_reply:
+    authors:
+    - affiliation: /
+      given_name: Jane
+      is_corresponding_author: true
+      orcid:
+        id: https://orcid.org/0000-0012-3456-7890
+        is_authenticated: true
+      surname: Doe
+    - affiliation: /
+      given_name: John
+      is_corresponding_author: false
+      orcid: null
+      surname: Doe
+    doi: 10.15252/rc.2020123456
+    publication_date: 2020-10-20 00:00:00
+    text: {}
+  reviews:
+  - authors: []
+    doi: 10.15252/rc.2020123456
+    publication_date: 2020-10-20 00:00:00
+    text:
+      Estimated time to Complete Revisions (Required): Cannot tell / Not applicable
+      Significance (Required): Lorem ipsum dolor sit amet.
+  - authors: []
+    doi: 10.15252/rc.2020123456
+    publication_date: 2020-10-20 00:00:00
+    text:
+      Estimated time to Complete Revisions (Required): Between 3 and 6 months
+      Evidence, reproducibility and clarity (Required): 'Summary: this is a test'
+      Significance (Required): 'Signification: also a test'
+title: An article with a single revision round.
diff --git a/tests/resources/meca/no-institution/688_1_merged_1608565201.pdf b/tests/resources/meca/no-institution/688_1_merged_1608565201.pdf
diff --git a/tests/resources/meca/no-institution/Article Tracked Changes.doc b/tests/resources/meca/no-institution/Article Tracked Changes.doc
diff --git a/tests/resources/meca/no-institution/Article-1.doc b/tests/resources/meca/no-institution/Article-1.doc
diff --git a/tests/resources/meca/no-institution/Article.doc b/tests/resources/meca/no-institution/Article.doc
diff --git a/tests/resources/meca/no-institution/article.xml b/tests/resources/meca/no-institution/article.xml
@@ -0,0 +1,137 @@
+<?xml version="1.0" encoding="UTF-8"?>
+<!DOCTYPE article SYSTEM "JATS-archivearticle1.dtd">
+<article xmlns:xlink="http://www.w3.org/1999/xlink" dtd-version="1.2d1">
+  <front>
+    <journal-meta>
+      <journal-id journal-id-type="publisher">reviewcommonstest</journal-id>
+      <journal-title-group>
+        <journal-title>Review Commons - TEST</journal-title>
+        <abbrev-journal-title abbrev-type="nlm-ta" />
+      </journal-title-group>
+      <issn pub-type="ppub">2575-1077</issn>
+      <issn pub-type="epub">2575-1077</issn>
+    </journal-meta>
+    <article-meta>
+      <article-id pub-id-type="doi">10.12345/single-revision-round.1234567890</article-id>
+      <article-id pub-id-type="manuscript">single-revision-round-2020-1234</article-id>
+      <article-categories>
+        <subj-group subj-group-type="Article Type">
+          <subject>Research Article</subject>
+        </subj-group>
+      </article-categories>
+      <title-group>
+        <article-title>An article with a single revision round.</article-title>
+      </title-group>
+      <contrib-group>
+        <contrib contrib-type="author" corresp="yes">
+          <contrib-id contrib-id-type="orcid" specific-use="authenticated">https://orcid.org/0000-0012-3456-7890</contrib-id>
+          <name>
+            <surname>Doe</surname>
+            <given-names>Jane</given-names>
+            <prefix>Dr.</prefix>
+          </name>
+          <email>[email protected]</email>
+          <role content-type="1" />
+          <xref ref-type="aff" rid="aff1" />
+        </contrib>
+        <contrib contrib-type="author">
+          <name>
+            <surname>Doe</surname>
+            <given-names>John</given-names>
+          </name>
+          <email>[email protected]</email>
+          <role content-type="2" />
+          <xref ref-type="aff" rid="aff2" />
+        </contrib>
+        <aff id="aff1">
+          <institution>/</institution>
+          <addr-line content-type="addrline1">None</addr-line>
+          <addr-line content-type="city">/</addr-line>
+          <addr-line content-type="zipcode">/</addr-line>
+          <country>/</country>
+        </aff>
+        <aff id="aff2">
+          <institution>/</institution>
+          <addr-line content-type="addrline1">None</addr-line>
+          <addr-line content-type="city">/</addr-line>
+          <addr-line content-type="zipcode">/</addr-line>
+        </aff>
+      </contrib-group>
+      <author-notes>
+        <corresp id="cor1">
+          <label>*</label>
+          <bold>Corresponding Author:</bold>
+          Jane Doe, Medicinal Chemistry, EMBL
+        </corresp>
+        <fn fn-type="con">
+          <p>Author contributions: J.D. conceptualization; J.D. data curation.</p>
+        </fn>
+      </author-notes>
+      <pub-date pub-type="epub" />
+      <pub-date pub-type="ppub" />
+      <elocation-id>1234-5678.single-revision-round-2020-1234</elocation-id>
+      <history>
+        <date date-type="received">
+          <day>12</day>
+          <month>12</month>
+          <year>2020</year>
+        </date>
+      </history>
+      <permissions>
+        <copyright-statement></copyright-statement>
+        <copyright-year>2020</copyright-year>
+      </permissions>
+      <abstract abstract-type="primary">
+        <p>This article has a single revision round and thus in ...</p>
+      </abstract>
+      <kwd-group>
+        <kwd>test</kwd>
+        <kwd>test1</kwd>
+        <kwd>test2</kwd>
+      </kwd-group>
+      <funding-group>
+        <award-group>
+          <funding-source>
+            EMBO
+            <named-content content-type="funder-id">http://dx.doi.org/10.12345/asdf1234567890</named-content>
+          </funding-source>
+          <principal-award-recipient></principal-award-recipient>
+        </award-group>
+      </funding-group>
+      <custom-meta-group>
+        <custom-meta>
+          <meta-name>dual_publication</meta-name>
+          <meta-value>0</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>human_subjects</meta-name>
+          <meta-value>0</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>animal_subjects</meta-name>
+          <meta-value>0</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>Rebuttal Submitted to Hypothes.is</meta-name>
+          <meta-value>Yes</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>Pre-existing BioRxiv Preprint</meta-name>
+          <meta-value>Yes</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>Pre-existing BioRxiv Preprint DOI</meta-name>
+          <meta-value>10.1101/single-revision-round.123.456.7890</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>BioRxiv Preprint Submitted on Author's Behalf</meta-name>
+          <meta-value>Yes</meta-value>
+        </custom-meta>
+        <custom-meta>
+          <meta-name>Acknowledgment of Policy on Original Data</meta-name>
+          <meta-value>EJ</meta-value>
+        </custom-meta>
+      </custom-meta-group>
+    </article-meta>
+  </front>
+</article>
diff --git a/tests/resources/meca/no-institution/decision.pdf b/tests/resources/meca/no-institution/decision.pdf