def test_citekey_to_csl_item_pubmed_book():
"""
Extracting CSL metadata from books in PubMed is not supported.
Logic not implemented to parse XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=29227604&rettype=full
"""
with pytest.raises(NotImplementedError):
citekey_to_csl_item("pmid:29227604")
def test_citekey_to_csl_item_pubmed_2():
"""
Generated from XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=27094199&rettype=full
"""
citekey = "pmid:27094199"
csl_item = citekey_to_csl_item(citekey)
print(csl_item)
assert csl_item["id"] == "alaFV9OY"
assert csl_item["type"] == "article-journal"
assert csl_item["URL"] == "https://www.ncbi.nlm.nih.gov/pubmed/27094199"
assert csl_item[
"container-title"] == "Circulation. Cardiovascular genetics"
assert csl_item["container-title-short"] == "Circ Cardiovasc Genet"
assert csl_item["page"] == "179-84"
assert (
csl_item["title"] ==
"Genetic Association-Guided Analysis of Gene Networks for the Study of Complex Traits."
)
assert csl_item["issued"]["date-parts"] == [[2016, 4]]
authors = csl_item["author"]
assert authors[0]["given"] == "Casey S"
assert authors[0]["family"] == "Greene"
assert csl_item["PMID"] == "27094199"
assert csl_item["DOI"] == "10.1161/circgenetics.115.001181"
def test_citekey_to_csl_item_clinical_trial():
"""
Test clinicaltrials.gov citation support using CURIEs.
https://github.com/manubot/manubot/issues/216
"""
csl_item = citekey_to_csl_item("clinicaltrials:NCT04292899")
assert csl_item["title"].startswith("A Phase 3 Randomized Study")
assert csl_item["source"].startswith("clinicaltrials.gov")
assert csl_item["URL"] == "https://clinicaltrials.gov/ct2/show/NCT04292899"
def test_citekey_to_csl_item_pubmed_book(caplog):
"""
Extracting CSL metadata from books in PubMed is not supported.
Logic not implemented to parse XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=29227604&rettype=full
"""
csl_item = citekey_to_csl_item("pmid:29227604", log_level="ERROR")
assert csl_item is None
assert "Unsupported PubMed record: no <Article> element" in caplog.text
def test_citekey_to_csl_item_pubmed_with_numeric_month():
"""
Generated from XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=29028984&rettype=full
See https://github.com/manubot/manubot/issues/69
"""
citekey = "pmid:29028984"
csl_item = citekey_to_csl_item(citekey)
print(csl_item)
assert csl_item["issued"]["date-parts"] == [[2018, 3, 15]]
def test_citekey_to_csl_item_pubmed_book(caplog):
"""
Extracting CSL metadata from books in PubMed is not supported.
Logic not implemented to parse XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=29227604&rettype=full
"""
csl_item = citekey_to_csl_item("pmid:29227604", log_level="ERROR")
assert csl_item is None
assert (
"Expected article to be an XML element with tag PubmedArticle, received tag 'PubmedBookArticle'"
in caplog.text)
def test_citekey_to_csl_item_doi_datacite():
citekey = "doi:10.7287/peerj.preprints.3100v1"
csl_item = citekey_to_csl_item(citekey)
assert csl_item["id"] == "11cb5HXoY"
assert csl_item["URL"] == "https://doi.org/10.7287/peerj.preprints.3100v1"
assert csl_item["DOI"] == "10.7287/peerj.preprints.3100v1"
assert csl_item["type"] == "report"
assert (csl_item["title"] ==
"Sci-Hub provides access to nearly all scholarly literature")
authors = csl_item["author"]
assert authors[0]["family"] == "Himmelstein"
assert authors[-1]["family"] == "Greene"
def test_citekey_to_csl_item_doi_datacite():
citekey = 'doi:10.7287/peerj.preprints.3100v1'
csl_item = citekey_to_csl_item(citekey)
assert csl_item['id'] == '11cb5HXoY'
assert csl_item['URL'] == 'https://doi.org/10.7287/peerj.preprints.3100v1'
assert csl_item['DOI'] == '10.7287/peerj.preprints.3100v1'
assert csl_item['type'] == 'report'
assert csl_item[
'title'] == 'Sci-Hub provides access to nearly all scholarly literature'
authors = csl_item['author']
assert authors[0]['family'] == 'Himmelstein'
assert authors[-1]['family'] == 'Greene'
def test_citekey_to_csl_item_arxiv():
citekey = "arxiv:cond-mat/0703470v2"
csl_item = citekey_to_csl_item(citekey)
assert csl_item["id"] == "ES92tcdg"
assert csl_item["URL"] == "https://arxiv.org/abs/cond-mat/0703470v2"
assert csl_item["number"] == "cond-mat/0703470v2"
assert csl_item["version"] == "v2"
assert csl_item["type"] == "report"
assert csl_item["container-title"] == "arXiv"
assert csl_item["title"] == "Portraits of Complex Networks"
authors = csl_item["author"]
assert authors[0]["literal"] == "J. P. Bagrow"
assert csl_item["DOI"] == "10.1209/0295-5075/81/68004"
def test_citekey_to_csl_item_arxiv():
citekey = 'arxiv:cond-mat/0703470v2'
csl_item = citekey_to_csl_item(citekey)
assert csl_item['id'] == 'ES92tcdg'
assert csl_item['URL'] == 'https://arxiv.org/abs/cond-mat/0703470v2'
assert csl_item['number'] == 'cond-mat/0703470v2'
assert csl_item['version'] == '2'
assert csl_item['type'] == 'report'
assert csl_item['container-title'] == 'arXiv'
assert csl_item['title'] == 'Portraits of Complex Networks'
authors = csl_item['author']
assert authors[0]['literal'] == 'J. P. Bagrow'
assert csl_item['DOI'] == '10.1209/0295-5075/81/68004'
def test_citekey_to_csl_item_pmc():
"""
https://api.ncbi.nlm.nih.gov/lit/ctxp/v1/pmc/?format=csl&id=3041534
"""
citekey = f'pmcid:PMC3041534'
csl_item = citekey_to_csl_item(citekey)
assert csl_item['id'] == 'RoOhUFKU'
assert csl_item[
'URL'] == 'https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3041534/'
assert csl_item['container-title-short'] == 'Summit Transl Bioinform'
assert csl_item[
'title'] == 'Secondary Use of EHR: Data Quality Issues and Informatics Opportunities'
authors = csl_item['author']
assert authors[0]['family'] == 'Botsis'
assert csl_item['PMID'] == '21347133'
assert csl_item['PMCID'] == 'PMC3041534'
assert 'generated by Manubot' in csl_item['note']
assert 'standard_id: pmcid:PMC3041534' in csl_item['note']
def getPaperTitleFromIssue(issue):
""" gets the papers title using manubot; if manubot can't get title, extract from issue title """
try:
# Try using manubot
citekey = getCitationFromIssue(issue)
csl_item = cite.citekey_to_csl_item(citekey)
title = csl_item["title"]
return title
except:
# On error, try getting from issue title
try:
title = issue["title"].split(":")[1]
return title
except:
print(
"the paper title could not be automatically extracted from the following issue: \n",
issue["title"])
return "unknown"
def test_citekey_to_csl_item_pmc():
"""
https://api.ncbi.nlm.nih.gov/lit/ctxp/v1/pmc/?format=csl&id=3041534
"""
citekey = f"pmcid:PMC3041534"
csl_item = citekey_to_csl_item(citekey)
assert csl_item["id"] == "RoOhUFKU"
assert csl_item["URL"] == "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3041534/"
assert csl_item["container-title-short"] == "Summit Transl Bioinform"
assert (
csl_item["title"]
== "Secondary Use of EHR: Data Quality Issues and Informatics Opportunities"
)
authors = csl_item["author"]
assert authors[0]["family"] == "Botsis"
assert csl_item["PMID"] == "21347133"
assert csl_item["PMCID"] == "PMC3041534"
assert "generated by Manubot" in csl_item["note"]
assert "standard_id: pmcid:PMC3041534" in csl_item["note"]
def test_citekey_to_csl_item_pubmed_1():
"""
Generated from XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=21347133&rettype=full
"""
citekey = 'pmid:21347133'
csl_item = citekey_to_csl_item(citekey)
assert csl_item['id'] == 'y9ONtSZ9'
assert csl_item['type'] == 'article-journal'
assert csl_item['URL'] == 'https://www.ncbi.nlm.nih.gov/pubmed/21347133'
assert csl_item[
'container-title'] == 'Summit on translational bioinformatics'
assert csl_item[
'title'] == 'Secondary Use of EHR: Data Quality Issues and Informatics Opportunities.'
assert csl_item['issued']['date-parts'] == [[2010, 3, 1]]
authors = csl_item['author']
assert authors[0]['given'] == 'Taxiarchis'
assert authors[0]['family'] == 'Botsis'
assert csl_item['PMID'] == '21347133'
assert csl_item['PMCID'] == 'PMC3041534'
def test_citekey_to_csl_item_pubmed_1():
"""
Generated from XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=21347133&rettype=full
"""
citekey = "pmid:21347133"
csl_item = citekey_to_csl_item(citekey)
assert csl_item["id"] == "y9ONtSZ9"
assert csl_item["type"] == "article-journal"
assert csl_item["URL"] == "https://www.ncbi.nlm.nih.gov/pubmed/21347133"
assert csl_item["container-title"] == "Summit on translational bioinformatics"
assert (
csl_item["title"]
== "Secondary Use of EHR: Data Quality Issues and Informatics Opportunities."
)
assert csl_item["issued"]["date-parts"] == [[2010, 3, 1]]
authors = csl_item["author"]
assert authors[0]["given"] == "Taxiarchis"
assert authors[0]["family"] == "Botsis"
assert csl_item["PMID"] == "21347133"
assert csl_item["PMCID"] == "PMC3041534"
def test_citekey_to_csl_item_pubmed_2():
"""
Generated from XML returned by
https://eutils.ncbi.nlm.nih.gov/entrez/eutils/efetch.fcgi?db=pubmed&id=27094199&rettype=full
"""
citekey = 'pmid:27094199'
csl_item = citekey_to_csl_item(citekey)
print(csl_item)
assert csl_item['id'] == 'alaFV9OY'
assert csl_item['type'] == 'article-journal'
assert csl_item['URL'] == 'https://www.ncbi.nlm.nih.gov/pubmed/27094199'
assert csl_item[
'container-title'] == 'Circulation. Cardiovascular genetics'
assert csl_item['container-title-short'] == 'Circ Cardiovasc Genet'
assert csl_item['page'] == '179-84'
assert csl_item[
'title'] == 'Genetic Association-Guided Analysis of Gene Networks for the Study of Complex Traits.'
assert csl_item['issued']['date-parts'] == [[2016, 4]]
authors = csl_item['author']
assert authors[0]['given'] == 'Casey S'
assert authors[0]['family'] == 'Greene'
assert csl_item['PMID'] == '27094199'
assert csl_item['DOI'] == '10.1161/circgenetics.115.001181'
def test_citekey_to_csl_item_isbn():
csl_item = citekey_to_csl_item("isbn:9780387950693")
assert csl_item["type"] == "book"
assert csl_item["title"] == "Complex analysis"
def test_citekey_to_csl_item_isbn():
csl_item = citekey_to_csl_item('isbn:9780387950693')
assert csl_item['type'] == 'book'
assert csl_item['title'] == 'Complex analysis'
