mirror of
https://github.com/bookwyrm-social/bookwyrm.git
synced 2024-05-20 17:28:55 +00:00
228 lines
7.7 KiB
Python
228 lines
7.7 KiB
Python
"""ISNI author checking utilities"""
|
|
import xml.etree.ElementTree as ET
|
|
from typing import Union, Optional
|
|
|
|
import requests
|
|
|
|
from bookwyrm import activitypub, models
|
|
|
|
|
|
def get_element_text(element: Optional[ET.Element]) -> str:
|
|
"""If the element is not None and there is a text attribute return this"""
|
|
if element is not None and element.text is not None:
|
|
return element.text
|
|
return ""
|
|
|
|
|
|
def request_isni_data(search_index: str, search_term: str, max_records: int = 5) -> str:
|
|
"""Request data from the ISNI API"""
|
|
|
|
search_string = f'{search_index}="{search_term}"'
|
|
query_params: dict[str, Union[str, int]] = {
|
|
"query": search_string,
|
|
"version": "1.1",
|
|
"operation": "searchRetrieve",
|
|
"recordSchema": "isni-b",
|
|
"maximumRecords": max_records,
|
|
"startRecord": "1",
|
|
"recordPacking": "xml",
|
|
"sortKeys": "RLV,pica,0,,",
|
|
}
|
|
result = requests.get("http://isni.oclc.org/sru/", params=query_params, timeout=15)
|
|
# the OCLC ISNI server asserts the payload is encoded
|
|
# in latin1, but we know better
|
|
result.encoding = "utf-8"
|
|
return result.text
|
|
|
|
|
|
def make_name_string(element: ET.Element) -> str:
|
|
"""create a string of form 'personal_name surname'"""
|
|
|
|
# NOTE: this will often be incorrect, many naming systems
|
|
# list "surname" before personal name
|
|
forename = element.find(".//forename")
|
|
surname = element.find(".//surname")
|
|
|
|
forename_text = get_element_text(forename)
|
|
surname_text = get_element_text(surname)
|
|
|
|
return "".join(
|
|
[forename_text, " " if forename_text and surname_text else "", surname_text]
|
|
)
|
|
|
|
|
|
def get_other_identifier(element: ET.Element, code: str) -> str:
|
|
"""Get other identifiers associated with an author from their ISNI record"""
|
|
|
|
identifiers = element.findall(".//otherIdentifierOfIdentity")
|
|
for section_head in identifiers:
|
|
if (
|
|
(section_type := section_head.find(".//type")) is not None
|
|
and section_type.text is not None
|
|
and section_type.text == code
|
|
and (identifier := section_head.find(".//identifier")) is not None
|
|
and identifier.text is not None
|
|
):
|
|
return identifier.text
|
|
|
|
# if we can't find it in otherIdentifierOfIdentity,
|
|
# try sources
|
|
for source in element.findall(".//sources"):
|
|
if (
|
|
(code_of_source := source.find(".//codeOfSource")) is not None
|
|
and code_of_source.text is not None
|
|
and code_of_source.text.lower() == code.lower()
|
|
and (source_identifier := source.find(".//sourceIdentifier")) is not None
|
|
and source_identifier.text is not None
|
|
):
|
|
return source_identifier.text
|
|
|
|
return ""
|
|
|
|
|
|
def get_external_information_uri(element: ET.Element, match_string: str) -> str:
|
|
"""Get URLs associated with an author from their ISNI record"""
|
|
|
|
sources = element.findall(".//externalInformation")
|
|
for source in sources:
|
|
information = source.find(".//information")
|
|
uri = source.find(".//URI")
|
|
if (
|
|
uri is not None
|
|
and uri.text is not None
|
|
and information is not None
|
|
and information.text is not None
|
|
and information.text.lower() == match_string.lower()
|
|
):
|
|
return uri.text
|
|
return ""
|
|
|
|
|
|
def find_authors_by_name(
|
|
name_string: str, description: bool = False
|
|
) -> list[activitypub.Author]:
|
|
"""Query the ISNI database for possible author matches by name"""
|
|
|
|
payload = request_isni_data("pica.na", name_string)
|
|
# parse xml
|
|
root = ET.fromstring(payload)
|
|
# build list of possible authors
|
|
possible_authors = []
|
|
for element in root.iter("responseRecord"):
|
|
|
|
# TODO: we don't seem to do anything with the
|
|
# personal_name variable - is this code block needed?
|
|
personal_name = element.find(".//forename/..")
|
|
if not personal_name:
|
|
continue
|
|
|
|
author = get_author_from_isni(
|
|
get_element_text(element.find(".//isniUnformatted"))
|
|
)
|
|
if author is None:
|
|
continue
|
|
|
|
if bool(description):
|
|
|
|
titles = []
|
|
# prefer title records from LoC+ coop, Australia, Ireland, or Singapore
|
|
# in that order
|
|
for source in ["LCNACO", "NLA", "N6I", "NLB"]:
|
|
for parent in element.findall(f'.//titleOfWork/[@source="{source}"]'):
|
|
titles.append(parent.find(".//title"))
|
|
for parent in element.findall(f'.//titleOfWork[@subsource="{source}"]'):
|
|
titles.append(parent.find(".//title"))
|
|
# otherwise just grab the first title listing
|
|
titles.append(element.find(".//title"))
|
|
|
|
if titles:
|
|
# some of the "titles" in ISNI are a little ...iffy
|
|
# @ is used by ISNI/OCLC to index the starting point ignoring stop words
|
|
# (e.g. "The @Government of no one")
|
|
author.bio = ""
|
|
for title in titles:
|
|
if (
|
|
title is not None
|
|
and hasattr(title, "text")
|
|
and title.text is not None
|
|
and not title.text.replace("@", "").isnumeric()
|
|
):
|
|
author.bio = title.text.replace("@", "")
|
|
break
|
|
|
|
possible_authors.append(author)
|
|
|
|
return possible_authors
|
|
|
|
|
|
def get_author_from_isni(isni: str) -> Optional[activitypub.Author]:
|
|
"""Find data to populate a new author record from their ISNI"""
|
|
|
|
payload = request_isni_data("pica.isn", isni)
|
|
# parse xml
|
|
root = ET.fromstring(payload)
|
|
# there should only be a single responseRecord
|
|
# but let's use the first one just in case
|
|
element = root.find(".//responseRecord")
|
|
if element is None:
|
|
return None
|
|
|
|
name = (
|
|
make_name_string(forename)
|
|
if (forename := element.find(".//forename/..")) is not None
|
|
else ""
|
|
)
|
|
viaf = get_other_identifier(element, "viaf")
|
|
# use a set to dedupe aliases in ISNI
|
|
aliases = set()
|
|
aliases_element = element.findall(".//personalNameVariant")
|
|
for entry in aliases_element:
|
|
aliases.add(make_name_string(entry))
|
|
bio = get_element_text(element.find(".//nameTitle"))
|
|
wikipedia = get_external_information_uri(element, "Wikipedia")
|
|
|
|
author = activitypub.Author(
|
|
id=get_element_text(element.find(".//isniURI")),
|
|
name=name,
|
|
isni=isni,
|
|
viafId=viaf,
|
|
# aliases needs to be list not set
|
|
aliases=list(aliases),
|
|
bio=bio,
|
|
wikipediaLink=wikipedia,
|
|
)
|
|
|
|
return author
|
|
|
|
|
|
def build_author_from_isni(match_value: str) -> dict[str, activitypub.Author]:
|
|
"""Build basic author class object from ISNI URL"""
|
|
|
|
# if it is an isni value get the data
|
|
if match_value.startswith("https://isni.org/isni/"):
|
|
isni = match_value.replace("https://isni.org/isni/", "")
|
|
author = get_author_from_isni(isni)
|
|
if author is not None:
|
|
return {"author": author}
|
|
# otherwise it's a name string
|
|
return {}
|
|
|
|
|
|
def augment_author_metadata(author: models.Author, isni: str) -> None:
|
|
"""Update any missing author fields from ISNI data"""
|
|
|
|
isni_author = get_author_from_isni(isni)
|
|
if isni_author is None:
|
|
return
|
|
|
|
isni_author.to_model(model=models.Author, instance=author, overwrite=False)
|
|
|
|
# we DO want to overwrite aliases because we're adding them to the
|
|
# existing aliases and ISNI will usually have more.
|
|
# We need to dedupe because ISNI records often have lots of dupe aliases
|
|
aliases = set(isni_author.aliases)
|
|
for alias in author.aliases:
|
|
aliases.add(alias)
|
|
author.aliases = list(aliases)
|
|
author.save()
|