The returned wiki url in response object by calling analyze_entities references to Non-English Wikipedia page
FrederickXZhang opened this issue · 1 comments
Hello,
The returned wiki url in response object by calling analyze_entities references to Non-English Wikipedia page. I think this method functioned properly before, i.e., always referencing to English Wikipedia page. However, it starts returning non-English wiki URLs right now. Please check details below.
Environment details
- OS type and version: Ubuntu 20.04.4 LTS
- Python version: python 3.9.7
- pip version: pip 22.0.3
google-cloud-language
version: google-cloud-language Version: 2.4.1
Steps to reproduce
from google.cloud import language_v1
os.environ["GOOGLE_APPLICATION_CREDENTIALS"] = "AUTHENTICATION PRIVATE KEY"
client = language_v1.LanguageServiceClient()
text_content = "USA"
type_ = language_v1.Document.Type.PLAIN_TEXT
language = "en"
document = {"content": text_content, "type_": type_, "language": language}
encoding_type = language_v1.EncodingType.UTF32
response = client.analyze_entities(request = {'document': document, 'encoding_type': encoding_type})
Code example
The returned response object correctly recognizes the correct entity (USA), but references to non-English wiki URL even though I have specified "en" for the language option.
entities {
name: "USA"
type_: LOCATION
metadata {
key: "mid"
value: "/m/09c7w0"
}
metadata {
key: "wikipedia_url"
value: "https://de.wikipedia.org/wiki/Vereinigte_Staaten"
}
salience: 1.0
mentions {
text {
content: "USA"
}
type_: PROPER
}
}
language: "en"
The error is also reproducible on your online demo. Please take a look at the url circled by me in the following screenshot.
Is there a way to revert the current release to an older version or revert your database?
Thanks!