Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

langchain[patch]: make BooleanOutputParser check words not substrings #20064

Merged
merged 2 commits into from
Apr 9, 2024
Merged

langchain[patch]: make BooleanOutputParser check words not substrings #20064

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
932d234
langchain[patch]: make BooleanOutputParser check words not substrings
casperdcl Apr 5, 2024
adf9807
x
eyurtsev Apr 9, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Jump to
Jump to file
Failed to load files.
Diff view
Diff view
34 changes: 19 additions & 15 deletions libs/langchain/langchain/output_parsers/boolean.py
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
import re

from langchain_core.output_parsers import BaseOutputParser


Expand All @@ -19,24 +21,26 @@ def parse(self, text: str) -> bool:
boolean

"""
cleaned_upper_text = text.strip().upper()
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I believe adding .split() is sufficient to make the required change?

cleaned_upper_text = text.strip().upper().split()

Copy link
Contributor Author

@casperdcl casperdcl Apr 7, 2024 •
edited

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Not really, as "I believe so (YES)" will .split() into [..., "(YES)"]. Meanwhile the re.findall will remove the surrounding punctuation i.e. [..., "YES"].

if (
self.true_val.upper() in cleaned_upper_text
and self.false_val.upper() in cleaned_upper_text
):
raise ValueError(
f"Ambiguous response. Both {self.true_val} and {self.false_val} in "
f"received: {text}."
truthy = {
val.upper() for val in re.findall(
rf"\b({self.true_val}|{self.false_val})\b",
text,
flags=re.IGNORECASE | re.MULTILINE
)
elif self.true_val.upper() in cleaned_upper_text:
}
if self.true_val.upper() in truthy:
if self.false_val.upper() in truthy:
raise ValueError(
f"Ambiguous response. Both {self.true_val} and {self.false_val} "
f"in received: {text}."
)
return True
elif self.false_val.upper() in cleaned_upper_text:
elif self.false_val.upper() in truthy:
return False
else:
raise ValueError(
f"BooleanOutputParser expected output value to include either "
f"{self.true_val} or {self.false_val}. Received {text}."
)
raise ValueError(
f"BooleanOutputParser expected output value to include either "
f"{self.true_val} or {self.false_val}. Received {text}."
)

@property
def _type(self) -> str:
Expand Down
4 changes: 4 additions & 0 deletions libs/langchain/tests/unit_tests/output_parsers/test_boolean_parser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -24,6 +24,10 @@ def test_boolean_output_parser_parse() -> None:
result = parser.parse("Not relevant (NO)")
assert result is False

# Test valid input
result = parser.parse("NOW this is relevant (YES)")
assert result is True

# Test ambiguous input
try:
parser.parse("yes and no")
Expand Down