Skip to content

fix(repl_env): handle nested parens in FINAL(...) extraction #653

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

Open
kashif wants to merge 5 commits into
base: main
Choose a base branch
from
Open

fix(repl_env): handle nested parens in FINAL(...) extraction #653

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
257ae85
fix(repl_env): handle nested parens in FINAL(...) extraction
kashif May 11, 2026
936d030
Merge branch 'main' into fix/repl-env-final-regex-nested-parens
sergiopaniego May 11, 2026
27fb59f
Merge branch 'main' into fix/repl-env-final-regex-nested-parens
sergiopaniego May 11, 2026
52fa299
update based on reviews
sergiopaniego May 18, 2026
a06f55e
Merge branch 'main' into fix/repl-env-final-regex-nested-parens
sergiopaniego May 18, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 12 additions & 4 deletions envs/repl_env/runner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -243,10 +243,18 @@ def _default_answer(
]
try:
response = self._chat(final_prompt, model)
# Try to extract FINAL(...) from the response
match = re.search(r"FINAL\((.*?)\)", response, re.DOTALL)
if match:
return match.group(1).strip()
# Try to extract FINAL(...) from the response. Paren-counting handles
# nested parens and mid-sentence FINAL without regex flag trade-offs.
idx = response.find("FINAL(")
if idx != -1:
depth, start = 0, idx + len("FINAL")
for i, ch in enumerate(response[idx + len("FINAL"):], start=idx + len("FINAL")):
if ch == "(":
depth += 1
elif ch == ")":
depth -= 1
if depth == 0:
return response[start + 1 : i].strip()
# If no FINAL pattern, return the raw response as best-effort
return response.strip() if response.strip() else None
except Exception:
Expand Down
16 changes: 12 additions & 4 deletions envs/repl_env/server/repl_environment.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -536,10 +536,18 @@ def _extract_final_answer(self, stdout: str) -> Optional[str]:
Returns:
Final answer string or None if not found
"""
# Pattern 1: RLM-style FINAL(answer)
final_match = re.search(r"FINAL\((.*?)\)", stdout, re.DOTALL)
if final_match:
return final_match.group(1).strip()
# Pattern 1: RLM-style FINAL(answer). Paren-counting handles nested
# parens (e.g. FINAL(f(x))) and multi-line values without regex flag trade-offs.
idx = stdout.find("FINAL(")
if idx != -1:
depth, start = 0, idx + len("FINAL")
for i, ch in enumerate(stdout[idx + len("FINAL"):], start=idx + len("FINAL")):
if ch == "(":
depth += 1
elif ch == ")":
depth -= 1
if depth == 0:
return stdout[start + 1 : i].strip()

# Pattern 2: RLM-style FINAL_VAR(variable_name)
final_var_match = re.search(r"FINAL_VAR\((\w+)\)", stdout)
Expand Down
26 changes: 26 additions & 0 deletions tests/envs/test_repl_env.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -229,6 +229,32 @@ def test_final_pattern_basic(self):
assert obs.done
assert obs.metadata["final_answer"] == "42"

@pytest.mark.parametrize(
"code, expected",
[
# Nested function calls inside FINAL(...).
("print('FINAL(f(x))')", "f(x)"),
# Tuple as the final answer.
("print('FINAL((1, 2, 3))')", "(1, 2, 3)"),
# Math expression with multiple nested parens (e2b_repl_example).
(
"print('FINAL(2^(2^(2^(2))) = 65536)')",
"2^(2^(2^(2))) = 65536",
),
# Dict containing a tuple value.
("print(\"FINAL({'a': (1, 2)})\")", "{'a': (1, 2)}"),
# Output after FINAL must not bleed into the extracted answer.
("print('FINAL(42)\\nresult: (ok)')", "42"),
],
)
def test_final_pattern_nested_parentheses(self, code, expected):
"""FINAL(...) extraction must handle nested parentheses (rlm #75)."""
env = REPLEnvironment()
env.reset()
obs = env.step(REPLAction(code=code))
assert obs.done
assert obs.metadata["final_answer"] == expected
Comment on lines +250 to +256
Copy link
Copy Markdown
Contributor

@greptile-apps greptile-apps Bot May 11, 2026

Choose a reason for hiding this comment

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

P2 Test suite doesn't cover the cross-line over-match regression

All parametrized cases print FINAL(...) as an isolated single-line string, so they pass with either the current (DOTALL) or the corrected (no DOTALL) regex. Adding one case where stdout contains a ) on a subsequent line — e.g. print('FINAL(42)\nresult: (ok)') expecting "42" — would have caught the DOTALL bug and will guard against regressions.

Prompt To Fix With AI 
This is a comment left during a code review.
Path: tests/envs/test_repl_env.py
Line: 248-254

Comment:
**Test suite doesn't cover the cross-line over-match regression**

All parametrized cases print `FINAL(...)` as an isolated single-line string, so they pass with *either* the current (DOTALL) or the corrected (no DOTALL) regex. Adding one case where `stdout` contains a `)` on a subsequent line — e.g. `print('FINAL(42)\nresult: (ok)')` expecting `"42"` — would have caught the DOTALL bug and will guard against regressions.

How can I resolve this? If you propose a fix, please make it concise.

Copy link
Copy Markdown
Member

@sergiopaniego sergiopaniego May 18, 2026

Choose a reason for hiding this comment

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

updated


def test_final_var_pattern(self):
"""Test FINAL_VAR() pattern."""
env = REPLEnvironment()
Expand Down