Update apply_guardrail to allow tool call deletion
#24262Comparing
predibase:apply_guardrail_deletion
(c22b17d
) withmain
(b64b0d4
)0%
- Untouched
- 16
Benchmarks
16 total
test_token_counter_long_content
tests/benchmarks/test_benchmarks.py+1%
1.7 ms1.7 ms
test_token_counter_simple_message
tests/benchmarks/test_benchmarks.py+1%
240.9 µs239.5 µs
test_get_llm_provider_azure
tests/benchmarks/test_benchmarks.py0%
148.9 µs148.7 µs
test_token_counter_with_tools
tests/benchmarks/test_benchmarks.py0%
417.3 µs417.6 µs
test_get_model_info_anthropic
tests/benchmarks/test_benchmarks.py0%
83.6 µs83.8 µs
test_get_model_cost_key_exact_match
tests/benchmarks/test_benchmarks.py0%
84.8 µs85.1 µs
test_get_llm_provider_openai
tests/benchmarks/test_benchmarks.py0%
142.7 µs143.2 µs
test_token_counter_raw_text
tests/benchmarks/test_benchmarks.py0%
186.9 µs187.6 µs
test_token_counter_multi_turn
tests/benchmarks/test_benchmarks.py-1%
570.5 µs573.4 µs
test_get_model_info_with_provider
tests/benchmarks/test_benchmarks.py-1%
85.2 µs85.7 µs
test_get_model_cost_key_case_insensitive
tests/benchmarks/test_benchmarks.py-1%
87.1 µs87.6 µs
test_cost_per_token_openai
tests/benchmarks/test_benchmarks.py-1%
542.1 µs545.5 µs
test_get_llm_provider_with_prefix
tests/benchmarks/test_benchmarks.py-1%
141.6 µs142.7 µs
test_get_llm_provider_anthropic
tests/benchmarks/test_benchmarks.py-1%
146.1 µs147.5 µs
test_cost_per_token_anthropic
tests/benchmarks/test_benchmarks.py-1%
541.2 µs546.3 µs
test_get_model_info_openai
tests/benchmarks/test_benchmarks.py-1%
86.3 µs87.4 µs
Commits
Click on a commit to change the comparison rangeBase
main
b64b0d4
-0.42%
Small style fixes
3394bc3
1 day ago
by seph-barker+0.07%
Small style fixes
c1ab361
1 day ago
by seph-barker+0.04%
feat: add guardrail_handles_tool_calls opt-in flag for tool call processing
308795e
24 hours ago
by seph-barker-0.25%
fix: address PR review feedback — fastapi import, streaming path, metadata stripping
f75e16f
23 hours ago
by seph-barker+0.01%
fix: gate guardrail_handles_tool_calls on apply-back only, not inspection
97490a4
23 hours ago
by seph-barker-0.06%
chore: re-export GUARDRAIL_DELETED_KEY from custom_guardrail module
e3cd998
20 hours ago
by seph-barker+0.1%
fix: use typed objects for injected message output items
15697f1
20 hours ago
by seph-barker+0.1%
fix: use unique IDs for injected guardrail message output items
a349ba8
19 hours ago
by seph-barker-0.1%
fix: suppress mypy arg-type for GenericResponseOutputItem append
2df5afd
19 hours ago
by seph-barker-0.1%
fix: move type ignore to correct line for mypy, remove unused loop var
3c2c49d
19 hours ago
by seph-barker+0.19%
fix: add isinstance guards for tool call dict access in OpenAI Chat handler
d798b05
19 hours ago
by seph-barker-0.08%
style: apply Black formatting
8715785
19 hours ago
by seph-barker+0.15%
fix: Anthropic input IndexError, invalid assistant message, multi-text injection
c22b17d
8 hours ago
by seph-barker