mirror/nanobot
1
0
Fork 0
mirror of https://github.com/HKUDS/nanobot.git synced 2026-04-26 12:55:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(providers): normalize anthropic cached token usage

Browse Source
This commit is contained in:
Xubin Ren 2026-04-02 04:48:11 +00:00
parent 4ce0a8a68e
commit feb04a046d
2 changed files with 10 additions and 5 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

9
nanobot/providers/anthropic_provider.py
View File

@ -370,17 +370,20 @@ class AnthropicProvider(LLMProvider):
usage: dict[str, int] = {} usage: dict[str, int] = {}
if response.usage: if response.usage:
input_tokens = response.usage.input_tokens
cache_creation = getattr(response.usage, "cache_creation_input_tokens", 0) or 0
cache_read = getattr(response.usage, "cache_read_input_tokens", 0) or 0
total_prompt_tokens = input_tokens + cache_creation + cache_read
usage = { usage = {
"prompt_tokens": response.usage.input_tokens, "prompt_tokens": total_prompt_tokens,
"completion_tokens": response.usage.output_tokens, "completion_tokens": response.usage.output_tokens,
"total_tokens": response.usage.input_tokens + response.usage.output_tokens, "total_tokens": total_prompt_tokens + response.usage.output_tokens,
} }
for attr in ("cache_creation_input_tokens", "cache_read_input_tokens"): for attr in ("cache_creation_input_tokens", "cache_read_input_tokens"):
val = getattr(response.usage, attr, 0) val = getattr(response.usage, attr, 0)
if val: if val:
usage[attr] = val usage[attr] = val
# Normalize to cached_tokens for downstream consistency. # Normalize to cached_tokens for downstream consistency.
cache_read = usage.get("cache_read_input_tokens", 0)
if cache_read: if cache_read:
usage["cached_tokens"] = cache_read usage["cached_tokens"] = cache_read

6
tests/providers/test_cached_tokens.py
View File

@ -198,7 +198,7 @@ def test_anthropic_maps_cache_fields_to_cached_tokens():
usage_obj = FakeUsage( usage_obj = FakeUsage(
input_tokens=800, input_tokens=800,
output_tokens=200, output_tokens=200,
cache_creation_input_tokens=0, cache_creation_input_tokens=300,
cache_read_input_tokens=1200, cache_read_input_tokens=1200,
) )
content_block = FakeUsage(type="text", text="hello") content_block = FakeUsage(type="text", text="hello")
@ -211,7 +211,9 @@ def test_anthropic_maps_cache_fields_to_cached_tokens():
) )
result = AnthropicProvider._parse_response(response) result = AnthropicProvider._parse_response(response)
assert result.usage["cached_tokens"] == 1200 assert result.usage["cached_tokens"] == 1200
assert result.usage["prompt_tokens"] == 800 assert result.usage["prompt_tokens"] == 2300
assert result.usage["total_tokens"] == 2500
assert result.usage["cache_creation_input_tokens"] == 300
def test_anthropic_no_cache_fields(): def test_anthropic_no_cache_fields():