Skip to content

add feature: config summarisation, context compression, caching #1198

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
Insomniac2904 wants to merge 3 commits into
base: main
Choose a base branch
from
+149 −2
Open

add feature: config summarisation, context compression, caching #1198

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
91c5e26
add feature: config summarisation, context compression, caching
Insomniac2904 Jan 11, 2026
b3398dd
adding feature for python files
Insomniac2904 Jan 14, 2026
1fb2b2b
Merge branch 'main' into feature/querydoc-summarization
Insomniac2904 Jan 14, 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
2 changes: 1 addition & 1 deletion go/api/v1alpha2/zz_generated.deepcopy.go
View file
Open in desktop

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

40 changes: 40 additions & 0 deletions go/config/crd/bases/kagent.dev_agents.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6060,6 +6060,46 @@ spec:
minItems: 1
type: array
type: object
context:
properties:
cache:
properties:
cacheIntervals:
type: integer
enabled:
type: boolean
minTokens:
type: integer
ttlSeconds:
type: integer
required:
- enabled
type: object
compression:
properties:
compactionInterval:
type: integer
enabled:
type: boolean
overlapSize:
type: integer
summarizer:
properties:
model:
description: Model to use for summarization (required
if type is llm)
type: string
type:
default: llm
enum:
- llm
- text
type: string
type: object
required:
- enabled
type: object
type: object
deployment:
properties:
affinity:
Expand Down
40 changes: 40 additions & 0 deletions helm/kagent-crds/templates/kagent.dev_agents.yaml
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -6060,6 +6060,46 @@ spec:
minItems: 1
type: array
type: object
context:
properties:
cache:
properties:
cacheIntervals:
type: integer
enabled:
type: boolean
minTokens:
type: integer
ttlSeconds:
type: integer
required:
- enabled
type: object
compression:
properties:
compactionInterval:
type: integer
enabled:
type: boolean
overlapSize:
type: integer
summarizer:
properties:
model:
description: Model to use for summarization (required
if type is llm)
type: string
type:
default: llm
enum:
- llm
- text
type: string
type: object
required:
- enabled
type: object
type: object
deployment:
properties:
affinity:
Expand Down
10 changes: 9 additions & 1 deletion python/packages/kagent-adk/src/kagent/adk/_a2a.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -55,6 +55,7 @@ def __init__(
app_name: str,
lifespan: Optional[Callable[[Any], Any]] = None,
plugins: List[BasePlugin] = None,
context_configs: Optional[Dict[str, Any]] = None,
stream: bool = False,
):
self.root_agent = root_agent
Expand All @@ -63,6 +64,7 @@ def __init__(
self.agent_card = agent_card
self._lifespan = lifespan
self.plugins = plugins if plugins is not None else []
self.context_configs = context_configs or {"compaction": None, "cache": None}
self.stream = stream

def build(self, local=False) -> FastAPI:
Expand All @@ -77,7 +79,13 @@ def build(self, local=False) -> FastAPI:
)
session_service = KAgentSessionService(http_client)

adk_app = App(name=self.app_name, root_agent=self.root_agent, plugins=self.plugins)
adk_app = App(
name=self.app_name,
root_agent=self.root_agent,
plugins=self.plugins,
events_compaction_config=self.context_configs.get("compaction"),
context_cache_config=self.context_configs.get("cache")
)

def create_runner() -> Runner:
return Runner(
Expand Down
59 changes: 59 additions & 0 deletions python/packages/kagent-adk/src/kagent/adk/types.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -5,6 +5,9 @@
from agentsts.adk import ADKTokenPropagationPlugin
from google.adk.agents import Agent
from google.adk.agents.base_agent import BaseAgent
from google.adk.apps.app import App, EventsCompactionConfig
from google.adk.apps.llm_event_summarizer import LlmEventSummarizer
from google.adk.agents.context_cache_config import ContextCacheConfig as AdkCacheConfig
from google.adk.agents.llm_agent import ToolUnion
from google.adk.agents.remote_a2a_agent import AGENT_CARD_WELL_KNOWN_PATH, DEFAULT_TIMEOUT, RemoteA2aAgent
from google.adk.code_executors.base_code_executor import BaseCodeExecutor
Expand Down Expand Up @@ -95,6 +98,60 @@ class Gemini(BaseLLM):
type: Literal["gemini"]


class ContextSummarizerSettings(BaseModel):
model: Optional[str] = None
type: Literal["llm", "text"] = "llm"

class ContextCompressionSettings(BaseModel):
enabled: bool
compaction_interval: Optional[int] = Field(None, alias="compactionInterval")
overlap_size: Optional[int] = Field(None, alias="overlapSize")
summarizer: Optional[ContextSummarizerSettings] = None

class ContextCacheSettings(BaseModel):
enabled: bool
min_tokens: Optional[int] = Field(None, alias="minTokens")
ttl_seconds: Optional[int] = Field(None, alias="ttlSeconds")
cache_intervals: Optional[int] = Field(None, alias="cacheIntervals")

class ContextConfig(BaseModel):
cache: Optional[ContextCacheSettings] = None
compression: Optional[ContextCompressionSettings] = None


def build_adk_context_configs(ctx: Optional[ContextConfig], default_llm: Any) -> Dict[str, Any]:
configs = {
"events_compaction_config": None,
"context_cache_config": None
}

if not ctx:
return configs

if ctx.compression and ctx.compression.enabled:
summarizer = None
if ctx.compression.summarizer and ctx.compression.summarizer.type == "llm":
# Use the provided model for summarization, or fall back to the main agent model
summerizer_model_name = ctx.compression.summarizer.model
summarizer_llm = GeminiLLM(model=summerizer_model_name) if summerizer_model_name else default_llm
summarizer = LlmEventSummarizer(llm=summarizer_llm)

configs["events_compaction_config"] = EventsCompactionConfig(
compaction_interval=ctx.compression.compaction_interval or 3,
overlap_size=ctx.compression.overlap_size or 1,
summarizer=summarizer,
)

if ctx.cache and ctx.cache.enabled:
configs["context_cache_config"] = AdkCacheConfig(
min_tokens=ctx.cache.min_tokens or 2048,
ttl_seconds=ctx.cache.ttl_seconds or 3600,
cache_intervals=ctx.cache.cache_intervals or 5,
)

return configs


class AgentConfig(BaseModel):
model: Union[OpenAI, Anthropic, GeminiVertexAI, GeminiAnthropic, Ollama, AzureOpenAI, Gemini] = Field(
discriminator="type"
Expand All @@ -105,6 +162,8 @@ class AgentConfig(BaseModel):
sse_tools: list[SseMcpServerConfig] | None = None # SSE MCP tools
remote_agents: list[RemoteAgentConfig] | None = None # remote agents
execute_code: bool | None = None
context: Optional[ContextConfig] = None

# This stream option refers to LLM response streaming, not A2A streaming
stream: bool | None = None

Expand Down