mirror of https://github.com/All-Hands-AI/OpenHands.git synced 2026-04-29 03:00:45 -04:00

Files

openhands 6649be08a7 Fix circular import issue by extracting RecallType to standalone module

- Created new module openhands/events/recall_type.py with RecallType enum
- Removed RecallType from openhands/events/event.py to break circular dependency
- Updated all import statements across 13 files to use new module path
- Resolves circular import chain: sync/enrich_user_interaction_data.py ->
  integrations.github.data_collector -> ... -> openhands.events.event ->
  openhands.llm.metrics -> ... -> storage.conversation_callback ->
  openhands.events.observation.agent -> openhands.events.event (circular)

The RecallType enum now has minimal dependencies and can be imported
without triggering the heavy dependency chain that caused the circular import.

Co-authored-by: openhands <openhands@all-hands.dev>

2025-12-10 00:46:31 +00:00

condenser

Optimize the condense conditions of the condenser (#11332 )

2025-10-26 11:23:22 -04:00

__init__.py

fix: Context window truncation using CondensationAction (#7578 )

2025-03-31 13:47:00 -06:00

conversation_memory.py

Fix circular import issue by extracting RecallType to standalone module

2025-12-10 00:46:31 +00:00

memory.py

Fix circular import issue by extracting RecallType to standalone module

2025-12-10 00:46:31 +00:00

README.md

remove llamaindex (#7151 )

2025-03-11 22:28:56 +00:00

view.py

Condensation request signal in event stream (#9097 )

2025-06-27 09:57:39 -06:00

README.md

Memory Component

Short Term History
Memory Condenser

Short Term History

Short term history filters the event stream and computes the messages that are injected into the context
It filters out certain events of no interest for the Agent, such as AgentChangeStateObservation or NullAction/NullObservation
When the context window or the token limit set by the user is exceeded, history starts condensing: chunks of messages into summaries.
Each summary is then injected into the context, in the place of the respective chunk it summarizes

Memory Condenser

Memory condenser is responsible for summarizing the chunks of events
It summarizes the earlier events first
It starts with the earliest agent actions and observations between two user messages
Then it does the same for later chunks of events between user messages
If there are no more agent events, it summarizes the user messages, this time one by one, if they're large enough and not immediately after an AgentFinishAction event (we assume those are tasks, potentially important)
Summaries are retrieved from the LLM as AgentSummarizeAction, and are saved in State.