add: file attachments, run_id grouping, and invocation log UI improve…#555
Closed
shahharsh176 wants to merge 28 commits into
Closed
add: file attachments, run_id grouping, and invocation log UI improve…#555shahharsh176 wants to merge 28 commits into
shahharsh176 wants to merge 28 commits into
Conversation
- Add AppLLMMemorySession/AppLLMMemoryMessage models with
migrations for
per-session conversation history storage
- Add memory_enabled, memory_max_messages fields to AppLLMAgent
- Add session_id field to AppLLMInvocation for cross-round grouping
- AgentClient.run() now accepts input= (plain string) and
session_id=, loads
prior history and persists exchanges when memory_enabled
- Invocation log UI: improved filtering, session grouping, and
detail views
- Agent UI: expose memory settings in agent builder form
- API: new serializer fields and endpoints for memory/session
management
- Add __all__ and noqa comment to ai/tools/__init__.py for intentional re-exports - Remove unused datetime import from apps/ai/models/provider.py Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…pendency validation
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…ment - Add OutputParseError and OutputValidationError exceptions for structured output failures - Switch Anthropic provider to use native output_config JSON schema instead of prompt-engineering prefill - Add parsed_content field to LLMResponse for validated JSON output - Add provider logo SVGs (Anthropic, OpenAI, Azure, Bedrock) replacing colored initials in Agents UI - Add collapsible JSON schema viewer in agent detail row - Fix InvocationLogs to skip tool_result messages when identifying user input Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…, and invocation date filtering
- AppAi: add staleness-aware tab activation with per-tab STALE_AFTER thresholds (2-5 min), refreshSignals counter, lastFetchedAt tracking, and cross-tab invalidation via markStale/onInvalidate - InvocationLogs: manual Refresh button + "Updated X ago" indicator; tab activation only re-fetches invocations (not stats); stats refreshed only on manual refresh; searchable paginated agent filter dropdown replaces static select; removes upfront agents pre-fetch - Prompts: Refresh button in header; background re-fetch on activation without skeleton or filter reset - Agents: thread onInvalidate through AgentBuilder → PromptSelector → InlinePromptCreator so creating a prompt marks Prompts tab stale - Providers, Tools: accept refreshSignal/onFetchComplete; background re-fetch on tab activation - Fix: stabilise onFetchComplete refs with useCallback to prevent infinite fetch loop from inline lambda deps Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- Drop estimate_tokens() from all providers and base ABC (unused) - Remove compute_cost() in favour of compute_anthropic_cost only - Remove unused guardrails field from AppLLMAgent + serializers - Remove ProviderUsage, ProviderResetBudget, ProviderDependencies, and PromptDependencies views + URL patterns (dead API surface) - Remove tool_calls / tool_call_id columns from AppLLMMemoryMessage - Fix tool stats update to use F() expressions for atomicity - Add optional ModalFooter prop with sticky pinned footer to StandardModal - Sticky action buttons in Add/Edit provider wizard steps - Mark Bedrock and Azure OpenAI as "Coming Soon" in provider picker - Re-fetch providers on each agent-builder open for freshness - Rename "Avg Time" label to "Avg Time (24h)" in Tools list/detail - Register update_tool_usage_stats in Celery beat (daily at 2 AM) Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
…cation_logs Improve grouped invocation logs, use pagination size 10 always
Tool call and result messages are now saved to AppLLMMemoryMessage and
replayed verbatim on subsequent agent.run() calls, giving the LLM full
awareness of prior tool interactions within a session.
- Added memory_policy ("include"/"exclude") to @tool decorator and
AppLLMTool model — exclude drops side-effect tools from loaded history
- Migration 0002 adds memory_policy column to AppLLMTool
- sync_tools() syncs memory_policy from decorator to DB
- InvocationLogs UI: memory-injected tool calls no longer shown as live
- Tools UI: memory_policy badge added to tool metadata section
- Serializers expose memory_policy on list and detail endpoints
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
feat(ai): persist tool call/result rounds in session memory
…ent_json fix json schema copy for duplicate agent
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
…ments