Merged
Conversation
Always display and save all chunker configuration parameters (chunk_size, overlap_size, method, threshold, pattern) regardless of default chunker selection, since documents may use different chunkers.
… cleanup, and chunker UI improvements LLM Config: - Refactor test_llm_config to use get_xxx_config() getters + _build_test_config() overlay pattern instead of manual config assembly with _create_llm_service() - Rewrite _unmask_auth to resolve real credentials via getters instead of comparing against static llm_config dict - Fix Bedrock multimodal test: replace 1x1 test PNG (rejected by Bedrock image validation) with 20x20 PNG - Fix provider-aware image format in multimodal test and image_data_extractor - Remove hardcoded DEFAULT_MULTIMODAL_MODELS — inherit now uses completion model - Add get_embedding_config() to common/config.py for parity with other getters - Migrate all consumers to get_xxx_config(graphname) (root.py, report-service, ecc/main.py, ui.py); fix report-service bug where llm_config was never imported - Strip null service values before reload (null = inherit, key should be absent) - LLMConfig UI: add inherit checkbox + amber warning for multimodal, reorder sections, red asterisks on required fields, shared LLM_PROVIDERS constant, sync state on single/multi-provider toggle - Rename agent_generation.py params for clarity (llm_model -> llm_service) GraphRAG Config & Chunkers: - Show all chunker settings in GraphRAG config page regardless of selected chunker - Use empty defaults instead of hardcoded values for chunker_config fields - Track loaded config via useRef to only save changed values - Fix boolean field defaults (use ?? instead of || for false-valued fields) - Update README and CHANGELOG
- Add resolve_llm_services() for shared config resolution across save and test - Unify buildLLMConfigPayload() in frontend for both save and test connection - Promote auth and region_name from completion_service to top-level config - Strip redundant per-service auth/region_name copies on save - Graph-specific config stores delta only (resolved diff against global) - Eliminate duplicate single/multi-provider state in LLMConfig.tsx - Inherit checkbox is a pure toggle; model names persist when toggling
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
1 participant
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.
PR Type
Enhancement, Bug fix, Tests, Documentation
Description
Unify LLM service resolution paths
Normalize auth and region inheritance
Overhaul LLM UI; inherit multimodal
Standardize chunker defaults, parameter names
Diagram Walkthrough
File Walkthrough
14 files
Unify LLM save/test; normalize and resolve servicesAdd resolver, embedding getter, region inheritanceDefault sizes; overlap param renamed and autoDefault chunk size and auto overlapUse completion getter for logging detailsAdapt chunker params to new schemaChunker construction with new defaultsRename constructor arg; clarify token configHealth and root use config gettersDefault chunk size and auto overlapDefault to 2048; auto-overlap; runtime checksRoot endpoint returns model via getterRefactor UI, shared providers, unified payloadsShow all chunker settings; save deltas1 files
Update tests for defaults and validations1 files
Provider-aware multimodal image content formatting2 files
Document config, UI, and chunker updatesUpdate docs for LLM and chunkers