fix: switch to Qwen2.5-7B-Instruct with provider=auto — confirmed available on HF

fix: switch provider from hf-inference to nebius — hf-inference dropped large LLM support (July 2025 per docs)

fix: switch HF model to mistral-7b

fix: update HuggingFace model ID

updates

updates

updates

feat: deduplicate knowledge graph nodes by source URL

updates

feat: Nuclear visibility fix (Black on White) + Phase 2 features (Radar table, Feed sorting)

feat: refocus Strategic Advisor on Fendt/AGCO marketing & sales

fix: start scheduler at module load so gunicorn workers run the scout

merge: resolve conflicts with origin/main — keep improved colours and layout

fix: switch LLM provider cerebras->novita, harden empty-response, fix PDF encoding

fix: remove stale Gemini/ChromaDB refs, improve chart contrast, harden error handling

Stability: Fix duplicate Dash output error and add graceful Astra DB connection handling

Fix: UI unreadable text, Knowledge Graph errors, and PDF auto-download

fix: resolve runtime bugs, stale references, and config drift

test: add test_graph_engine, test_agents, test_summary_engine, test_integration_pipeline

fix(tests): repair broken pipeline tests and delete stale q2/ suite

feat: production-grade Knowledge Graph refactor

fix: switch HuggingFace provider to cerebras + Llama-3.1-8B-Instruct

fix: restore missing stats keys and update astrapy 2.x API

feat: multi-agent relational brain, triples ontology, dict_to_string

feat: migrate vector store from ChromaDB to Astra DB + Astra Vectorize

feat: model unification, node truncation, generate brief button

fix(pipeline): sanitise LLM JSON output, gate short inputs

fix(pipeline): switch to InferenceClient.chat_completion, drop LangChain

chore: add core/ modules, archive legacy code, update gitignore

refactor: replace Gemini with HuggingFace/LangGraph stack