added config.py

feat(agent): rag_search returns structured JSON, agent synthesizes cited report

Revert "fix(agent): pass topic keyword not meta-request to rag_search"

fix(agent): pass topic keyword not meta-request to rag_search

fix(agent): forbid multi-call rag_search loop per article

fix(agent,web): fix citation routing and sources rendering

fix: replace BM25 coverage counting with reranker-based pipeline

refactor(coverage): replace hybrid+rerank count with BM25-only keyword search

feat(agent): surface momentum_percentile, label, coverage in trending_topics

feat(agent): remove scope guardrail from system prompt

feat(agent): generic system prompt, 10s estimate_gap cap, scope guardrail

fix(signals): guard zero baseline only on denominator

fix(topic_index): use query_points not search (client API)

refactor(agent): semantic topic index, label-free gaps, decluttered prompt

Add Langfuse observability via LangChain CallbackHandler

feat(pipeline): RAG quality gate, live coverage, streaming API, recursion limit

fix(agent): prompt routing + determinism guard + gap table

fix(agent): rank gaps by score, fuzzy topic filter, stream in web API

feat(observability): INFO-level logging across all agent components

feat(agent): fix broken tools, add coverage_lookup, rewrite system prompt

refactor(pipeline): metadata-driven coverage, per-topic scroll, pure gap fn

feat(kg): COVERED_BY edges, wipe script, scroll_all, docs_for_topic

feat(topics): mine 667 GDELT WB-taxonomy topics + orgs for 2026-06-14

refactor(signals): 180-day weekly horizon, negative velocity, no cache

feat(gap): Phase 1 — Gap v2 (Demand×(1−Coverage)) + anchor-normalized signals

perf: block huggingface network checks on startup

feat(kb): Solid Gartner KB — cross-industry Wikidata seed, Wikidata-native relationships, 7-day signal cache

feat(agent): implement comprehensive logging and fix recursion loop

feat(agent): implement production-ready LangGraph ReAct agent

feat(graph): implement lean Neo4j Knowledge Graph build with spaCy NER

feat(gap): implement ultra-lean GapScore engine and reporting

feat(signals): implement csv-driven topic seeding and cache relocation

feat(signals): implement external trends signals with cache-aside TTL

feat(rag): implement cited RAG answer generation

feat(retrieval): implement hybrid search with cross-encoder reranking

feat(p2): implement data ingestion pipeline with dense and BM25 sparse vectors

fix(phase-1): switch pyproject.toml to setuptools.build_meta for compatibility

feat(phase-1): bootstrap project scaffold and connectivity layer