How We Combined Vector Search, Knowledge Graphs, and Chain-of-Thought Reasoning to Build a Hallucination-Resistant Document Intelligence System We built a local-first RAG system that doesn't just retrieve documents it validates responses using semantic similarity, critiques its own answers before returning them, and runs hybrid vector+graph retrieval in parallel. This article breaks down the architecture, shows the code, and explains why these design decisions matter for production AI systems. Beyond the Hype: A Builder's Blueprint for RAG The AI space is flooded with articles and slick video demos about RAG. They're inspiring, but often leave you wondering, "Where's the code?" I'm a firm believer in the 'do it to learn it' philosophy. You don't truly understand a complex system until you've built it, broken it, and fixed it yourself. That's the spirit behind DocuChat. I built this project to move beyond the tutorials and create a tangible, scalable system that tackles real-world RAG challenges like hallucination and entity relationships. My goal was to share an open-source codebase that serves as a learning tool for the entire community one designed to be tinkered with, broken, and improved. This article is the blueprint to that system. Let's dive in. Table of Contents New to RAG? Start Here Why Another RAG System? System Architecture The Six Key Innovations Deep Dive: Implementation Details Performance & Benchmarks Use Cases & Customization Industry Trends & Differentiation Getting Started New to RAG? Start Here Why Another RAG System? System Architecture The Six Key Innovations Deep Dive: Implementation Details Performance & Benchmarks Use Cases & Customization Industry Trends & Differentiation Getting Started New to RAG? Start Here If you're new to the world of AI, terms like "RAG" and "Agentic Architecture" might seem complex. Here's a simple breakdown: Retrieval-Augmented Generation (RAG): Think of it as giving an AI a library to read before it answers your question. Instead of just using its pre-existing knowledge, it first retrieves relevant information from documents you provide and then generates an answer based on that specific context. This makes the AI's responses more accurate and grounded in your data. Retrieval-Augmented Generation (RAG): Agentic Architecture: This is like upgrading from a simple tool to a smart assistant. An "agent" is an AI system that can plan, reason, and use different tools to accomplish a goal. DocuChat isn't just a simple Q&A bot; it's an agent that analyzes your query, decides the best way to find the answer (using vector search, graph search, or both), and even critiques its own answer for accuracy before showing it to you. Agentic Architecture: DocuChat is the perfect sandbox to learn these concepts. It's designed to run entirely on your laptop for free, using efficient, small language models (like those under 1 billion parameters via Ollama) and state-of-the-art embedding models. You can look at the code, see how it works, and build a powerful, private document intelligence system without needing expensive cloud services. Why Another RAG System? The RAG (Retrieval-Augmented Generation) space is crowded. LangChain, LlamaIndex, Haystack excellent frameworks exist. But when we analyzed production requirements for document intelligence systems, we found three critical gaps: 1. Hallucination Detection is Mostly Missing Most RAG systems generate responses and return them immediately. There's no validation layer checking if the LLM just made something up. 2. Vector Search Alone Misses Entity Relationships "What companies did Microsoft acquire in 2023?" requires understanding relationships, not just semantic similarity. Pure vector search struggles here. 3. No Transparency into Reasoning Users get answers but don't see why the system chose those documents or how it reached that conclusion. DocuChat addresses these gaps with a production-ready, open-source implementation. System Architecture High-Level Component Diagram Suggested diagram content: A flowchart showing the ingestion pipeline on the left (Documents → Loader → Text Extraction → Entity Extraction with spaCy → Vector Embeddings with BAAI/bge-m3 + Knowledge Graph with Neo4j) and the query pipeline on the right (User Query → Query Analyzer → Reasoning Planner → Parallel Retriever [Vector + Graph] → Context Builder → Response Generator → Self-Critique → Confidence Scoring → Final Response). Use different colors for parallel processes and validation steps. Suggested diagram content: A flowchart showing the ingestion pipeline on the left (Documents → Loader → Text Extraction → Entity Extraction with spaCy → Vector Embeddings with BAAI/bge-m3 + Knowledge Graph with Neo4j) and the query pipeline on the right (User Query → Query Analyzer → Reasoning Planner → Parallel Retriever [Vector + Graph] → Context Builder → Response Generator → Self-Critique → Confidence Scoring → Final Response). Use different colors for parallel processes and validation steps. Data Flow: Query Processing Suggested diagram content: A sequence diagram showing interactions between User, Query Analyzer, Vector Store, Graph Store, Context Builder, LLM Generator, and Self-Critique components. Show parallel retrieval with simultaneous arrows, and highlight the self-critique loop with a different color. Suggested diagram content: A sequence diagram showing interactions between User, Query Analyzer, Vector Store, Graph Store, Context Builder, LLM Generator, and Self-Critique components. Show parallel retrieval with simultaneous arrows, and highlight the self-critique loop with a different color. The Six Key Innovations 1. Dual Storage Architecture (Hybrid Retrieval) Problem: Vector search finds semantically similar content but misses structured relationships. Graph search finds entity connections but may miss semantically related content. Problem: Solution: Run both in parallel and fuse results with cross-modal validation. Solution: Implementation: Implementation: 📄 View Code: rag_workflow.py (lines 387-458) View Code: View Code: rag_workflow.py The _parallel_retriever_node executes vector and graph retrieval simultaneously using asyncio.gather, maximizing throughput by running both operations concurrently. _parallel_retriever_node asyncio.gather Performance Impact: Performance Impact: Sequential: Vector (150ms) + Graph (120ms) = 270ms Parallel: max(150ms, 120ms) = 150ms (44% faster) Sequential: Vector (150ms) + Graph (120ms) = 270ms Parallel: max(150ms, 120ms) = 150ms (44% faster) Cross-Modal Validation: Cross-Modal Validation: 📄 View Code: context_builder.py (lines 527-583) View Code: View Code: context_builder.py The _calculate_hybrid_boost function boosts results appearing in BOTH vector and graph search. It analyzes content similarity, entity overlap, and source matching to assign confidence scores. High cross-modal agreement indicates the information appears in multiple retrieval modalities. _calculate_hybrid_boost Why This Matters: Results found through multiple retrieval methods have higher confidence they appear in both semantic space AND relationship graph. Why This Matters: 2. Semantic Grounding with Embeddings Problem: LLMs hallucinate. They generate plausible-sounding text that's not supported by retrieved documents. Problem: Solution: Calculate how much of the response is semantically grounded in the context using embedding-based similarity. Solution: Implementation: Implementation: 📄 View Code: response_generator.py (lines 1219-1333) View Code: View Code: response_generator.py The _calculate_semantic_grounding_score function performs sentence-level semantic similarity analysis: _calculate_semantic_grounding_score Splits the response into individual sentences Generates embeddings for each sentence using BAAI/bge-m3 Calculates cosine similarity against all context chunks Scores as: grounded_sentences / total_sentences Splits the response into individual sentences Generates embeddings for each sentence using BAAI/bge-m3 Calculates cosine similarity against all context chunks Scores as: grounded_sentences / total_sentences A sentence is considered "grounded" if it achieves 70%+ similarity with at least one context chunk. Fallback Strategy: If embeddings fail, falls back to phrase-based matching. Never breaks response generation. Fallback Strategy: 3. Chain-of-Thought Self-Critique Problem: LLMs should validate their own responses before returning them, especially for medical/legal domains. Problem: Solution: The LLM critiques its own response, rates it, and revises if below threshold. Solution: Implementation: Implementation: 📄 View Code: response_generator.py (lines 993-1089) View Code: View Code: response_generator.py The _self_critique_response function generates a self-critique and potentially revises the response. The LLM critiques its own output for: _self_critique_response Accuracy against provided context Completeness in addressing the query Potential fabricated details not supported by context Overall coherence and logical flow Accuracy against provided context Completeness in addressing the query Potential fabricated details not supported by context Overall coherence and logical flow If the self-assigned rating falls below 3/5, the response is automatically revised before being returned to the user. Production Impact: Catches errors before they reach users, especially critical for healthcare/legal applications. Production Impact: 4. Dynamic Confidence Thresholds Problem: A 70% grounding score means different things for different query types. Factual queries need higher standards than procedural ones. Problem: Solution: Adaptive thresholds based on query type and context availability. Solution: Implementation: Implementation: 📄 View Code: response_generator.py (lines 773-808) View Code: View Code: response_generator.py The _get_dynamic_grounding_threshold function adjusts confidence thresholds based on: _get_dynamic_grounding_threshold Query type: Factual queries require 0.4 threshold, procedural queries 0.25, comparative queries 0.35 Context availability: Fewer chunks (< 2) reduce expectations by 30%, while rich context (> 5) increases standards by 15% Query type: Factual queries require 0.4 threshold, procedural queries 0.25, comparative queries 0.35 Query type: Context availability: Fewer chunks (< 2) reduce expectations by 30%, while rich context (> 5) increases standards by 15% Context availability: Context-Aware Confidence: Context-Aware Confidence: 📄 View Code: response_generator.py (lines 835-951) View Code: View Code: response_generator.py The _generate_confidence_indicators function handles both sparse and rich context scenarios: _generate_confidence_indicators SPARSE CONTEXT: Uses response appropriateness (acknowledges limitations) RICH CONTEXT: Uses semantic grounding (document support) SPARSE CONTEXT: Uses response appropriateness (acknowledges limitations) SPARSE CONTEXT: RICH CONTEXT: Uses semantic grounding (document support) RICH CONTEXT: The system intelligently adapts its confidence calculation based on context quality. With sparse context, a good "I don't have that information" response receives high confidence, not low. Why This is Smart: The system adapts its confidence calculation based on context quality. With sparse context, a good "I don't have that information" response gets high confidence, not low. Why This is Smart: 5. Persona-Driven Response Adaptation Problem: Different users need different response styles. A financial analyst needs quantitative metrics and ROI calculations, while a product manager needs strategic insights and user impact analysis. Problem: Solution: Configurable personas that shape how the system interprets context, prioritizes information, and generates responses. Solution: Personas aren't just prompt wrappers they influence three critical stages of the pipeline: three critical stages Stage 1: Context Prioritization Different personas weight different entities and concepts during retrieval. 📄 View Code: context_builder.py - Persona Boost Logic View Code: View Code: context_builder.py The _apply_persona_boost function adjusts chunk scores based on persona-specific priorities. For example: _apply_persona_boost Financial Analyst: Boosts revenue (+40%), EBITDA (+40%), margins (+30%) Medical Professional: Boosts diagnosis (+50%), treatment (+40%), contraindications (+50%) Product Manager: Boosts user feedback (+40%), feature requests (+40%), market opportunities (+30%) Financial Analyst: Boosts revenue (+40%), EBITDA (+40%), margins (+30%) Financial Analyst: Medical Professional: Boosts diagnosis (+50%), treatment (+40%), contraindications (+50%) Medical Professional: Product Manager: Boosts user feedback (+40%), feature requests (+40%), market opportunities (+30%) Product Manager: Stage 2: Response Generation Style Personas define the tone, structure, and focus of responses. 📄 View Code: response_generator.py - Persona Prompts View Code: View Code: response_generator.py The system includes built-in personas for: Financial Analyst: Leads with quantitative metrics, references accounting standards, structures as Key Metrics → Analysis → Risk Factors Medical Professional: Prioritizes patient safety, structures as Diagnosis → Treatment → Contraindications, uses precise medical terminology Product Manager: Focuses on user impact and business value, structures as User Impact → Business Value → Action Items Legal Researcher: Emphasizes citations and precedents, structures as Legal Issue → Relevant Law → Analysis → Conclusion Financial Analyst: Leads with quantitative metrics, references accounting standards, structures as Key Metrics → Analysis → Risk Factors Financial Analyst: Medical Professional: Prioritizes patient safety, structures as Diagnosis → Treatment → Contraindications, uses precise medical terminology Medical Professional: Product Manager: Focuses on user impact and business value, structures as User Impact → Business Value → Action Items Product Manager: Legal Researcher: Emphasizes citations and precedents, structures as Legal Issue → Relevant Law → Analysis → Conclusion Legal Researcher: Stage 3: Confidence Threshold Adjustment High-stakes personas require stricter validation. 📄 View Code: response_generator.py - Persona Confidence Config View Code: View Code: response_generator.py The PERSONA_CONFIDENCE_CONFIG defines validation standards for each persona: PERSONA_CONFIDENCE_CONFIG Medical Professional: Base grounding 0.5 (vs 0.3 default), self-critique threshold 4.0/5, requires 2+ sources Legal Researcher: Base grounding 0.45, self-critique threshold 4.0/5, requires citations Financial Analyst: Base grounding 0.4, self-critique threshold 3.5/5, requires numerical evidence Default: Base grounding 0.3, self-critique threshold 3.0/5, single source acceptable Medical Professional: Base grounding 0.5 (vs 0.3 default), self-critique threshold 4.0/5, requires 2+ sources Medical Professional: Legal Researcher: Base grounding 0.45, self-critique threshold 4.0/5, requires citations Legal Researcher: Financial Analyst: Base grounding 0.4, self-critique threshold 3.5/5, requires numerical evidence Financial Analyst: Default: Base grounding 0.3, self-critique threshold 3.0/5, single source acceptable Default: Real-World Impact: Same Query, Different Personas Query: "What are the main findings in the Q4 2024 report?" Query: Financial Analyst Persona: Financial Analyst Persona: Q4 2024 FINANCIAL PERFORMANCE Key Metrics: • Revenue: $45.2M (+23% YoY, +8% QoQ) • EBITDA: $12.1M (26.8% margin, +340bps YoY) • Operating Cash Flow: $8.7M • Free Cash Flow: $6.2M Margin Analysis: • Gross Margin: 58.2% (-180bps YoY) due to increased COGS • Operating Margin: 22.1% (+210bps YoY) from SG&A efficiency • SG&A as % of revenue: 22% (improved from 25% in Q4 2023) Working Capital: • Accounts Receivable: +$2.1M (DSO increased to 47 days from 42) • Inventory: Stable at $3.2M Risk Factors: • Customer concentration: Top 3 clients = 47% of revenue • AR aging: $1.2M past 60 days (needs attention) Confidence: 0.89 ✅ Sources: Q4_2024_Financial_Report.pdf (pg. 3, 7, 12) Q4 2024 FINANCIAL PERFORMANCE Key Metrics: • Revenue: $45.2M (+23% YoY, +8% QoQ) • EBITDA: $12.1M (26.8% margin, +340bps YoY) • Operating Cash Flow: $8.7M • Free Cash Flow: $6.2M Margin Analysis: • Gross Margin: 58.2% (-180bps YoY) due to increased COGS • Operating Margin: 22.1% (+210bps YoY) from SG&A efficiency • SG&A as % of revenue: 22% (improved from 25% in Q4 2023) Working Capital: • Accounts Receivable: +$2.1M (DSO increased to 47 days from 42) • Inventory: Stable at $3.2M Risk Factors: • Customer concentration: Top 3 clients = 47% of revenue • AR aging: $1.2M past 60 days (needs attention) Confidence: 0.89 ✅ Sources: Q4_2024_Financial_Report.pdf (pg. 3, 7, 12) Product Manager Persona: Product Manager Persona: Q4 2024 STRATEGIC INSIGHTS Market Performance: • Enterprise segment: +34% growth → now 40% of revenue (strategic win) • New customer acquisition: +28% → strong product-market fit signal • Average contract value: +19% → successful premium tier positioning User Adoption & Satisfaction: • Feature X adoption: 67% of active users (launched Q3) • Customer satisfaction: 8.1/10 (up from 7.2/10 in Q3) • NPS Score: 58 (industry benchmark: 45) Competitive Position: • Won 3 strategic enterprise deals vs. Competitor A • Feature parity achieved in 2/3 competitive gaps identified in Q3 Action Items for Product Roadmap: 1. Double down on enterprise features (driving 34% growth) 2. Address top 3 customer pain points from feedback analysis 3. Expand Feature X capabilities based on 67% adoption success 4. Investigate premium tier features driving 19% ACV increase User Impact Priority: HIGH Business Value: $4.2M incremental revenue opportunity identified Confidence: 0.86 ✅ Sources: Q4_2024_Business_Review.pdf, Customer_Feedback_Analysis_Q4.xlsx Q4 2024 STRATEGIC INSIGHTS Market Performance: • Enterprise segment: +34% growth → now 40% of revenue (strategic win) • New customer acquisition: +28% → strong product-market fit signal • Average contract value: +19% → successful premium tier positioning User Adoption & Satisfaction: • Feature X adoption: 67% of active users (launched Q3) • Customer satisfaction: 8.1/10 (up from 7.2/10 in Q3) • NPS Score: 58 (industry benchmark: 45) Competitive Position: • Won 3 strategic enterprise deals vs. Competitor A • Feature parity achieved in 2/3 competitive gaps identified in Q3 Action Items for Product Roadmap: 1. Double down on enterprise features (driving 34% growth) 2. Address top 3 customer pain points from feedback analysis 3. Expand Feature X capabilities based on 67% adoption success 4. Investigate premium tier features driving 19% ACV increase User Impact Priority: HIGH Business Value: $4.2M incremental revenue opportunity identified Confidence: 0.86 ✅ Sources: Q4_2024_Business_Review.pdf, Customer_Feedback_Analysis_Q4.xlsx Medical Professional Persona: Medical Professional Persona: [For a medical case document] CLINICAL SUMMARY Diagnosis: • Primary: Type 2 Diabetes Mellitus (confirmed via HbA1c 8.2%, fasting glucose 167 mg/dL) • Secondary: Hypertension (BP 145/92 mmHg, sustained over 3 readings) Current Treatment Plan: • Metformin 500mg BID (initiated 01/15/2024) • Lisinopril 10mg QD for hypertension • Lifestyle modifications: dietary counseling, exercise plan Contraindications Noted: ⚠️ Patient allergic to sulfa drugs → avoid sulfonylureas ⚠️ eGFR 55 mL/min/1.73m² → Metformin dose appropriate, monitor renal function Clinical Evidence: • Treatment plan aligns with ADA 2024 guidelines (Level A evidence) • Blood pressure target <140/90 appropriate for this patient profile Missing Critical Information: ⚠️ No documented retinal exam (recommended annually for diabetics) ⚠️ Foot examination status not documented ⚠️ Last lipid panel date not mentioned Follow-up Required: • HbA1c recheck in 3 months • Renal function panel in 6 months (due to Metformin) • Ophthalmology referral for retinal screening Confidence: 0.91 ✅ (High - Clinical data well-documented) Sources: Patient_Chart_20240315.pdf, Lab_Results_20240315.pdf [For a medical case document] CLINICAL SUMMARY Diagnosis: • Primary: Type 2 Diabetes Mellitus (confirmed via HbA1c 8.2%, fasting glucose 167 mg/dL) • Secondary: Hypertension (BP 145/92 mmHg, sustained over 3 readings) Current Treatment Plan: • Metformin 500mg BID (initiated 01/15/2024) • Lisinopril 10mg QD for hypertension • Lifestyle modifications: dietary counseling, exercise plan Contraindications Noted: ⚠️ Patient allergic to sulfa drugs → avoid sulfonylureas ⚠️ eGFR 55 mL/min/1.73m² → Metformin dose appropriate, monitor renal function Clinical Evidence: • Treatment plan aligns with ADA 2024 guidelines (Level A evidence) • Blood pressure target <140/90 appropriate for this patient profile Missing Critical Information: ⚠️ No documented retinal exam (recommended annually for diabetics) ⚠️ Foot examination status not documented ⚠️ Last lipid panel date not mentioned Follow-up Required: • HbA1c recheck in 3 months • Renal function panel in 6 months (due to Metformin) • Ophthalmology referral for retinal screening Confidence: 0.91 ✅ (High - Clinical data well-documented) Sources: Patient_Chart_20240315.pdf, Lab_Results_20240315.pdf Why Personas Matter for Production Systems 1. Domain Expertise 1. Domain Expertise Medical, legal, and financial applications require domain-specific language, priorities, and validation standards. A general-purpose system can't meet these specialized needs. 2. User Segmentation 2. User Segmentation Different stakeholders view the same data differently: CFO needs financial metrics Product VP needs market insights CTO needs technical details CFO needs financial metrics Product VP needs market insights CTO needs technical details 3. Compliance & Risk Management 3. Compliance & Risk Management Regulated industries need: Specific disclaimers (medical: "consult a physician") Citation requirements (legal: statute references) Stricter validation (financial: auditable sources) Specific disclaimers (medical: "consult a physician") Citation requirements (legal: statute references) Stricter validation (financial: auditable sources) 4. Response Quality 4. Response Quality Persona-specific critique prompts catch domain errors: 📄 View Code: response_generator.py - Medical Critique Validation View Code: View Code: response_generator.py For medical personas, the critique prompt includes additional validation for unsupported medical claims, contraindications, side effects, and requires a 4/5 rating minimum for responses to pass. Using Personas in Production # CLI usage docuchat chat --persona financial_analyst docuchat chat --persona medical_professional # API usage response = docuchat_client.query( question="What are the Q4 findings?", persona="financial_analyst" ) # CLI usage docuchat chat --persona financial_analyst docuchat chat --persona medical_professional # API usage response = docuchat_client.query( question="What are the Q4 findings?", persona="financial_analyst" ) Creating Custom Personas 📄 View Example: examples/custom_persona.py View Example: View Example: examples/custom_persona.py The example demonstrates how to create a custom "security_analyst" persona by: Defining the persona prompt with specific instructions Setting confidence thresholds appropriate for security analysis Requiring multiple source documents for validation Defining the persona prompt with specific instructions Setting confidence thresholds appropriate for security analysis Requiring multiple source documents for validation Persona Impact Summary: Persona Impact Summary: Persona Context Boost Critique Threshold Min Confidence Default None 3.0/5 0.30 Financial Analyst Financial entities +40% 3.5/5 0.40 Medical Professional Clinical terms +50% 4.0/5 0.50 Legal Researcher Legal citations +40% 4.0/5 0.45 Product Manager User feedback +40% 3.0/5 0.35 Persona Context Boost Critique Threshold Min Confidence Default None 3.0/5 0.30 Financial Analyst Financial entities +40% 3.5/5 0.40 Medical Professional Clinical terms +50% 4.0/5 0.50 Legal Researcher Legal citations +40% 4.0/5 0.45 Product Manager User feedback +40% 3.0/5 0.35 Persona Context Boost Critique Threshold Min Confidence Persona Persona Context Boost Context Boost Critique Threshold Critique Threshold Min Confidence Min Confidence Default None 3.0/5 0.30 Default Default None None 3.0/5 3.0/5 0.30 0.30 Financial Analyst Financial entities +40% 3.5/5 0.40 Financial Analyst Financial Analyst Financial entities +40% Financial entities +40% 3.5/5 3.5/5 0.40 0.40 Medical Professional Clinical terms +50% 4.0/5 0.50 Medical Professional Medical Professional Clinical terms +50% Clinical terms +50% 4.0/5 4.0/5 0.50 0.50 Legal Researcher Legal citations +40% 4.0/5 0.45 Legal Researcher Legal Researcher Legal citations +40% Legal citations +40% 4.0/5 4.0/5 0.45 0.45 Product Manager User feedback +40% 3.0/5 0.35 Product Manager Product Manager User feedback +40% User feedback +40% 3.0/5 3.0/5 0.35 0.35 6. LangGraph Conditional Workflow Problem: Not all queries need the same processing path. Simple questions shouldn't go through expensive multi-step reasoning. Problem: Solution: Adaptive routing based on query complexity and persona requirements. Solution: Architecture: Architecture: 📄 View Code: rag_workflow.py (lines 210-263) View Code: View Code: rag_workflow.py The LangGraph workflow defines nodes for each processing stage and uses conditional edges to route queries based on complexity. The workflow supports multiple paths: Reasoning path: For complex, multi-step queries Vector-only path: For simple semantic searches Graph-only path: For entity-focused queries Parallel path: For hybrid retrieval combining both approaches Reasoning path: For complex, multi-step queries Reasoning path: Vector-only path: For simple semantic searches Vector-only path: Graph-only path: For entity-focused queries Graph-only path: Parallel path: For hybrid retrieval combining both approaches Parallel path: Routing Logic: Routing Logic: 📄 View Code: rag_workflow.py (lines 521-550) View Code: View Code: rag_workflow.py The _route_after_analysis function makes intelligent routing decisions based on: _route_after_analysis Query intent (explanation, comparison, procedural) Query complexity (word count, query type) Required retrieval methods (vector, graph, or both) Query intent (explanation, comparison, procedural) Query complexity (word count, query type) Required retrieval methods (vector, graph, or both) Simple queries bypass expensive reasoning nodes, reducing latency by approximately 40%. Performance Optimization: Simple queries bypass expensive reasoning nodes, reducing latency by ~40%. Performance Optimization: Deep Dive: Implementation Details Tech Stack Component Technology Reason Vector Store ChromaDB Local-first, 10M+ vectors supported Graph Database Neo4j Community Industry-standard graph queries Embeddings BAAI/bge-m3 SOTA multilingual embeddings (1024 dims) Entity Extraction spaCy (en_core_web_sm) Fast NER, preserves entity boundaries LLM Integration Ollama + Gemini API Local privacy + cloud performance options Workflow Engine LangGraph Conditional routing, state management CLI Framework Rich + Click Professional terminal UI Component Technology Reason Vector Store ChromaDB Local-first, 10M+ vectors supported Graph Database Neo4j Community Industry-standard graph queries Embeddings BAAI/bge-m3 SOTA multilingual embeddings (1024 dims) Entity Extraction spaCy (en_core_web_sm) Fast NER, preserves entity boundaries LLM Integration Ollama + Gemini API Local privacy + cloud performance options Workflow Engine LangGraph Conditional routing, state management CLI Framework Rich + Click Professional terminal UI Component Technology Reason Component Component Technology Technology Reason Reason Vector Store ChromaDB Local-first, 10M+ vectors supported Vector Store Vector Store ChromaDB ChromaDB Local-first, 10M+ vectors supported Local-first, 10M+ vectors supported Graph Database Neo4j Community Industry-standard graph queries Graph Database Graph Database Neo4j Community Neo4j Community Industry-standard graph queries Industry-standard graph queries Embeddings BAAI/bge-m3 SOTA multilingual embeddings (1024 dims) Embeddings Embeddings BAAI/bge-m3 BAAI/bge-m3 SOTA multilingual embeddings (1024 dims) SOTA multilingual embeddings (1024 dims) Entity Extraction spaCy (en_core_web_sm) Fast NER, preserves entity boundaries Entity Extraction Entity Extraction spaCy (en_core_web_sm) spaCy (en_core_web_sm) Fast NER, preserves entity boundaries Fast NER, preserves entity boundaries LLM Integration Ollama + Gemini API Local privacy + cloud performance options LLM Integration LLM Integration Ollama + Gemini API Ollama + Gemini API Local privacy + cloud performance options Local privacy + cloud performance options Workflow Engine LangGraph Conditional routing, state management Workflow Engine Workflow Engine LangGraph LangGraph Conditional routing, state management Conditional routing, state management CLI Framework Rich + Click Professional terminal UI CLI Framework CLI Framework Rich + Click Rich + Click Professional terminal UI Professional terminal UI Entity-Aware Chunking Standard chunking breaks text at token limits (e.g., every 512 tokens), often splitting entities mid-sentence. Our approach preserves entity boundaries: 📄 View Code: document_processor.py (lines 645-720) View Code: View Code: document_processor.py The _chunk_text_with_entity_awareness function: _chunk_text_with_entity_awareness Runs spaCy NER to identify all entities in the document Calculates chunk boundaries that avoid splitting entities Creates overlapping chunks (50 token overlap) with entity metadata preserved Runs spaCy NER to identify all entities in the document Calculates chunk boundaries that avoid splitting entities Creates overlapping chunks (50 token overlap) with entity metadata preserved Impact: Entity relationships remain intact, improving graph query accuracy by ~20%. Impact: Performance & Benchmarks Query Pipeline Breakdown (typical query) Stage Time Notes Query Analysis 50ms spaCy NER + intent classification Parallel Retrieval 150ms Vector + Graph (concurrent) Context Building 30ms Deduplication, ranking, fusion Response Generation 2000ms LLM-dependent (streaming) Confidence Calculation 80ms Semantic grounding analysis Total Latency ~2.3s Target: 2-5s Stage Time Notes Query Analysis 50ms spaCy NER + intent classification Parallel Retrieval 150ms Vector + Graph (concurrent) Context Building 30ms Deduplication, ranking, fusion Response Generation 2000ms LLM-dependent (streaming) Confidence Calculation 80ms Semantic grounding analysis Total Latency ~2.3s Target: 2-5s Stage Time Notes Stage Stage Time Time Notes Notes Query Analysis 50ms spaCy NER + intent classification Query Analysis Query Analysis 50ms 50ms spaCy NER + intent classification spaCy NER + intent classification Parallel Retrieval 150ms Vector + Graph (concurrent) Parallel Retrieval Parallel Retrieval 150ms 150ms Vector + Graph (concurrent) Vector + Graph (concurrent) Context Building 30ms Deduplication, ranking, fusion Context Building Context Building 30ms 30ms Deduplication, ranking, fusion Deduplication, ranking, fusion Response Generation 2000ms LLM-dependent (streaming) Response Generation Response Generation 2000ms 2000ms LLM-dependent (streaming) LLM-dependent (streaming) Confidence Calculation 80ms Semantic grounding analysis Confidence Calculation Confidence Calculation 80ms 80ms Semantic grounding analysis Semantic grounding analysis Total Latency ~2.3s Target: 2-5s Total Latency Total Latency Total Latency ~2.3s ~2.3s ~2.3s Target: 2-5s Target: 2-5s Memory Footprint Component Memory Base system 1.2GB Embedding model (BAAI/bge-m3) 2.0GB ChromaDB (10K chunks) 0.8GB Neo4j (10K nodes) 1.0GB Peak Usage ~5.2GB Component Memory Base system 1.2GB Embedding model (BAAI/bge-m3) 2.0GB ChromaDB (10K chunks) 0.8GB Neo4j (10K nodes) 1.0GB Peak Usage ~5.2GB Component Memory Component Component Memory Memory Base system 1.2GB Base system Base system 1.2GB 1.2GB Embedding model (BAAI/bge-m3) 2.0GB Embedding model (BAAI/bge-m3) Embedding model (BAAI/bge-m3) 2.0GB 2.0GB ChromaDB (10K chunks) 0.8GB ChromaDB (10K chunks) ChromaDB (10K chunks) 0.8GB 0.8GB Neo4j (10K nodes) 1.0GB Neo4j (10K nodes) Neo4j (10K nodes) 1.0GB 1.0GB Peak Usage ~5.2GB Peak Usage Peak Usage Peak Usage ~5.2GB ~5.2GB ~5.2GB Target Hardware: Intel i7-6500U (2015-era dual-core laptop) @ 5.5GB RAM Target Hardware: Use Cases & Customization 1. A Hands-On Learning Lab for AI Enthusiasts Audience: Anyone curious about building modern AI systems, from students to experienced engineers. Audience: DocuChat is more than just a tool; it's a transparent, running example of a sophisticated RAG system. Because it runs locally on your machine, you can interact with it, read the code, and see cause-and-effect in real-time. It's the best way to bridge the gap between theory and practice. What to study: What to study: /docuchat/agents/rag_workflow.py: See how a LangGraph-based agent makes decisions /docuchat/agents/nodes/context_builder.py: Understand how different sources of information (vectors and graphs) are fused together /docuchat/agents/nodes/response_generator.py: Look at the code that makes the AI critique its own answers to ensure accuracy /docuchat/agents/rag_workflow.py: See how a LangGraph-based agent makes decisions /docuchat/agents/rag_workflow.py /docuchat/agents/nodes/context_builder.py: Understand how different sources of information (vectors and graphs) are fused together /docuchat/agents/nodes/context_builder.py /docuchat/agents/nodes/response_generator.py: Look at the code that makes the AI critique its own answers to ensure accuracy /docuchat/agents/nodes/response_generator.py Example Exercise: Example Exercise: # Run in verbose mode to see the agent's thought process docuchat chat --verbose # Observe: # - How the agent analyzes your question # - The parallel retrieval in action # - The semantic grounding scores that prevent hallucinations # - The self-critique ratings where the AI grades itself # Run in verbose mode to see the agent's thought process docuchat chat --verbose # Observe: # - How the agent analyzes your question # - The parallel retrieval in action # - The semantic grounding scores that prevent hallucinations # - The self-critique ratings where the AI grades itself 2. Mid-Sized Enterprise Document Intelligence Use Case: A law firm with 50,000 legal documents or a company with an internal knowledge base. Use Case: DocuChat's local-first design provides the security needed for proprietary information. Its modular architecture allows it to be adapted for specialized enterprise needs. Scaling from Laptop to Enterprise: Scaling from Laptop to Enterprise: The same system you run on your laptop can be scaled for enterprise use. The key is swapping out components and expanding the data sources. Ingest Proprietary Data Securely: Since DocuChat is local-first, you can ingest confidential documents without them ever leaving your network. Ingest Proprietary Data Securely: Connect to Internal Knowledge Bases: Use the built-in URL ingestion feature to scrape and index internal websites, like a company wiki or documentation portal. Connect to Internal Knowledge Bases: # Index an internal Confluence or SharePoint page docuchat url http://internal-wiki.mycompany.com/important-docs # Index an internal Confluence or SharePoint page docuchat url http://internal-wiki.mycompany.com/important-docs Scale Your Models: Start with a small, local model (e.g., via Ollama) for development and then switch to a powerful, managed API (like Gemini or a private Azure OpenAI endpoint) for production by changing the configuration. Scale Your Models: Customize for Your Domain: Customize for Your Domain: Legal Entity Extraction: Replace the general-purpose NER with a model fine-tuned on legal text Citation Tracking: Extend the knowledge graph schema to link legal citations between documents Compliance Checks: Add a final validation node in the workflow to check responses against regulatory rules Legal Entity Extraction: Replace the general-purpose NER with a model fine-tuned on legal text Legal Entity Extraction: Citation Tracking: Extend the knowledge graph schema to link legal citations between documents Citation Tracking: Compliance Checks: Add a final validation node in the workflow to check responses against regulatory rules Compliance Checks: 3. Educational Institution Research Assistant Use Case: University library with 100K research papers. Use Case: Customizations: Customizations: Citation Graph: Build academic citation network in Neo4j Author Tracking: Link papers to authors, institutions Research Trends: Time-series analysis of topics Citation Graph: Build academic citation network in Neo4j Citation Graph: Author Tracking: Link papers to authors, institutions Author Tracking: Research Trends: Time-series analysis of topics Research Trends: Graph Schema Extension: Graph Schema Extension: // Neo4j schema for academic papers CREATE (p:Paper {title: "...", year: 2024, doi: "..."}) CREATE (a:Author {name: "...", institution: "..."}) CREATE (c:Concept {name: "machine learning"}) // Relationships CREATE (p)-[:WRITTEN_BY]->(a) CREATE (p)-[:CITES]->(other_paper) CREATE (p)-[:DISCUSSES]->(c) // Neo4j schema for academic papers CREATE (p:Paper {title: "...", year: 2024, doi: "..."}) CREATE (a:Author {name: "...", institution: "..."}) CREATE (c:Concept {name: "machine learning"}) // Relationships CREATE (p)-[:WRITTEN_BY]->(a) CREATE (p)-[:CITES]->(other_paper) CREATE (p)-[:DISCUSSES]->(c) 4. Healthcare Provider Clinical Documentation Use Case: Hospital with 500K patient records (HIPAA-compliant). Use Case: Why Local-First Matters: All data stays on-premises, no cloud API calls. Why Local-First Matters: Customizations: Customizations: Medical NER: Use BioBERT or ClinicalBERT for entity extraction SNOMED CT Integration: Map entities to medical ontologies Audit Trail: Log all queries for compliance Medical NER: Use BioBERT or ClinicalBERT for entity extraction Medical NER: SNOMED CT Integration: Map entities to medical ontologies SNOMED CT Integration: Audit Trail: Log all queries for compliance Audit Trail: Self-Critique for Medical Accuracy: Self-Critique for Medical Accuracy: 📄 View Code: response_generator.py - Medical Critique View Code: View Code: response_generator.py The system includes special validation for medical topics, paying attention to accuracy of medical claims, completeness of important medical information, and any unsupported medical advice. Ratings below 3/5 trigger automatic revision before the response is returned. Industry Trends & Differentiation RAG Evolution Timeline Suggested content: A horizontal timeline showing the evolution from 2020 (Simple RAG - Vector search only) → 2021 (Semantic search improvements) → 2022 (Multi-modal retrieval) → 2023 (Graph-augmented RAG, Self-RAG) → 2025 (Agentic RAG with validation, DocuChat) → 2026 (Future: Multi-agent collaboration) Suggested content: A horizontal timeline showing the evolution from 2020 (Simple RAG - Vector search only) → 2021 (Semantic search improvements) → 2022 (Multi-modal retrieval) → 2023 (Graph-augmented RAG, Self-RAG) → 2025 (Agentic RAG with validation, DocuChat) → 2026 (Future: Multi-agent collaboration) Competitive Landscape Feature DocuChat LangChain LlamaIndex Haystack Knowledge Graph ✅ Built-in Neo4j ❌ External ⚠️ Plugin ⚠️ Plugin Parallel Retrieval ✅ asyncio ❌ Sequential ⚠️ Custom ⚠️ Custom Self-Critique ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Semantic Grounding ✅ Sentence-level ❌ None ❌ None ❌ None Local-First ✅ Complete ⚠️ Partial ⚠️ Partial ⚠️ Partial Entity-Aware Chunking ✅ spaCy NER ❌ Generic ❌ Generic ⚠️ Plugin Adaptive Confidence ✅ Query-type based ❌ None ❌ None ❌ None Persona System ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Feature DocuChat LangChain LlamaIndex Haystack Knowledge Graph ✅ Built-in Neo4j ❌ External ⚠️ Plugin ⚠️ Plugin Parallel Retrieval ✅ asyncio ❌ Sequential ⚠️ Custom ⚠️ Custom Self-Critique ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Semantic Grounding ✅ Sentence-level ❌ None ❌ None ❌ None Local-First ✅ Complete ⚠️ Partial ⚠️ Partial ⚠️ Partial Entity-Aware Chunking ✅ spaCy NER ❌ Generic ❌ Generic ⚠️ Plugin Adaptive Confidence ✅ Query-type based ❌ None ❌ None ❌ None Persona System ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Feature DocuChat LangChain LlamaIndex Haystack Feature Feature DocuChat DocuChat LangChain LangChain LlamaIndex LlamaIndex Haystack Haystack Knowledge Graph ✅ Built-in Neo4j ❌ External ⚠️ Plugin ⚠️ Plugin Knowledge Graph Knowledge Graph ✅ Built-in Neo4j ✅ Built-in Neo4j ❌ External ❌ External ⚠️ Plugin ⚠️ Plugin ⚠️ Plugin ⚠️ Plugin Parallel Retrieval ✅ asyncio ❌ Sequential ⚠️ Custom ⚠️ Custom Parallel Retrieval Parallel Retrieval ✅ asyncio ✅ asyncio ❌ Sequential ❌ Sequential ⚠️ Custom ⚠️ Custom ⚠️ Custom ⚠️ Custom Self-Critique ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Self-Critique Self-Critique ✅ Built-in ✅ Built-in ❌ Manual ❌ Manual ❌ Manual ❌ Manual ❌ Manual ❌ Manual Semantic Grounding ✅ Sentence-level ❌ None ❌ None ❌ None Semantic Grounding Semantic Grounding ✅ Sentence-level ✅ Sentence-level ❌ None ❌ None ❌ None ❌ None ❌ None ❌ None Local-First ✅ Complete ⚠️ Partial ⚠️ Partial ⚠️ Partial Local-First Local-First ✅ Complete ✅ Complete ⚠️ Partial ⚠️ Partial ⚠️ Partial ⚠️ Partial ⚠️ Partial ⚠️ Partial Entity-Aware Chunking ✅ spaCy NER ❌ Generic ❌ Generic ⚠️ Plugin Entity-Aware Chunking Entity-Aware Chunking ✅ spaCy NER ✅ spaCy NER ❌ Generic ❌ Generic ❌ Generic ❌ Generic ⚠️ Plugin ⚠️ Plugin Adaptive Confidence ✅ Query-type based ❌ None ❌ None ❌ None Adaptive Confidence Adaptive Confidence ✅ Query-type based ✅ Query-type based ❌ None ❌ None ❌ None ❌ None ❌ None ❌ None Persona System ✅ Built-in ❌ Manual ❌ Manual ❌ Manual Persona System Persona System ✅ Built-in ✅ Built-in ❌ Manual ❌ Manual ❌ Manual ❌ Manual ❌ Manual ❌ Manual Market Positioning vs. LangChain: More opinionated, production-ready architecture (not a framework) vs. LlamaIndex: Focus on validation and transparency, not just retrieval vs. Haystack: Built-in graph integration, simpler deployment vs. GraphRAG: Fully implemented system, not research framework vs. LangChain: More opinionated, production-ready architecture (not a framework) vs. LangChain: vs. LlamaIndex: Focus on validation and transparency, not just retrieval vs. LlamaIndex: vs. Haystack: Built-in graph integration, simpler deployment vs. Haystack: vs. GraphRAG: Fully implemented system, not research framework vs. GraphRAG: Getting Started Installation # Prerequisites: Python 3.10+, Docker (for Neo4j), 8GB RAM # Note for beginners: While there are a few components, the setup is highly automated. # The system is designed to run efficiently on modern laptops without special hardware. # 1. Clone repository git clone https://github.com/[GITHUB_USERNAME]/docuchat-agent.git cd docuchat-agent # 2. Setup environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate pip install -e . # 3. Start Neo4j (Docker) docker run -d \ --name neo4j-docuchat \ -p 7474:7474 -p 7687:7687 \ -e NEO4J_AUTH=neo4j/password \ neo4j:5.12-community # 4. Initialize system docuchat status --check-services # Prerequisites: Python 3.10+, Docker (for Neo4j), 8GB RAM # Note for beginners: While there are a few components, the setup is highly automated. # The system is designed to run efficiently on modern laptops without special hardware. # 1. Clone repository git clone https://github.com/[GITHUB_USERNAME]/docuchat-agent.git cd docuchat-agent # 2. Setup environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate pip install -e . # 3. Start Neo4j (Docker) docker run -d \ --name neo4j-docuchat \ -p 7474:7474 -p 7687:7687 \ -e NEO4J_AUTH=neo4j/password \ neo4j:5.12-community # 4. Initialize system docuchat status --check-services Quick Start Example # 1. Ingest sample documents docuchat ingest ./examples/sample_docs # Expected output: # Processing: research_paper.pdf ✓ # Processing: technical_spec.docx ✓ # Processing: meeting_notes.md ✓ # # Results: # - 47 chunks created # - 152 entities extracted # - 3 documents indexed # Time: 8.2s # 2. Interactive chat with verbose mode docuchat chat --verbose # Example conversation: You: What machine learning models were mentioned? 🔍 VERBOSE: [RAGWorkflow] Executing query analysis... 🔍 VERBOSE: [QueryAnalyzer] Detected entities: ["machine learning", "models"] 🔍 VERBOSE: [QueryAnalyzer] Query type: factual 🔍 VERBOSE: [RAGWorkflow] Starting parallel retrieval... 🔍 VERBOSE: [RAGWorkflow] Vector search: 8 results (150ms) 🔍 VERBOSE: [RAGWorkflow] Graph search: 6 results (120ms) 🔍 VERBOSE: [ContextBuilder] Building context from 14 results... 🔍 VERBOSE: [ContextBuilder] After deduplication: 10 results 🔍 VERBOSE: [ContextBuilder] Hybrid boost applied: 3 cross-modal matches 🔍 VERBOSE: [ResponseGenerator] Generating response... 🔍 VERBOSE: [ResponseGenerator] Self-critique rating: 4/5 (passed) 🔍 VERBOSE: [ResponseGenerator] Grounding Analysis: 🔍 VERBOSE: [ResponseGenerator] - Method: Semantic similarity (embeddings) 🧠 🔍 VERBOSE: [ResponseGenerator] - Total sentences analyzed: 6 🔍 VERBOSE: [ResponseGenerator] - Semantically grounded sentences: 5 (83.3%) 🔍 VERBOSE: [ResponseGenerator] - Dynamic threshold: 0.40 (factual query) 🔍 VERBOSE: [ResponseGenerator] - Confidence level: HIGH ✅ The documents mention three machine learning models: 1. **Random Forest Classifier** (research_paper.pdf, p.12) - Used for binary classification with 87% accuracy - 100 estimators, max_depth=10 2. **LSTM Neural Network** (technical_spec.docx, section 3.2) - Sequence prediction model - 128 hidden units, dropout=0.2 3. **BERT-base-uncased** (research_paper.pdf, p.15) - Fine-tuned for sentiment analysis - 12 layers, 768 hidden dimensions --- Confidence: 0.833 --- # 1. Ingest sample documents docuchat ingest ./examples/sample_docs # Expected output: # Processing: research_paper.pdf ✓ # Processing: technical_spec.docx ✓ # Processing: meeting_notes.md ✓ # # Results: # - 47 chunks created # - 152 entities extracted # - 3 documents indexed # Time: 8.2s # 2. Interactive chat with verbose mode docuchat chat --verbose # Example conversation: You: What machine learning models were mentioned? 🔍 VERBOSE: [RAGWorkflow] Executing query analysis... 🔍 VERBOSE: [QueryAnalyzer] Detected entities: ["machine learning", "models"] 🔍 VERBOSE: [QueryAnalyzer] Query type: factual 🔍 VERBOSE: [RAGWorkflow] Starting parallel retrieval... 🔍 VERBOSE: [RAGWorkflow] Vector search: 8 results (150ms) 🔍 VERBOSE: [RAGWorkflow] Graph search: 6 results (120ms) 🔍 VERBOSE: [ContextBuilder] Building context from 14 results... 🔍 VERBOSE: [ContextBuilder] After deduplication: 10 results 🔍 VERBOSE: [ContextBuilder] Hybrid boost applied: 3 cross-modal matches 🔍 VERBOSE: [ResponseGenerator] Generating response... 🔍 VERBOSE: [ResponseGenerator] Self-critique rating: 4/5 (passed) 🔍 VERBOSE: [ResponseGenerator] Grounding Analysis: 🔍 VERBOSE: [ResponseGenerator] - Method: Semantic similarity (embeddings) 🧠 🔍 VERBOSE: [ResponseGenerator] - Total sentences analyzed: 6 🔍 VERBOSE: [ResponseGenerator] - Semantically grounded sentences: 5 (83.3%) 🔍 VERBOSE: [ResponseGenerator] - Dynamic threshold: 0.40 (factual query) 🔍 VERBOSE: [ResponseGenerator] - Confidence level: HIGH ✅ The documents mention three machine learning models: 1. **Random Forest Classifier** (research_paper.pdf, p.12) - Used for binary classification with 87% accuracy - 100 estimators, max_depth=10 2. **LSTM Neural Network** (technical_spec.docx, section 3.2) - Sequence prediction model - 128 hidden units, dropout=0.2 3. **BERT-base-uncased** (research_paper.pdf, p.15) - Fine-tuned for sentiment analysis - 12 layers, 768 hidden dimensions --- Confidence: 0.833 --- Project Structure for Contributors docuchat-agent/ ├── docuchat/ # Main application package │ ├── agents/ # LangGraph workflow nodes │ │ ├── rag_workflow.py # 🔥 Start here: Main orchestration │ │ └── nodes/ │ │ ├── query_analyzer.py # Intent & entity extraction │ │ ├── reasoning_planner.py # Q*-inspired planning │ │ ├── vector_retriever.py # ChromaDB integration │ │ ├── graph_retriever.py # Neo4j integration │ │ ├── context_builder.py # 🔥 Multi-modal fusion │ │ └── response_generator.py # 🔥 Self-critique & grounding │ │ │ ├── core/ # Business logic │ │ ├── document_processor.py # Entity-aware chunking │ │ ├── vector_store.py # ChromaDB wrapper │ │ ├── knowledge_graph.py # Neo4j wrapper │ │ └── shared_embedding_service.py # BAAI/bge-m3 embeddings │ │ │ ├── integrations/ # External service clients │ │ ├── ollama_client.py # Local LLM integration │ │ └── gemini_client.py # Gemini API integration │ │ │ └── cli/ # Command-line interface │ ├── chat.py # Interactive chat command │ └── ingest.py # Document ingestion command │ ├── docs/ # Documentation │ ├── architecture.md # Detailed system design │ ├── explanation_guide.md # Reasoning modes guide │ └── development.md # Development setup │ ├── tests/ # Integration tests │ └── test_rag_workflow.py │ └── examples/ # Sample code and documents └── custom_persona.py # How to add custom personas docuchat-agent/ ├── docuchat/ # Main application package │ ├── agents/ # LangGraph workflow nodes │ │ ├── rag_workflow.py # 🔥 Start here: Main orchestration │ │ └── nodes/ │ │ ├── query_analyzer.py # Intent & entity extraction │ │ ├── reasoning_planner.py # Q*-inspired planning │ │ ├── vector_retriever.py # ChromaDB integration │ │ ├── graph_retriever.py # Neo4j integration │ │ ├── context_builder.py # 🔥 Multi-modal fusion │ │ └── response_generator.py # 🔥 Self-critique & grounding │ │ │ ├── core/ # Business logic │ │ ├── document_processor.py # Entity-aware chunking │ │ ├── vector_store.py # ChromaDB wrapper │ │ ├── knowledge_graph.py # Neo4j wrapper │ │ └── shared_embedding_service.py # BAAI/bge-m3 embeddings │ │ │ ├── integrations/ # External service clients │ │ ├── ollama_client.py # Local LLM integration │ │ └── gemini_client.py # Gemini API integration │ │ │ └── cli/ # Command-line interface │ ├── chat.py # Interactive chat command │ └── ingest.py # Document ingestion command │ ├── docs/ # Documentation │ ├── architecture.md # Detailed system design │ ├── explanation_guide.md # Reasoning modes guide │ └── development.md # Development setup │ ├── tests/ # Integration tests │ └── test_rag_workflow.py │ └── examples/ # Sample code and documents └── custom_persona.py # How to add custom personas Key Files to Study (🔥): Key Files to Study (🔥): agents/rag_workflow.py - LangGraph workflow architecture agents/nodes/context_builder.py - Multi-modal result fusion agents/nodes/response_generator.py - Self-critique & grounding agents/rag_workflow.py - LangGraph workflow architecture agents/rag_workflow.py agents/nodes/context_builder.py - Multi-modal result fusion agents/nodes/context_builder.py agents/nodes/response_generator.py - Self-critique & grounding agents/nodes/response_generator.py Conclusion: Why This Architecture Matters For Learning This codebase demonstrates production RAG patterns that go beyond tutorials: Parallel retrieval with asyncio Semantic validation using embeddings Self-supervised quality control with LLM critique Adaptive confidence scoring based on query characteristics Persona-driven response customization Parallel retrieval with asyncio Semantic validation using embeddings Self-supervised quality control with LLM critique Adaptive confidence scoring based on query characteristics Persona-driven response customization For Production The architecture handles real-world concerns: Hallucination detection (semantic grounding) Confidence transparency (dynamic thresholds) Performance optimization (parallel retrieval, conditional routing) Privacy compliance (local-first processing) Domain adaptation (persona system) Hallucination detection (semantic grounding) Confidence transparency (dynamic thresholds) Performance optimization (parallel retrieval, conditional routing) Privacy compliance (local-first processing) Domain adaptation (persona system) For Research Novel contributions to RAG: Cross-modal validation for hybrid retrieval Context-aware confidence (sparse vs. rich context) Entity-aware chunking preserving relationship boundaries Self-critique loop before response finalization Persona-influenced information prioritization Cross-modal validation for hybrid retrieval Context-aware confidence (sparse vs. rich context) Entity-aware chunking preserving relationship boundaries Self-critique loop before response finalization Persona-influenced information prioritization Open Source & Community Repository: https://github.com/rdondeti/docuchat-agent_cli Repository: https://github.com/rdondeti/docuchat-ag License: MIT (Free with attribution - see LICENSE) License: Contributing: Contributing: Study the architecture in docs/architecture.md Review open issues tagged good-first-issue Join discussions in GitHub Discussions Submit PRs following the contribution guide Study the architecture in docs/architecture.md docs/architecture.md Review open issues tagged good-first-issue good-first-issue Join discussions in GitHub Discussions Submit PRs following the contribution guide Roadmap: Roadmap: [ ]Web UI (Streamlit-based) [ ]Multi-lingual support (expand beyond English) [ ]Fine-tuned reranker for context building [ ]Export to ONNX for embedding inference [ ]Kubernetes deployment manifests [ ]Additional domain-specific personas (HR, Sales, Customer Support) [ ]Web UI (Streamlit-based) [ ] [ ]Multi-lingual support (expand beyond English) [ ] [ ]Fine-tuned reranker for context building [ ] [ ]Export to ONNX for embedding inference [ ] [ ]Kubernetes deployment manifests [ ] [ ]Additional domain-specific personas (HR, Sales, Customer Support) [ ] Acknowledgments Built with: LangChain, LangGraph, ChromaDB, Neo4j, spaCy, Sentence-Transformers, Ollama, Rich Built with: Inspired by: Self-RAG (Asai et al., 2023), GraphRAG (Microsoft Research), Q* algorithm concepts Inspired by: If this architecture helps your project, give us a ⭐ on GitHub! Questions? Open an issue or discussion. We're building in public and learning together. Questions? Appendix: Code Snippets Library A. Custom Persona Implementation 📄 View Full Example: examples/custom_persona.py View Full Example: View Full Example: examples/custom_persona.py This example demonstrates how to create a custom "security_analyst" persona with specialized prompts and confidence thresholds. The persona is configured to prioritize vulnerabilities and security incidents, structure responses around threat analysis, and require stricter validation with multiple source documents. Usage: Usage: # Use the custom persona in chat docuchat chat --persona security_analyst # Use the custom persona in chat docuchat chat --persona security_analyst B. Custom Embedding Model 📄 View Full Example: examples/custom_embeddings.py View Full Example: View Full Example: examples/custom_embeddings.py This example shows how to extend the base EmbeddingService class to use domain-specific embedding models, such as medical or legal embeddings, instead of the default BAAI/bge-m3 model. EmbeddingService C. Custom Graph Schema 📄 View Full Example: examples/custom_medical_schema.cypher View Full Example: View Full Example: examples/custom_medical_schema.cypher This Cypher script demonstrates how to extend the knowledge graph schema for medical documents, creating custom entity types (MedicalCondition, Medication) and relationships (DIAGNOSED_WITH, TREATED_WITH) that capture domain-specific connections.
