Technical Report (EN): ChatDKU Agent RAG System

Project: ChatDKU Date: 2026-02-23 Scope: Static analysis of /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23 Security Note: All keys/tokens are redacted as [REDACTED]. The report explicitly recommends centralizing secrets in a single managed file.

Chapter 1: Project Overview & Motivation

ChatDKU is an agentic Retrieval-Augmented Generation (RAG) system built for Duke Kunshan University (DKU) stakeholders. Its core goal is to answer DKU-specific questions using curated institutional data with traceable references, while controlling latency and maintaining response quality. The codebase reveals a hybrid retrieval stack (vector + BM25), a DSPy-driven agent loop, and production-aware backend features (auth, rate limiting, observability, and user-upload indexing).

On the client side, a new Next.js frontend at chatdku/chatdku/frontend/ChatDKU-web serves as the primary UI. In development it proxies /api and /user via Next rewrites, while production uses next build with static export (output: 'export'), decoupling UI delivery from the Django API.

The primary research and engineering challenges are: retrieving correct domain-specific evidence at scale, resolving ambiguous short queries, compressing long multi-turn context without losing fidelity, and coordinating tool calls safely under real operational constraints. Within the RAG and LLM-Agent landscape, ChatDKU positions itself as a lightweight, production-oriented agentic RAG, focusing on “retrieval sufficiency judgment” rather than multi-agent autonomy.

Relative to existing paradigms:

Compared with Naive RAG, ChatDKU adds hybrid retrieval, reranking, and a judge-controlled retrieval loop.
Compared with Advanced RAG (Self-RAG/Adaptive/CRAG), it provides a judge-driven re-retrieval path but lacks explicit factuality scoring.
Compared with full Agentic RAG frameworks, tool planning is present but only one tool is currently active.

Takeaway: ChatDKU is an engineered, campus-specific agentic RAG prototype that bridges research ideas with deployable system constraints.

Chapter 2: System Architecture

The system consists of a Next.js frontend, a Django API layer, a DSPy-based agent core, a hybrid retrieval stack (ChromaDB + Redis), external LLM/embedding services, and observability/monitoring utilities. The frontend proxies requests to Django via Next.js API routes and development rewrites. The end-to-end flow is: user query → query rewrite → retrieval → judge → synthesis → memory update.

flowchart LR
  U[User/Client] --> FE[Next.js Frontend]
  FE -->|/api/* proxy| NAPI[Next.js API Routes]
  NAPI --> API[Django Chat API]
  API --> A[Agent]
  A --> QR[QueryRewrite]
  A --> P[Planner]
  P --> DR[DocumentRetriever]
  DR --> VR[VectorRetriever]
  DR --> KR[KeywordRetriever]
  DR --> RR[Reranker]
  VR --> C[(ChromaDB)]
  KR --> R[(Redis + BM25)]
  RR --> V[vLLM /v1/rerank]
  A --> J[Judge]
  J -->|if No| QR
  A --> S[Synthesizer]
  S --> LLM[(LLM Server)]
  A <--> CM[ConversationMemory]
  A <--> TM[ToolMemory]
  A --> O[Phoenix/OTel]
  S --> API --> NAPI --> FE --> U

sequenceDiagram
  participant U as User
  participant FE as Next.js Frontend
  participant NAPI as Next.js API Routes
  participant API as Django ChatView
  participant A as Agent
  participant QR as QueryRewrite
  participant DR as DocumentRetriever
  participant VR as VectorRetriever
  participant KR as KeywordRetriever
  participant RR as Reranker
  participant J as Judge
  participant S as Synthesizer
  U->>FE: query
  FE->>NAPI: POST /api/chat
  NAPI->>API: proxy request
  API->>A: current_user_message
  A->>QR: rewrite(query, memory, tool_history)
  A->>DR: semantic_query + keyword_query
  DR->>VR: Chroma query
  DR->>KR: Redis BM25 query
  DR->>RR: rerank(top_k)
  A->>J: judge(retrieved_context)
  J-->>A: Yes/No
  A->>S: synthesize(context, history)
  S-->>API: streaming response
  API-->>NAPI: response stream
  NAPI-->>FE: response stream
  FE-->>U: render response

Design patterns: Chain-of-Thought (Planner/Judge), Tool-Use, Plan-and-Execute (lightweight), Judge-gated iterative retrieval. There is no explicit multi-agent communication; all orchestration is in-process.

End-to-end example (from tests): “How often should I visit my advisor?” is rewritten by QueryRewrite, retrieved via hybrid search, re-ranked, judged for sufficiency, and synthesized with memory and references.

Chapter 3: Core Components Deep Dive

3.1 Agent (`Agent`)

Responsibility: Orchestrates the full loop of rewrite → plan → retrieve → judge → synthesize, with memory handling and streaming output. Manages internal_memory for deduplication. Interface: Agent.forward(current_user_message, question_id="", user_id="Chat_DKU", search_mode=0, files=None) returns dspy.Prediction or a generator. Internal logic:

initialize memories and tools
for each iteration:
  plan = Planner(...)
  execute tool calls
  update ToolMemory and internal_memory
  if Judge == Yes: break
  if rewrite_query: query = QueryRewrite(...)
response = Synthesizer(...)
update ConversationMemory
return response

Design choice: Judge-gated looping reduces unnecessary context while keeping recall high. Dependencies: DSPy, OpenTelemetry/Phoenix, retrievers, config. Code quality: Clean modular design, but references to CustomClient are unresolved in repo.

3.2 Planner (`Planner`)

Responsibility: Produces tool call plans given question and memory context. Interface: Planner.forward(current_user_message, tools, conversation_memory, tool_memory, max_calls=5). Logic: dspy.ChainOfThought + dspy.Refine validates tool names. Design choice: Refinement rejects invalid tool calls to reduce runtime errors. Dependencies: DSPy, ToolMemory, ConversationMemory. Code quality: Well-structured but underutilized due to single-tool setup.

3.3 QueryRewrite (`QueryRewrite`)

Responsibility: Cleans and contextualizes the query to improve retrieval. Interface: QueryRewrite.forward(current_user_message, conversation_memory, tool_memory). Logic: Injects history and tool summaries, then truncates by token budget. Design choice: Token ratio budgeting protects context window. Dependencies: DSPy, memory modules. Code quality: Clear, but effectiveness depends on prompt quality.

3.4 Judge (`Judge`)

Responsibility: Decides whether retrieved context is sufficient. Interface: Judge.forward(current_user_message, conversation_memory, tool_memory). Logic: CoT + Refine enforces Yes/No; defaults to No on repeated failure. Design choice: LLM-based sufficiency avoids hard-coded thresholds. Dependencies: DSPy, memory modules. Code quality: Robust parsing; still prompt-dependent.

3.5 Synthesizer (`Synthesizer` + `ResponseGen`)

Responsibility: Generates the final answer with DKU-specific constraints and references. Interface: Synthesizer.forward(current_user_message, conversation_memory, tool_memory, streaming, final=False). Logic: Builds prompt, optionally streams via dspy.streamify, logs spans. Design choice: Enforces reference formatting and forbids tool mentions. Dependencies: DSPy, OpenTelemetry, LLM API. Code quality: Strong guardrails, but long prompts may increase latency.

3.6 ToolMemory (`ToolMemory`)

Responsibility: Stores tool call history and summarizes overflow. Interface: ToolMemory.forward(current_user_message, conversation_memory, call, result, max_history_size). Logic: Append history, compress old entries via LLM summarization. Design choice: Summarization trades recall for context budget stability. Dependencies: DSPy, ConversationMemory. Code quality: Effective but uses fixed history size override.

3.7 ConversationMemory (`ConversationMemory`)

Responsibility: Manages chat history and summary for long conversations. Interface: ConversationMemory.forward(role, content, max_history_size=1000). Logic: Append and summarize when exceeding token limit. Design choice: Summary + recent history balances context limits. Dependencies: DSPy. Code quality: Reasonable, but summary accuracy is LLM-dependent.

3.8 Retrieval Entrypoint (`DocRetrieverOuter`)

Responsibility: Coordinates vector + keyword retrieval and reranking. Interface: DocumentRetriever(semantic_query, keyword_query=""). Logic: Runs both retrievers, reranks, merges results, returns ids. Design choice: Hybrid retrieval improves coverage for short and long queries. Dependencies: VectorRetriever, KeywordRetriever, Reranker, timeout. Code quality: Good error handling; mixed output types complicate downstream.

3.9 VectorRetriever

Responsibility: Semantic search via ChromaDB + TEI embeddings. Interface: VectorRetriever.query(query). Logic: Chroma query with metadata filters for user/file scope. Design choice: search_mode enables per-user isolation. Dependencies: ChromaDB, TEI. Code quality: Clear filtering logic; external service dependency is strong.

3.10 KeywordRetriever

Responsibility: BM25 keyword search in Redis. Interface: KeywordRetriever.query(query). Logic: NLTK tokenization + keyword combination boosting. Design choice: Weighted combinations improve multi-term precision. Dependencies: Redis, NLTK, redisvl schema. Code quality: Effective but parameterized heuristics are fixed.

3.11 Reranker

Responsibility: Calls vLLM /v1/rerank and reorders results. Interface: rerank(nodes, query, reranker_top_n). Logic: Send prompt+docs, sort by relevance scores, fallback on error. Design choice: Backup to top-n by embedding scores ensures robustness. Dependencies: vLLM rerank endpoint. Code quality: Useful, but lacks explicit timeout and retries.

3.12 Base Retriever (`BaseDocRetriever` / `NodeWithScore`)

Responsibility: Shared tracing wrapper and node structure. Interface: query_with_tell wraps query with OTEL spans. Logic: Adds OpenInference-compliant document metadata to spans. Design choice: Built-in telemetry for retrieval analysis. Dependencies: OpenTelemetry/OpenInference. Code quality: Clean and instrumented.

3.13 Config & Setup (`config.py`, `setup.py`)

Responsibility: Central configuration and initialization. Interface: Config singleton with .get/.set/.update; setup() loads embedding and tokenizer. Logic: Reads env vars, provides defaults, enforces write-protected views. Design choice: Discourages direct mutation of config dicts. Dependencies: dotenv, LlamaIndex, Transformers. Code quality: Solid, but hard-coded paths reduce portability.

3.14 DSPy Patch (`dspy_patch.py`, `dspy_common.py`)

Responsibility: Patch DSPy template formatting and backtracking. Interface: Monkey-patches DSPy internals; get_template builds final prompt. Logic: Custom system/user framing and error handling. Design choice: Practical workaround for DSPy limitations. Dependencies: DSPy internals. Code quality: Powerful but brittle across DSPy updates.

3.15 Ingestion Pipeline (`ingestion/*`)

Responsibility: Parse, chunk, and index documents into vector stores. Interface: update_data.update, load_chroma.load_chroma, load_redis.load_redis. Logic: LlamaParse for PDFs, UnstructuredReader for HTML/CSV/XLSX, chunk size 1024 overlap 20. Design choice: Incremental updates via log.json reduce recomputation. Dependencies: LlamaParse, Unstructured, Chroma, Redis, LlamaIndex. Code quality: Functional but contains hard-coded API keys (redacted here).

3.16 User Upload Indexing (`backend/user_data_interface.py`)

Responsibility: Index user-uploaded files into Chroma and Redis. Interface: update(data_dir, user_id). Logic: Same chunking logic with user-scoped metadata filters. Design choice: Enables personal corpus retrieval via search_mode. Dependencies: Chroma, Redis, LlamaParse. Code quality: Overlaps with ingestion code; could be refactored.

3.17 Django Backend (Primary)

Responsibility: Chat API, file upload, rate limiting, Celery tasks. Interface: /api/chat, /api/feedback, /user/upload, /user/health. Logic: Redis locks for uploads, streaming responses from Agent, IP-based rate limiting. Design choice: Treat Django as primary production layer. Dependencies: Django, DRF, Celery, Redis, Prometheus. Code quality: Production-ready but testing coverage is thin.

3.18 Flask Backend (Legacy/Auxiliary)

Responsibility: Legacy chat endpoints and SocketIO STT flow. Interface: /chat, /feedback, /upload, audio_data. Logic: Agent instance per request, Whisper-based transcription. Design choice: Retained for backward compatibility and STT handling. Dependencies: Flask, SocketIO, Whisper. Code quality: Functional but redundant with Django.

3.19 Tools Ecosystem

Responsibility: Search tools, email tools, Python execution, GraphRAG, SQL course tool. Interface: DuckDuckGo, GoogleSearch, EmailTools, GraphragTool, QueryCurriculumDB. Logic: Mostly wrappers around external APIs; not yet integrated into Planner. Design choice: Tool extensibility planned, but integration incomplete. Dependencies: External APIs and services. Code quality: Mixed completeness; some missing imports.

3.20 Scraper

Responsibility: Crawl DKU web pages and build raw corpus. Interface: scraper.py CLI, filter_llm.filter_page. Logic: Async crawl with retry and optional LLM-based filtering. Design choice: LLM filtering reduces noisy pages. Dependencies: aiohttp, BeautifulSoup, OpenAI SDK. Code quality: Useful but SSL verification is disabled.

3.21 Benchmarks & Monitoring

Responsibility: Load testing and monitoring integration. Interface: k6 scripts, locust tasks, Prometheus/Phoenix docs. Logic: Weekly/daily test runs plus alert emails. Design choice: Embeds observability in system lifecycle. Dependencies: k6, locust, Prometheus, Phoenix. Code quality: Tooling is adequate, but depends on external datasets.

3.22 Frontend Layer (Next.js / ChatDKU-web)

Responsibility: Provides the ChatDKU web UI including sessions, chat rendering, uploads, and feedback entry points, while proxying backend requests through Next.js API routes. Interface: app/api/chat/route.ts proxies chat, app/api/feedback/route.ts proxies feedback, and app/api/get_session/route.ts / app/user/route.ts provide development mock session/user data. Internal logic: ChatPage.tsx renders the streaming chat UI, parses Markdown, and strips <think> blocks; lib/convosNew.ts manages session storage and history; components/* compose the UI via shadcn/ui primitives. Design decisions: Static export in production (output: 'export'), with rewrites or API routes handling SSL/CORS proxying in development. Dependencies: Next.js, React, TypeScript, Tailwind CSS, shadcn/ui, Radix UI, Jest. Code quality: Clean component structure with growing test coverage, but backend endpoints are still hard-coded and should be configuration-driven.

Chapter 4: Retrieval Pipeline

Ingestion & preprocessing: update_data.py parses PDF/HTML/CSV/XLSX, then chunks with SentenceSplitter(chunk_size=1024, chunk_overlap=20). PDFs are handled by LlamaParse; non-PDF uses UnstructuredReader. Custom filetype detection and partition logic address HTML misclassification issues.

Embedding & indexing: TEI serves embeddings (default BAAI/bge-m3). ChromaDB hosts vector embeddings, Redis hosts BM25 (and vector index in some flows). custom_schema.yaml defines fields and vector dimensions.

Query processing: QueryRewrite performs spelling/context cleanup. There is no explicit query decomposition, but the judge loop enables iterative re-retrieval.

Retrieval strategy: DocRetrieverOuter concurrently queries vector and keyword stores, optionally re-ranks with vLLM.

Context assembly: Results are converted to {text, metadata} and stored with internal_memory ids to avoid duplication.

Non-standard/innovative aspects: keyword combination boosting, judge-gated re-retrieval, reranker fallback on failure.

Chapter 5: Agent Framework & Reasoning

The agent implements a lightweight Plan-and-Execute loop with judge gating. Planner produces tool calls, Judge decides if more retrieval is needed, and Synthesizer renders the final answer. Memory is managed through ConversationMemory and ToolMemory with LLM-based compression.

Tools include DocumentRetriever as the primary one; other tools exist but are not wired into the planner. No multi-agent architecture is implemented.

Chapter 6: Prompt Engineering

Key prompts/templates include: SynthesizerSignature (system role and DKU constraints), QueryRewriteSignature, PlannerSignature, JudgeSignature, memory compression signatures, feedback summarization prompt, title generation prompt, crawler filter prompt, and SQL generation prompts.

Dynamic variables include current_user_message, conversation_history, tool_history, summaries, and current_date. Prompt chaining is explicit: rewrite → plan → tool calls → judge → synthesize.

Strengths include strict reference formatting and DKU contextualization; weaknesses include long prompts and lack of automatic prompt evaluation.

The frontend renderer strips <think> blocks before Markdown rendering to avoid exposing internal reasoning to users, acting as a UI-side safety guard.

Chapter 7: Technology Stack & Infrastructure

Stack (with versions): LlamaIndex 0.13.1, DSPy ~3.0.3, ChromaDB ~1.0.15, Redis + redisvl, OpenTelemetry/OpenInference/Phoenix, Django 5.2.3, Flask 3.0.3, Celery, Whisper, locust, k6. Frontend: Next.js ^16.1.6, React ^19, TypeScript ^5, Tailwind CSS ^4, shadcn/ui, Radix UI, Jest/Testing Library, next-themes.

Configuration: Central Config in chatdku/chatdku/config.py, plus Django/Flask config files. Environment variables in .env are assumed.

Deployment: TEI + vLLM via Docker; Django on gunicorn/Apache; Phoenix for tracing.

Performance: Streaming responses, caching via DSPy cache, ingestion batching, and asynchronous scraping.

Chapter 8: Error Handling & Robustness

There are explicit timeouts around retrieval (timeout wrapper), reranker fallbacks, and backend-level error handling. Judge defaults to “No” if output is malformed, reducing false early exits.

Potential failures include missing definitions (CustomClient, DocumentSummarizer), hard-coded API keys (redacted), and external dependency brittleness. Token budgets are controlled via ratio-based truncation and memory summarization.

Conclusion: The system is resilient in retrieval layers but lacks robust hallucination detection or answer faithfulness checks.

Chapter 9: Evaluation & Experiments

Existing evaluation consists of retriever tests, load tests, and benchmark scripts: test_retriever.py, test_load_retriever.py, k6 backend benchmark, and locust-based Django load tests. The frontend uses Jest (jest.config.js, jest.setup.js) with integration tests such as integration/chat-flow.test.tsx, component/unit tests, and a production-readiness-reporter.js for coverage gates. These focus on latency and throughput rather than answer correctness.

Missing evaluation: no systematic faithfulness/correctness metrics or retrieval quality metrics. Recommended metrics include Recall@k, nDCG, faithfulness, answer correctness, and latency distributions. Suggested frameworks include RAGAS or Phoenix evaluations.

Chapter 10: Novelty & Technical Contributions

Contribution 1: Judge-gated iterative retrieval with deduplication The Judge loop controls re-retrieval and uses internal_memory ids to avoid repeated evidence. This delivers adaptive retrieval without multi-agent complexity.

Contribution 2: Practical hybrid retrieval + rerank with fallback Parallel vector and keyword retrieval are re-ranked with vLLM; failures gracefully fall back to top-N vector results. This improves robustness relative to single-strategy RAG.

Contribution 3: Memory compression and prompt budget governance Conversation and tool memories are summarized with LLMs and constrained by token ratios, keeping long sessions stable. DSPy patches enforce predictable prompt formatting.

Chapter 11: Limitations & Future Work

Limitations include underutilized planning (single tool), incomplete evaluation, missing tool integrations, and hard-coded secrets/paths. Future work should integrate multiple tools, add factuality scoring, unify ingestion pipelines, and centralize secret management into a single secure file or service.

Chapter 12: Comparison with Related Work

Dimension	Naive RAG	Advanced RAG	Agent Methods	This System	Differentiator
Retrieval	Single vector	Adaptive/iterative	Tool-driven	Hybrid + rerank	Strong hybrid fusion
Iteration	No	Partial	Yes	Yes (Judge loop)	Judge-controlled loop
Tools	None	Few	Many	One (expandable)	Tool-ready but limited
Memory	Minimal	Compressed	Long-term	Conversation/Tool memory	Production-ready memory
Ops	Low	Medium	Medium	High	Phoenix + Prometheus + locust

Appendix

A. File-Level Summaries (2–3 sentences each)

All paths are absolute to guarantee full coverage.

File	Summary
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/README.md`	Declares the repository as a monorepo and points to subprojects. It serves as a navigation entry rather than implementation detail.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/TODO.md`	Tracks pending integration of the SQL syllabi tool with the planner. It documents current gaps in the roadmap.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/manage.py`	Loads `.env` at the repo root. Acts as a lightweight environment initializer.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/GUIDE.md`	Onboarding guide with dependency explanations and team roles. It highlights DSPy, LlamaIndex, Chroma, and Redis usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/Documentations/Database-Documentation.md`	Documents Redis and ChromaDB setup with Docker. Includes usage notes and password guidance.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/Documentations/Embedding-Server-Documentation.md`	Describes TEI embedding server setup and nginx routing. Lists valid model endpoints and metrics paths.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/Documentations/Prometheus-Monitoring-Documentation.md`	Explains Prometheus architecture, configuration paths, and sample queries. Details target services and alert setup.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/README.md`	Summarizes utility scripts in the utils directory. It is purely descriptive.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_backend.sh`	Curl-based test script for `/chat` endpoint. Accepts port and optional query content.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_feedback.sh`	Curl-based test script for feedback endpoint. Intended for quick API checks.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/data_count_size.sh`	Bash script to count files by extension and total size. Useful for data directory audits.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/generate_redis_schema.py`	Generates a sample Redis index schema via redisvl. Used for test indexing scenarios.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/save_tokenizer.py`	Downloads and saves Hugging Face tokenizer/config to disk. Enables offline or consistent tokenizer usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/visualization/dataVisualizer.py`	Implements 2D/3D visualization utilities using Matplotlib/Seaborn/Plotly. Includes example usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/visualization/README.md`	Explains how to use the DataVisualizer class. Provides example plotting calls.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_redis/bm25_search.py`	Simple BM25 search test in Redis. Demonstrates grouping filters.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_redis/bm25_search_improved.py`	Enhanced BM25 test with keyword combinations and highlights. Helps tune search weights.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_redis/chinese.py`	Tests Redis search with Chinese language mode. Shows summarize/highlight behavior.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_redis/load.py`	Minimal example that loads documents into RedisVectorStore. Uses TEI embeddings and a custom schema.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/utils/test_redis/vector_retriever.py`	Demonstrates LlamaIndex retrieval from Redis with metadata filters. Provides a test retrieval function.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/benchmarks/README.md`	Documents k6 and vLLM benchmark workflows. Notes data requirements and output locations.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/benchmarks/backend.js`	k6 load test targeting `/chat`. Generates an HTML report using k6-reporter.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/benchmarks/visualize_vllm.py`	Parses vLLM benchmark JSON and plots TTFT/TPOT/throughput. Outputs SVG plots.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/benchmarks/nohup.out`	Runtime log artifact from prior benchmark executions. Not an executable source file.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/README.md`	Usage guide for the DKU web scraper. Explains recursion, output, and progress reporting.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/pyproject.toml`	Dependency manifest for the scraper package. Separates scraper requirements from core.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/scraper/__init__.py`	Package marker for scraper module. Contains no logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/scraper/utils.py`	Defines download status enum and data structures. Provides summary printing.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/scraper/report.py`	Reads CSV download info and prints status summary. Used for crawl reporting.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/scraper/filter_llm.py`	Uses an LLM to keep/drop pages based on a strict prompt. Supports rate limiting and parsing heuristics.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/scraper/scraper/scraper.py`	Main async crawler with aiohttp and SAML login handling. Includes delay, retry, and URL normalization.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/README.md`	Project overview and setup steps for LLM/embedding services. Notes Phoenix observability usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/pyproject.toml`	Main dependency list and versions for ChatDKU. Declares Django, Flask, LlamaIndex, DSPy, etc.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/config.py`	Defines a singleton configuration with defaults and safe mutation controls. Central for all endpoints and paths.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/setup.py`	Initializes embedding model and tokenizer in LlamaIndex Settings. Also registers Phoenix tracing.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/README.md`	Explains the agent pipeline and DSPy submodules. Notes Planner is currently disabled for full agentic behavior.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/agent.py`	Implements the core agent loop and streaming response handling. Manages tool memory, conversation memory, and internal ids.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/compile.py`	DSPy compilation script for prompt optimization. References `CustomClient`, which is missing in the repo.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/utils.py`	Token truncation and helper utilities for prompt budgeting. Includes function-to-model mapping and conversation loading.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_common.py`	Utilities for constructing DSPy prompt templates. Provides custom rationale field.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_patch.py`	Monkey-patches DSPy templates and backtracking. Enables custom prompt formatting and assertion handling.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/prompt_settings.py`	Defines shared prompt fields and role prompt text. Centralizes description strings.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/conversation_memory.py`	Implements conversation memory with LLM-based summarization. Maintains both history and summary.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/tool_memory.py`	Implements tool memory and compression of tool history. Stores tool calls and results.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/query_rewrite.py`	Query rewrite DSPy module. Injects conversation and tool context into rewritten query.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/plan.py`	Planner module using Chain-of-Thought and refinement. Produces tool call plans.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/judge.py`	Judge module that outputs Yes/No for retrieval sufficiency. Filters reasoning tags from outputs.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/dspy_classes/synthesizer.py`	Synthesizer module with DKU-specific response rules. Supports streaming via ResponseGen.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/llama_index.py`	Wrapper around vector and keyword retrievers. Returns merged results and internal ids for deduplication.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/retriever/base_retriever.py`	Base retriever class with OpenTelemetry instrumentation. Defines `NodeWithScore` structure.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/retriever/vector_retriever.py`	Queries ChromaDB with TEI embeddings and metadata filters. Implements search_mode logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/retriever/keyword_retriever.py`	Performs Redis BM25 search with weighted keyword combinations. Builds RediSearch query strings.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/retriever/reranker.py`	Calls vLLM rerank endpoint and applies fallback on error. Emits retrieval spans.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/utils.py`	Provides timeout context manager and URL mapping helpers. Converts NodeWithScore to dictionaries.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/calculator.py`	Math utility class with JSON responses. Covers add/subtract/multiply/divide and optional functions.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/campus_service.py`	Excel lookup tool for campus services. Returns a dict of matched fields.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/pythonTool.py`	Executes or saves Python code within constraints. Includes file operations and pip install helpers.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/search/duckduckgo.py`	Wrapper for DuckDuckGo search and news. Returns JSON text results.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/search/python_googlesearch.py`	Google search wrapper via `googlesearch-python`. Supports language normalization.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/search/api_google_search.py`	Google Custom Search API wrapper. Returns LlamaIndex Document objects.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/search/brave_search.py`	Brave Search API wrapper sketch. Missing imports suggest incomplete implementation.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/email/email_tool.py`	SMTP email utility with attachments and inline images. Used for operational notifications.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/email/resend_tool.py`	Resend API email tool. Reads API key from environment variables.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/graphrag.py`	GraphRAG tool wrapper for global community search. References `DocumentSummarizer`, which is missing.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/redis_listener/redis_listener.py`	Listens for Redis key deletion/expiry and sends alert summaries. Maintains buffered notifications.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/redis_listener/redis_hanging_detector.py`	Detects Redis hanging via ping latency. Sends alerts when thresholds are exceeded.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/sql_agent.py`	DSPy-based SQL generation with table/column selection and sanitization. Prevents destructive queries.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/query_curriculum_db.py`	Executes SQL queries against the curriculum database. Returns natural language summaries.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/local_ingest.py`	Local ingestion tool for PDF/DOCX to structured data via SGLang. Includes schema validation and logging.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/update_db.py`	Uses LlamaExtract to parse syllabi PDFs and insert into PostgreSQL. Contains batch upsert logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/get_schema.py`	Reads and prints DB schema for `classes` table. Useful for debugging SQL generation.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/create_table.sql`	SQL schema definition for the syllabi database. Establishes course metadata fields.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/classes_schema.json`	JSON schema for syllabi extraction. Must stay synchronized with database.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/run_local_ingest.sh`	Shell runner for local ingestion. Encapsulates command-line usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/core/tools/syllabi_tool/syllabi-dev.zip`	Binary dataset/asset for syllabi tool development. Not executable code.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/README.md`	Ingestion pipeline documentation. Describes update_data, load_chroma, and load_redis.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/update_data.py`	Incrementally updates nodes.json based on file changes. Handles PDF and non-PDF parsing.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/load_chroma.py`	Loads nodes into ChromaDB with batch buffering. Supports custom collection names.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/load_redis.py`	Loads nodes into RedisVectorStore with a custom schema. Supports cache and reset.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/custom_filetype_detect.py`	Overrides Unstructured filetype detection for HTML edge cases. Improves MIME classification.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/custom_partation.py`	Adjusts Unstructured partition strategy to hi-res and HTML trimming. Enhances parse quality.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/documents_reprocess.py`	Cleans and reprocesses parsed documents. References a config object inconsistently.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/ingestion/menu2doc.ipynb`	Notebook for experimental data conversion. Not in the production pipeline.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/custom_schema.yaml`	Redis schema file with vector index parameters. Used by retrieval and ingestion.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/hash.pkl`	Binary hash cache artifact. Likely used for state tracking.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/tests/test_retriever.py`	Unit tests for retriever correctness and timeouts. Requires external services to pass.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/tests/test_load_retriever.py`	Concurrent load test for retriever performance. Prints latency statistics.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/readme.md`	Flask backend documentation and deployment notes. Covers Whisper and HTTPS setup.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/main.py`	Dev entry point to run Flask with Agent. Not intended for production.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/agent_app_parellel.py`	Parallel Flask app with Whisper and Agent streaming. Maintains legacy endpoints.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/stt_app.py`	SocketIO-based STT server with HTTPS. Routes audio to Whisper service.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/whisper_model.py`	Whisper inference API with audio preprocessing. Provides `/process_audio` and `/transcribe`.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/config.py`	Flask config for DB, secret key, and limits. Parallel to Django settings.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/user_data_interface.py`	Indexes user uploads into Chroma and Redis. Mirrors ingestion logic for user data.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/user_data_interface_readme.md`	Instructions for user data interface usage and search modes. Clarifies parameter semantics.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/__init__.py`	Initializes Flask app, SocketIO, DB, and Agent. Sets logging and admin configuration.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/routes.py`	Defines Flask routes for chat, feedback, upload, and audio. Streams responses from Agent.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/models.py`	SQLAlchemy models for feedback, requests, users, and uploads. Supports basic analytics.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/admin.py`	Flask-Admin views with Plotly charts for requests. Used in the admin dashboard.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/utils.py`	Helper functions for allowed file extensions. Used in upload flow.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/app/templates/admin.html`	Jinja template for admin dashboard with Plotly graphs. Extends Flask-Admin base template.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/README`	Alembic migration readme. Documents database migration usage.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/alembic.ini`	Alembic configuration file for migrations. Used by Flask backend.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/env.py`	Migration environment configuration for Alembic. Loads app context.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/script.py.mako`	Template for Alembic migration scripts. Used when generating new migrations.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/versions/ef34fed121b6_initial_migration.py`	Initial migration creating base tables. Defines core schema.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/versions/72e7656c297a_request_table.py`	Migration for request table adjustments. Supports request counting.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/versions/225612aaf33f_date_removed.py`	Migration that adjusts date-related fields. Ensures schema compatibility.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/versions/ae3073ac5fd4_time.py`	Migration that modifies time fields. Used for request time tracking.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/migrations/versions/cb48e322485c_.py`	Migration with truncated filename, indicating schema change. Needs Alembic history for detail.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/readme.md`	Comprehensive Django backend documentation. Covers env vars, deployment, and Celery.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/manage.py`	Django management entry point. Runs server and admin commands.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/__init__.py`	Django project package marker. No logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/asgi.py`	ASGI configuration for Django. Used in async deployments.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/wsgi.py`	WSGI configuration for Django. Used in production gunicorn deployments.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/settings.py`	Django settings including DB, cache, rate limits, and localization. Core production config.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/urls.py`	Root URL configuration for Django and API routes. Includes schema and Prometheus endpoints.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/views.py`	Empty placeholder views file. No routes implemented here.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/celery.py`	Celery configuration with Redis broker and scheduled tasks. Defines periodic load tests.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/locustfile.py`	Locust test script simulating chat traffic. Validates response length thresholds.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/locust_daily.conf`	Locust daily configuration file. Defines users, spawn rate, and runtime.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/locust_weekly.conf`	Locust weekly configuration for scheduled tests. Includes CSV output.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/start_locust_bg.sh`	Script to launch locust master/worker in background. Supports multi-worker setup.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/schema.yml`	OpenAPI schema configuration for drf-spectacular. Used for docs generation.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/__init__.py`	Core app package marker. No logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/apps.py`	Core app startup initializes DSPy and Phoenix. Enables caching for performance.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/admin.py`	Django admin registration for users and uploads. Enforces superuser-only edits.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/models.py`	User and UploadedFile models. Includes file deletion hook.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/serializers.py`	Upload serializer validating PDF and size. Limits files to 10MB.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/views.py`	Upload and health endpoints. Uploads enqueue Celery tasks.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/urls.py`	Routes for upload and health endpoints. Attached under `/user/`.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/utils.py`	Slugify helper and admin email lookup. Used in uploads and notifications.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/middleware.py`	NetID authentication middleware. Injects user into request.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/rate_limit_middleware.py`	Sliding-window rate limiter with IP-based identification. Adds rate limit headers.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/set_enqueue.py`	Queues user tasks in Redis and triggers Celery. Prevents parallel uploads.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/set_lock.py`	Redis lock helper for mutual exclusion. Used by ingestion tasks.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/tasks.py`	Celery tasks for user indexing, cleanup, and load tests. Includes retry logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/signals.py`	Placeholder for Django signals. Currently unused.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/tests.py`	Placeholder test file for core app. No implemented tests.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/migrations/0001_initial.py`	Initial migration for user and upload models. Defines primary schema.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/migrations/0002_activelm.py`	Migration adding ActiveLM or related model. Supports model switching logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/migrations/0003_delete_activelm_usermodel_email.py`	Migration removing ActiveLM and altering UserModel fields. Reflects model evolution.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/core/migrations/__init__.py`	Migration package marker. No logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/__init__.py`	Chat app package marker. No logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/apps.py`	Chat app config. No additional logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/models.py`	Defines chat sessions, messages, and feedback models. Includes Prometheus mixins.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/serializer.py`	Serializers for chat messages, sessions, and feedback. Computes search_mode from sources.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/views.py`	Django chat API with streaming responses. Persists messages in database and handles uploads.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/utils.py`	Feedback summarization, title generation, and LLM ping helpers. Contains async title generation.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/tasks.py`	Scheduled tasks for load testing and email reporting. Uses locust configs and thresholds.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/urls.py`	Routes for chat and feedback endpoints. Registers session routes via DRF router.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/tests.py`	Placeholder test file for chat app. No test cases defined.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/admin.py`	Admin registration for chat models. Enforces read-only behavior.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/mail.py`	Email utility for weekly reports and alerts. Supports HTML content and inline logo.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/migrations/0001_initial.py`	Initial migration for chat app models. Sets up session and message tables.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/migrations/0002_alter_feedback_question_id.py`	Adjusts Feedback question ID field. Ensures schema correctness.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/migrations/0003_usersession_chatmessages.py`	Migration altering session-message relationships. Supports message history.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/migrations/0004_alter_usersession_user.py`	Migration updating UserSession foreign key. Aligns with auth changes.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/migrations/__init__.py`	Migration package marker. No logic.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/templates/email/weekly_report.html`	HTML template for weekly load-test email. Rendered by Django templates.
`/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chat/templates/images/edge-intelligence.png`	Embedded logo asset for emails. Binary resource file.

| /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/.gitignore | Frontend file used by the Next.js application. Provides UI, config, or assets for the web client. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/README.md | Frontend developer documentation describing stack, workflow, and deployment steps. Emphasizes static export and testing before production. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/about/page.tsx | About page route for the frontend. Provides informational content about the project/team. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/api/chat/route.test.ts | Frontend test case for route behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/api/chat/route.ts | Next.js API route that proxies chat requests to the backend service. Used to avoid browser SSL limitations by server-side forwarding. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/api/feedback/route.test.ts | Frontend test case for route behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/api/feedback/route.ts | Next.js API route that validates feedback payloads and forwards them to the backend. Returns structured error responses on failure. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/api/get_session/route.ts | Development-only API route that returns mock session data. Supports local UI testing without backend dependencies. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/dev/page.tsx | Development page route for local testing. Used for internal tooling or debug UI. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/favicon-small.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/favicon.ico | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/globals.css | Global CSS for the frontend. Establishes baseline styles and Tailwind layers. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/icon.ico | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/landing/page.tsx | Landing page route shown before entering chat. Typically used for terms or onboarding. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/layout.tsx | Root layout for Next.js app, including metadata, fonts, and theme provider. Wraps all pages with global layout. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/page.tsx | Main entry page for the chat experience. Renders the ChatPage component. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/team-credits/page.tsx | Team credits page route. Displays contributors and acknowledgements. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/app/user/route.ts | Development-only API route that returns mock user profiles. Also accepts POST to update mock profile data. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components.json | shadcn/ui configuration file describing component paths and settings. Guides UI component generation and conventions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ChatPage.test.tsx | React component implementing the ChatPage.test UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ChatPage.tsx | Primary chat UI component handling session setup, message rendering, and streaming UI. Parses Markdown and strips <think> blocks before display. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/UploadSheet.tsx | React component implementing the UploadSheet UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/UserUploadSheet.tsx | React component implementing the UserUploadSheet UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/WelcomeBanner.tsx | React component implementing the WelcomeBanner UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/about.tsx | React component implementing the about UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/chatdku_remark.tsx | React component implementing the chatdku remark UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/doc-manager.tsx | React component implementing the doc manager UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/dynamic-logo.tsx | React component implementing the dynamic logo UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/feedback.test.tsx | React component implementing the feedback.test UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/hooks/use-auto-resize-textarea.tsx | React component implementing the use auto resize textarea UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/hooks/use-media-query.ts | React component implementing the use media query UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/hooks/useUpload.ts | React component implementing the useUpload UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/hooks/user.ts | React component implementing the user UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/navbar.tsx | React component implementing the navbar UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/prompt_recs.tsx | React component implementing the prompt recs UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/side.tsx | React component implementing the side UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/starter-buttons.tsx | React component implementing the starter buttons UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/starter.tsx | React component implementing the starter UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/theme-provider.tsx | React component implementing the theme provider UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/aboutModel.tsx | shadcn/ui wrapper component for aboutModel. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/ai-input.test.tsx | shadcn/ui wrapper component for ai input.test. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/ai-input.tsx | shadcn/ui wrapper component for ai input. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/alert.tsx | shadcn/ui wrapper component for alert. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/badge.tsx | shadcn/ui wrapper component for badge. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/button.test.tsx | shadcn/ui wrapper component for button.test. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/button.tsx | shadcn/ui wrapper component for button. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/card.tsx | shadcn/ui wrapper component for card. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/checkbox.tsx | shadcn/ui wrapper component for checkbox. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/combobox.tsx | shadcn/ui wrapper component for combobox. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/command.tsx | shadcn/ui wrapper component for command. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/dialog.tsx | shadcn/ui wrapper component for dialog. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/drawer.tsx | shadcn/ui wrapper component for drawer. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/dropdown-menu.tsx | shadcn/ui wrapper component for dropdown menu. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/input.test.tsx | shadcn/ui wrapper component for input.test. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/input.tsx | shadcn/ui wrapper component for input. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/label.tsx | shadcn/ui wrapper component for label. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/mode-toggle.tsx | shadcn/ui wrapper component for mode toggle. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/navigation-menu.tsx | shadcn/ui wrapper component for navigation menu. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/popover.tsx | shadcn/ui wrapper component for popover. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/progress.tsx | shadcn/ui wrapper component for progress. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/scroll-area.tsx | shadcn/ui wrapper component for scroll area. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/sheet.tsx | shadcn/ui wrapper component for sheet. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/terms.tsx | shadcn/ui wrapper component for terms. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/textarea.tsx | shadcn/ui wrapper component for textarea. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/ui/tooltip.tsx | shadcn/ui wrapper component for tooltip. Provides styled UI primitives used across the frontend. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/components/utils.tsx | React component implementing the utils UI element for the ChatDKU frontend. Encapsulates layout and interaction logic for that section. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/deploy.sh | Deployment script that runs tests, builds the static site, backs up existing deployment, and rsyncs output. Includes interactive confirmation steps. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/eslint.config.mjs | ESLint configuration for the frontend codebase. Enforces lint rules across TS/TSX files. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/hooks/use-auto-resize-textarea.tsx | Custom React hook for use auto resize textarea behavior. Encapsulates reusable client-side logic for the UI. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/integration/chat-flow.test.tsx | Frontend test case for chat flow behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/jest.config.js | Jest configuration for frontend tests with jsdom environment. Defines transforms, coverage, and path mapping. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/jest.setup.js | Jest setup file that loads DOM matchers and mocks Next.js Request/Response globals. Supports API route and UI tests. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/constants.ts | Centralized constants for API endpoints and configuration. Used across frontend utilities and components. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/convos.ts | Legacy conversation utilities for the frontend. Provides helper functions for chat/session handling. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/convosNew.test.ts | Frontend test case for convosNew behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/convosNew.ts | Session and conversation utilities for the frontend, including cookies and history retrieval. Encapsulates API calls and storage helpers. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/markdown-streaming.test.ts | Frontend test case for markdown streaming behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/utils.test.ts | Frontend test case for utils behavior. Validates UI logic, API routes, or utility functions. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/utils.ts | Shared utility functions for frontend components and styling. Supports common helpers and class merging. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/lib/utils.tsx | Shared utility functions for frontend components and styling. Supports common helpers and class merging. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/next.config.ts | Next.js configuration enabling static export in production and API rewrites in development. Also disables image optimization for static hosting. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/package-lock.json | Locks npm dependency versions for reproducible installs. Auto-generated by npm and not edited manually. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/package.json | Defines frontend package metadata, scripts, and dependencies. Controls dev/build/test workflows for the Next.js app. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/postcss.config.mjs | PostCSS configuration used by Tailwind and build tooling. Defines CSS processing pipeline. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/production-readiness-reporter.js | Custom Jest reporter that evaluates production readiness based on test success and coverage thresholds. Exits with pass/fail status to gate deployment. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/ChatDKU-1.0_Introduction.pdf | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Anar.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Bing.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Chenshuhao.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Haroon.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Jingxuan.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Mingxi.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Munish.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Ningyuan.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/RuihanYin.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Sean.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Temuulen.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Youran.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Yuxiang.jpeg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Zeyu.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/avatars/Zhiwei.jpg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/favicon-small.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/file.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/globe.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/BL_Edge Intelligence Lab_04.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/BL_Edge Intelligence Lab_06.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Dark-Logo-ChatDKU.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Dark-Logo-ChatDKU.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Dark-Logo.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Dark-Logo.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Light-Logo-ChatDKU.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Light-Logo-ChatDKU.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Light-Logo.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/Light-Logo.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/edge-intelligence.png | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/logos/new_logo.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/mdtest.md | Markdown content file used by the frontend. Serves as static content or test fixture. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/public/window.svg | Static asset for frontend UI/branding. This is a binary image/document file with no executable logic. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/tailwind.config.ts | Tailwind CSS configuration for design tokens and scanning paths. Controls utility generation for the UI. | | /Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/frontend/ChatDKU-web/tsconfig.json | TypeScript compiler configuration for the frontend. Sets module resolution and strictness defaults. |

B. Module Dependency Graph (Mermaid)

flowchart TD
  Core[core.agent + dspy_classes] --> Tools[core.tools.*]
  Core --> Config[chatdku.config]
  Core --> Setup[chatdku.setup]
  Tools --> Retriever[retriever.*]
  Retriever --> Stores[Chroma/Redis]
  Django[django chat/core apps] --> Core
  Django --> Celery[celery tasks]
  Celery --> Ingestion[backend.user_data_interface]
  Ingestion --> Chroma[ChromaDB]
  Ingestion --> Redis[Redis]
  Scraper[scraper.*] --> Data[raw web data]
  Benchmarks --> Backend

C. Glossary

RAG: Retrieval-Augmented Generation.
DSPy: Framework for programming and optimizing LLM pipelines.
TEI: Text Embeddings Inference server for embeddings.
BM25: Classical keyword ranking algorithm.
Phoenix/OpenTelemetry: Observability tooling for traces.
vLLM: OpenAI-compatible inference server with rerank endpoint.
LlamaIndex: Indexing and retrieval framework.
LlamaParse: Document parsing service for PDFs.
SGLang: LLM inference framework used for structured extraction.

D. Key Configuration Table

Core Config (/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/config.py)

Parameter	Default	Role
`llm`	`Qwen/Qwen3.5-4B`	Primary LLM model
`llm_url`	`http://localhost:18085/v1`	LLM API base URL
`embedding`	`BAAI/bge-m3`	Embedding model
`tei_url`	`http://localhost:18080`	TEI endpoint
`context_window`	`32000`	Context window size
`reranker_base_url`	`http://localhost:6767`	Rerank endpoint
`data_dir`	`/datapool/chat_dku_advising`	Data directory
`nodes_path`	`/datapool/chat_dku_advising/nodes.json`	Nodes file
`chroma_collection`	`dku_html_pdf`	Chroma collection
`index_name`	`chat_dku_advising`	Redis index

Django Settings Highlights (/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/django/chatdku_django/chatdku_django/settings.py)

Parameter	Default	Role
`DATABASES`	env-based	PostgreSQL config
`CACHES`	Redis	Cache and rate limits
`RATE_LIMIT_DEFAULT`	60/min	Default rate limit
`MEDIA_ROOT`	env-based	Upload storage path
`CELERY_RESULT_BACKEND`	django-db	Celery result backend

Flask Config Highlights (/Users/glitterc/Desktop/CodeX_codes/ChatDKU_2_23/chatdku/chatdku/backend/config.py)

Parameter	Default	Role
`SQLALCHEMY_DATABASE_URI`	sqlite/env	Flask DB
`SECRET_KEY`	local default	Session key (should be centralized)

Security Recommendation: All secrets should be consolidated into a single managed secrets file or service, replacing hard-coded values in code and docs.

FilesExpand file tree

Technical_Report_EN.md

Latest commit

History