Skip to content

Add all 5 GenAI learning projects from spec#2

Merged
nerdjerry merged 9 commits intocopilot/charming-turtlefrom
copilot/build-project-as-per-spec
Feb 26, 2026
Merged

Add all 5 GenAI learning projects from spec#2
nerdjerry merged 9 commits intocopilot/charming-turtlefrom
copilot/build-project-as-per-spec

Conversation

Copy link
Contributor

Copilot AI commented Feb 25, 2026

  • Root README.md
  • .gitignore
  • Project 1: RAG From Scratch
  • Project 2: Legal AI Assistant
  • Project 3: AI Research Agent
  • Project 4: Multimodal RAG — fix: removed base64 (stdlib, not pip-installable) from requirements.txt; pdfplumber==0.11.1 was already present
  • Project 5: Agentic RAG with Real-Time Data — wikipedia==1.4.0 and requests==2.31.0 were already present in requirements.txt
Original prompt

build project as per spec in @GenAI_Projects_Spec.md

Created from Copilot CLI via the copilot delegate command.


💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

Copilot AI and others added 7 commits February 25, 2026 18:57
Complete RAG from scratch implementation with:
- 6-step pipeline: load → chunk → embed → index → retrieve → generate
- HuggingFace all-MiniLM-L6-v2 embeddings (free, local, no API key)
- FAISS vector store with disk persistence
- OpenAI + Ollama LLM support
- argparse CLI with single-question and interactive modes
- Heavily commented, beginner-friendly code throughout

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Complete RAG-powered legal contract analysis tool with:
- document_parser.py: PDF/DOCX parsing with section detection
- indexer.py: FAISS vector indexing (same pattern as Project 1)
- summarizer.py: executive summary with structured JSON output
- clause_extractor.py: named clause extraction + plain-English translation
- risk_analyzer.py: HIGH/MEDIUM/LOW risk scoring with emoji indicators
- conflict_detector.py: internal contradiction detection (with disclaimers)
- qa_chain.py: grounded RAG Q&A with mandatory section citations
- main.py: full CLI pipeline using Rich for formatted output
- prompts/: summary, clause, and risk prompt templates
- README.md: setup guide, architecture diagram, limitations

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Complete ReAct research agent pipeline including:
- paper_parser.py: LLM-based PDF metadata extraction (PaperMetadata Pydantic model)
- paper_indexer.py: FAISS vector index with per-paper metadata filtering
- tools/search_tool.py: semantic search LangChain Tool
- tools/summary_tool.py: paper summary LangChain Tool
- tools/compare_tool.py: LLM-powered two-paper comparison Tool
- agent.py: ZERO_SHOT_REACT_DESCRIPTION AgentExecutor with verbose ReAct loop
- gap_analyzer.py: cross-paper synthesis and research gap identification
- report_generator.py: structured Markdown report writer
- main.py: CLI with --query, --report, --interactive flags
- README.md: agent concepts, ReAct loop, architecture diagram, sample queries

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Three-index multimodal RAG pipeline for PDF documents:
- multimodal_parser: pdfplumber extracts text blocks, PNG images, and tables
- text_indexer: FAISS index for text chunks (all-MiniLM-L6-v2)
- image_processor: GPT-4V base64 captioning with graceful fallback
- image_indexer: FAISS index over image captions with image_path metadata
- table_processor: LLM converts 2-D tables to prose descriptions + CSV export
- table_indexer: FAISS index over table descriptions with csv_path metadata
- query_router: LLM classifies query → TEXT / IMAGE / TABLE / ALL
- multi_retriever: fetches from relevant indexes, interleaves and de-duplicates
- generator: modality-labelled prompt → GPT-4 final answer
- main.py: argparse CLI with --skip-images, --skip-tables, --interactive flags

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Complete agentic RAG project with real-time tools:
- FAISS knowledge base indexer (reused from Project 1)
- Tool registry with 5 tools: RAG, finance (yfinance), Wikipedia,
  web search (Tavily), weather (OpenWeatherMap)
- LangChain agent with OPENAI_FUNCTIONS / ReAct loop
- Conversation memory (ConversationBufferWindowMemory k=5)
- Response formatter with reasoning trace display
- CLI with --query, --interactive, --no-memory, --verbose flags
- Mock fallbacks for all optional-key tools
- Comprehensive README with architecture diagram,
  Agentic vs Standard RAG comparison table,
  custom tool walkthrough, and troubleshooting guide

Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Co-authored-by: nerdjerry <7092764+nerdjerry@users.noreply.github.com>
Co-authored-by: nerdjerry <7092764+nerdjerry@users.noreply.github.com>
Copilot AI changed the title [WIP] Build project according to specifications Add all 5 GenAI learning projects from spec Feb 25, 2026
Copilot AI requested a review from nerdjerry February 25, 2026 19:20
@nerdjerry nerdjerry requested a review from Copilot February 25, 2026 20:00
Copy link

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This pull request implements a comprehensive educational curriculum for developers learning Generative AI, consisting of five progressively complex projects that demonstrate key concepts from basic RAG to production-grade agentic systems with real-time data integration.

Changes:

  • Added 5 self-contained GenAI projects with full documentation and working code
  • Added root README.md with learning path, glossary, and setup instructions
  • Added .gitignore to exclude generated files and sensitive data

Reviewed changes

Copilot reviewed 69 out of 78 changed files in this pull request and generated 3 comments.

Show a summary per file
File Description
README.md Project overview with difficulty ratings, learning path, and GenAI glossary
.gitignore Excludes Python artifacts, virtual environments, FAISS indexes, and generated content
01-rag-from-scratch/* Basic RAG implementation with HuggingFace embeddings and FAISS
02-legal-ai-assistant/* Domain-specific RAG with clause extraction, risk analysis, and conflict detection
03-research-agent/* ReAct agent with tools for paper search, summarization, and comparison
04-multimodal-rag/* Multi-index RAG handling text, images (GPT-4V), and tables
05-agentic-rag-realtime/* Production agent with 5 tools including live web search, finance, weather, and Wikipedia

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Co-authored-by: nerdjerry <7092764+nerdjerry@users.noreply.github.com>
@nerdjerry nerdjerry marked this pull request as ready for review February 26, 2026 01:59
@nerdjerry nerdjerry merged commit d2ab96a into copilot/charming-turtle Feb 26, 2026
2 checks passed
@nerdjerry nerdjerry deleted the copilot/build-project-as-per-spec branch February 26, 2026 02:00
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants