RAG Series (2): Building Your First RAG Pipeline with LangChain

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant Common SOC 2 Failures (Real World) Stop Vibe-Checking Your AI App: A Practical Guide to Evals How to Use SonarQube and SonarScanner Locally to Level Up Your Code Quality Your Next To-Do App Is Dead — I Replaced Mine with an OpenClaw AI Sign a Nostr event in 60 lines of Python using coincurve — no nostr-sdk, no nbxplorer, no rust toolchain ITGC Audit Explained Like You’re in Big 4 Patch Tuesday abril 2026: Microsoft parcha 163 vulnerabilidades y un zero-day en SharePoint Stop scraping everything: a better way to track competitor price changes Listing on MCPize + the Official MCP Registry while routing payments OUTSIDE the marketplace — how I kept 100% of my x402 revenue Building an AI-Powered Risk Intelligence System Using Serverless Architecture Why We Ripped Function Overloading Out of Our AI Toolchain Testing AI-Generated Code: How to Actually Know If It Works SaaS Churn Is Killing Your Business. Here Is What to Do About It (Without a Support Team) The Speed of AI Is No Longer Linear - And Self-Improving Models Are Why How to Implement RBAC for MCP Tools: A Practical Guide for Engineering Teams From Standard Quote to Persuasive Proposal: AI Automation for Arborists I built a CLI that scaffolds complete multi-tenant SaaS apps Axios CVE-2025–62718: The Silent SSRF Bug That Could Be Hiding in Your Node.js App Right Now The dashboard that ended our friendship Data Pipelines Explained Simply (and How to Build Them with Python) The Hidden Cost of AI Systems Nobody Talks About. undefined vs undeclared, and how typeof behaves Switching from file-based jobs to NATS/Kafka in Rust without changing code io_uring Adventures: Rust Servers That Love Syscalls Why Agentic AI is Killing the Traditional Database The POUR principles of web accessibility for developers and designers Quantum Neural Network 3D — A Deep Dive into Interactive WebGL Visualization How To Install Caveman In Codex On macOS And Windows Automation Pipeline Reliability: Why Your Workflow Breaks When Nobody Is Watching I Built an 'Open World' AI Coding Agent — It Works From ANY Folder From Freelancing to Product: A Tech Service Company's SaaS Transformation China's AI Giants: Adding Tencent Hunyuan & ByteDance Doubao to AI University (74 Providers) On the Vibe Coders and Their Lies clerk: Auto-Summarize Your Claude Code Sessions AI Weekly — 2026/04/10–04/17 | The Model Lockdown Is Here, but the Toolchain Is the Real Battleground AI 週報 — 2026/04/10–2026/04/17 模型封鎖潮來了，但工具鏈才是真戰場 Maybe this is how Open-Source apps are born... 🚀 Fine-Tune LLMs with LoRA and QLoRA: 2026 Guide tRPC v11 + Next.js App Router: End-to-End Type Safety Without the Boilerplate ShadCN UI in 2026: Why I Stopped Installing Component Libraries and Started Owning My Components SaaS Billing in React Server Components: Stripe + Supabase Without a Single `useEffect` Join our DEV Weekend Challenge — $1,000 in Prizes Across TEN winners! Submissions Due April 20 at 6:59 AM UTC. Implementing FSRS Spaced Repetition in Flutter + Supabase — Adding Memory Science to an AI Learning App "I Texted My Localhost From the Train — Claude Code Fixed the Bug Before I Got Home" I Built a Sales Prep AI and It Went Deeper Than Expected Design to Code #2: One JSON, Eleven Outputs Solving the 100M-Row Problem: A Summary Table Pattern for High-Volume Push Notification Logs Flutter Web With Wasm: What Actually Changes For Developers I Built 50 Royalty-Free Soundtracks for My Side Project in a Weekend Using AI Music Generation The Vibe Coding Security Checklist: 7 Things to Check Before You Ship Stop Letting Googlebot Guess Fix Your React App's SEO Right Desconstruindo o Streaming do LinkedIn: Como Criar um Engine de Extração de Vídeo de Alta Performance com HLS e FFmpeg (EDA Part-1) EDA (Exploratory Data Analysis) Explained With Real Life — Why Looking at Your Data Is the Most Important Step in Machine Learning Brand Relationship Management at Scale: Our 4-Touch Outreach System for 200+ Brands Why String.fromEnvironment() Might Return an Empty String in Dart JGuardrails 1.0.0 — Hardening Java LLM Apps Against Jailbreaks, Toxicity, and Prompt Injection Plan and Schedule a Full Week of Threads Content From One Claude Conversation Coding Cat Oran Ep3, Five Tables Changed Everything Updated: BFF Pattern I'm done watching freelancers get buried by 200 proposals. So I'm building the alternative. This is my first post BFS Algorithm in Java Step by Step Tutorial with Examples Tracking LLM Pricing Monthly: An Open Dataset for 22 AI Models How We Measure Content ROI on a Comparison Site: Revenue Attribution Without Perfect Data Introducing Nova AI Ops: The AI-Native Operating System for SRE Teams I built a free desktop video downloader for Windows — Grabbit How Talkie OCR Helps Vision-Impaired & Dyslexic Users Read the World Around Them VRCFaceTracking安装和iPhone面捕配置教程，有bug Even CrowdStrike Can't See Your Agents The Automation Gold Rush: What n8n Workflows and Claude Are Opening Up for Developers Right Now

WonderLab · 2026-05-01 · via DEV Community

<h2> From 100 Lines to a Production Pipeline </h2> In the last article, we built a minimal RAG system in 100 lines of pure Python. It worked. It demonstrated the core idea. But if you tried to productionize that code, you'd quickly run into a wall. Loading a PDF? You need <code>PyPDF2</code> or <code>pdfplumber</code>, and you'll discover that tables, headers, and footers are a nightmare to parse cleanly. Splitting text? Your naive <code>text.split("\n\n")</code> will cut sentences in half, break code blocks, or create chunks so large they blow past the token limit. Swapping the vector database? Say goodbye to your afternoon — every database has a different API, different distance metrics, different metadata handling. Switching LLM providers? OpenAI's client, Anthropic's client, local models via <code>llama.cpp</code> — each has its own message format, its own token counting, its own error handling. This is exactly why LangChain exists. It doesn't do anything magical. It doesn't replace the underlying models or databases. What it does is simple and valuable: it gives you a uniform interface for plugging components together, so you can focus on the logic of your RAG system instead of the plumbing. In this article, we'll rebuild our RAG pipeline using LangChain's modern API. By the end, you'll have a complete, runnable project that loads PDFs, splits them intelligently, stores them in ChromaDB, and answers questions using a multi-provider LLM — with about 30 lines of actual pipeline code. <blockquote> LangChain Version Note: The code in this article is based on <code>langchain 1.x</code> (current stable). LangChain 1.x performed a breaking reorganization of <code>0.3.x</code> — high-level APIs like <code>create_retrieval_chain</code> were removed. We use LCEL native syntax (<code>|</code> pipe operator) instead, which is functionally equivalent and version-agnostic. Full source code: <a href="https://github.com/chendongqi/llm-in-action/tree/main/02-langchain-basic" rel="noopener noreferrer">https://github.com/chendongqi/llm-in-action/tree/main/02-langchain-basic</a> </blockquote> <h2> The Six Moving Parts of a RAG Pipeline </h2> LangChain decomposes RAG into six components. Understanding what each one does — and where the quality risks hide — is the key to debugging RAG systems later. <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Component</th> <th>Role</th> <th>The Quality Risk</th> </tr> </thead> <tbody> <tr> <td>Document Loader</td> <td>Reads raw files (PDF, Word, Markdown, HTML) and extracts text</td> <td>Tables, images, and weird formatting get mangled</td> </tr> <tr> <td>Text Splitter</td> <td>Cuts long documents into semantically coherent chunks</td> <td>Chunks too large = low precision; too small = lost context</td> </tr> <tr> <td>Embedding Model</td> <td>Converts text chunks into high-dimensional vectors</td> <td>Poor model choice = semantically unrelated texts cluster together</td> </tr> <tr> <td>Vector Store</td> <td>Persists vectors and enables fast similarity search</td> <td>Wrong distance metric or no metadata filtering = bad retrieval</td> </tr> <tr> <td>Retriever</td> <td>Accepts a query, searches the vector store, returns relevant chunks</td> <td>Top-K too low = missing information; too high = noisy context</td> </tr> <tr> <td>Chain</td> <td>Orchestrates the full flow: query → retrieve → prompt → LLM → answer</td> <td>Prompt design and context assembly determine answer quality</td> </tr> </tbody> </table></div> Think of these six components as an assembly line in a factory. The Document Loader is the raw material intake. The Text Splitter is the precision cutting station. The Embedding Model and Vector Store form the warehouse and inventory system. The Retriever is the picker who fetches the right parts. The Chain is the foreman who coordinates everything and delivers the final product. If any station on the line is misconfigured, the final product suffers — and the tricky part is that the failure often looks like an LLM problem when it's actually a retrieval problem. <h2> Hands-On: A Complete LangChain RAG Project </h2> Let's build it. We'll create a RAG system that reads a directory of PDF files, indexes them, and lets you ask questions in natural language. <h3> Project Structure </h3> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>rag-project/ ├── requirements.txt ├── data/ │ └── sample.pdf # Your PDF documents go here └── rag_pipeline.py </code></pre> </div> <h3> Step 0: Dependencies </h3> <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>langchain>=0.3.0 langchain-text-splitters>=0.3.0 langchain-openai>=0.2.0 langchain-chroma>=0.1.0 langchain-community>=0.3.0 pypdf>=4.0.0 python-dotenv>=1.0.0 </code></pre> </div> <blockquote> Full source code (ready to run): <a href="https://github.com/chendongqi/llm-in-action/tree/main/02-langchain-basic" rel="noopener noreferrer">https://github.com/chendongqi/llm-in-action/tree/main/02-langchain-basic</a> Supports Zhipu AI / OpenAI / Ollama for LLM, SiliconFlow and local Ollama for Embedding. </blockquote> Install them: <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>pip install -r requirements.txt </code></pre> </div> Configure your API keys (copy <code>.env.example</code> to <code>.env</code>): <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>cp .env.example .env # Edit .env to fill in LLM_API_KEY and EMBEDDING_API_KEY </code></pre> </div> Supported Providers: <ul> <li> LLM: Zhipu AI (default), OpenAI, SiliconFlow, Ollama, Azure</li> <li> Embedding: SiliconFlow (default), OpenAI, Ollama</li> </ul> <h3> Step 1: Load Documents </h3> The <code>PyPDFLoader</code> handles PDF parsing for us. It extracts text page by page and returns a list of <code>Document</code> objects, each containing the page content and metadata (page number, source file, etc.). <div class="highlight js-code-highlight"> <pre class="highlight python"><code>from langchain_community.document_loaders import PyPDFLoader from pathlib import Path def load_pdfs(data_dir: str = "./data"): """Load all PDF files from the data directory.""" documents = [] pdf_paths = list(Path(data_dir).glob("*.pdf")) for pdf_path in pdf_paths: loader = PyPDFLoader(str(pdf_path)) pages = loader.load() documents.extend(pages) print(f"Loaded '{pdf_path.name}': {len(pages)} pages") print(f"Total documents loaded: {len(documents)}") return documents </code></pre> </div> <blockquote> Real-world note: PDFs are the wild west of document formats. If your PDFs contain scanned images, you'll need OCR (via <code>pdfplumber</code> with Tesseract or Azure Document Intelligence). If they contain tables, consider <code>UnstructuredPDFLoader</code> which preserves table structure better than raw text extraction. </blockquote> <h3> Step 2: Split Documents into Chunks </h3> Remember the chunking problem from Part 1? LangChain's <code>RecursiveCharacterTextSplitter</code> is the industry default for good reason. It tries to split on natural boundaries — paragraphs first, then newlines, then sentences, then words — so it avoids cutting mid-sentence whenever possible. <div class="highlight js-code-highlight"> <pre class="highlight python"><code>from langchain_text_splitters import RecursiveCharacterTextSplitter def split_documents(documents, chunk_size=200, chunk_overlap=30): """ Split documents into overlapping chunks. chunk_overlap ensures context continuity between adjacent chunks. """ splitter = RecursiveCharacterTextSplitter( chunk_size=chunk_size, chunk_overlap=chunk_overlap, length_function=len, separators=["\n\n", "\n", ". ", " ", ""] ) chunks = splitter.split_documents(documents) print(f"Split into {len(chunks)} chunks (chunk_size={chunk_size}, overlap={chunk_overlap})") return chunks </code></pre> </div> Why <code>chunk_overlap</code> matters: If a key concept spans two chunks — say, "The API rate limit is 100 requests per minute. Exceeding this limit returns a 429 status code" — an overlap of 50 characters ensures the second chunk still contains "100 requests per minute" as context. Without overlap, the retriever might fetch only one chunk and miss the causal relationship. Chunk size trade-offs: <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Chunk Size</th> <th>Precision</th> <th>Context</th> <th>Best For</th> </tr> </thead> <tbody> <tr> <td>256 tokens</td> <td>High</td> <td>Minimal</td> <td>Fact lookup, Q&A over structured docs</td> </tr> <tr> <td>512 tokens</td> <td>Balanced</td> <td>Moderate</td> <td>General-purpose RAG (good default)</td> </tr> <tr> <td>1024 tokens</td> <td>Lower</td> <td>Rich</td> <td>Long-form summarization, narrative documents</td> </tr> <tr> <td>2048+ tokens</td> <td>Low</td> <td>Very rich</td> <td>Only when the LLM context window is large and queries are broad</td> </tr> </tbody> </table></div> For most use cases, 512 tokens with 50-token overlap is a safe starting point. <h3> Step 3: Embed and Store in ChromaDB </h3> Now we convert each chunk into a vector and store them. ChromaDB is our choice here because it's persistent (data survives restarts), supports metadata filtering, and requires zero setup — it runs locally as an embedded database. <div class="highlight js-code-highlight"> <pre class="highlight python"><code>from langchain_openai import OpenAIEmbeddings from langchain_chroma import Chroma import os persist_directory = "./chroma_db" def build_vector_store(chunks): """Create embeddings and store in ChromaDB.""" embeddings = OpenAIEmbeddings( model="BAAI/bge-large-zh-v1.5", # SiliconFlow Chinese model api_key=os.getenv("EMBEDDING_API_KEY"), base_url="https://api.siliconflow.cn/v1", dimensions=1024, chunk_size=32 # SiliconFlow limit: max 32 per batch ) if os.path.exists(persist_directory): import shutil shutil.rmtree(persist_directory) vector_store = Chroma.from_documents( documents=chunks, embedding=embeddings, persist_directory=persist_directory ) print(f"Vector store built: {vector_store._collection.count()} vectors persisted") return vector_store </code></pre> </div> Embedding model note: We use <code>BAAI/bge-large-zh-v1.5</code> on SiliconFlow (excellent for Chinese). Access via the OpenAI-compatible interface in <code>langchain_openai</code>. Switch to <code>text-embedding-3-small</code> for OpenAI. The <code>chunk_size=32</code> parameter is critical — it's SiliconFlow's batch limit (max 32 per request), while most other providers default to 1000. <h3> Step 4: Build the Retriever </h3> The retriever is a thin wrapper around the vector store that handles the search logic. By default, it performs similarity search and returns the top-K most relevant chunks. <div class="highlight js-code-highlight"> <pre class="highlight python"><code>def get_retriever(vector_store, search_k=4): """Configure retriever with similarity search.""" retriever = vector_store.as_retriever( search_type="similarity", search_kwargs={"k": search_k} ) return retriever </code></pre> </div> <h3> Step 5: Build the RAG Chain </h3> Here's where LangChain's modern API shines. Instead of the older <code>RetrievalQA</code> class, we use LCEL (LangChain Expression Language) to compose a chain that is explicit, debuggable, and easy to modify. <div class="highlight js-code-highlight"> <pre class="highlight python"><code>from langchain_openai import ChatOpenAI from langchain_core.prompts import ChatPromptTemplate from langchain_core.output_parsers import StrOutputParser from langchain_core.runnables import RunnablePassthrough def build_rag_chain(retriever): """ Build the full RAG chain using LCEL (langchain 1.x compatible). retrieve → format_docs → assemble prompt → LLM → StrOutputParser """ llm = ChatOpenAI( model="glm-4-flash", # Zhipu AI, via SiliconFlow or direct api_key=os.getenv("LLM_API_KEY"), base_url="https://open.bigmodel.cn/api/paas/v4", temperature=0 ) # System prompt. {context} filled by format_docs, {question} by the user's input. system_prompt = ( "You are a precise knowledge assistant. Answer the user's question " "based solely on the provided reference content below. " "If the reference content does not contain the answer, say so clearly " "— do not make anything up.\n\n" "Reference content:\n{context}" ) prompt = ChatPromptTemplate.from_messages([ ("system", system_prompt), ("human", "{question}") ]) # Helper: convert Document list to a single string for {context} def format_docs(docs: list) -> str: return "\n\n".join(doc.page_content for doc in docs) # LCEL Chain: compose components with the pipe operator # 1. {"context": retriever | format_docs, "question": RunnablePassthrough()} # → retriever fetches docs, format_docs converts them to a string # 2. | prompt → assembles into a full prompt # 3. | llm → generates the answer # 4. | StrOutputParser() → returns plain text (not an AIMessage object) rag_chain = ( { "context": retriever | format_docs, "question": RunnablePassthrough() } | prompt | llm | StrOutputParser() ) return rag_chain </code></pre> </div> What's happening here? We're using LCEL (LangChain Expression Language) native syntax — the pipe <code>|</code> operator — instead of the high-level <code>create_retrieval_chain</code> (which was removed in langchain 1.x). The key line is <code>retriever | format_docs</code>: the retriever outputs a list of <code>Document</code> objects, <code>format_docs</code> converts them to a string that fills the <code>{context}</code> placeholder. <code>RunnablePassthrough()</code> passes the user's raw question through to the <code>{question}</code> placeholder. Three lines of declarative code that are functionally equivalent to the 50 lines of imperative Python from Part 1. <h3> Step 6: Query the Pipeline </h3> <div class="highlight js-code-highlight"> <pre class="highlight python"><code>def query(rag_chain, question: str, retriever): """Run a question through the RAG pipeline, print answer and sources.""" print(f"\nQuestion: {question}") answer = rag_chain.invoke(question) # LCEL chain returns plain text directly print(f"\nAnswer:\n{answer}") # Retrieve docs separately to show sources (rag_chain doesn't expose them) docs = retriever.invoke(question) print("\nRetrieved sources:") for i, doc in enumerate(docs, 1): source = doc.metadata.get("source", "unknown") page = doc.metadata.get("page", "?") preview = doc.page_content[:120].replace("\n", " ") print(f" [{i}] {source} (page {page}): {preview}...") return answer </code></pre> </div> <h3> Putting It All Together </h3> <div class="highlight js-code-highlight"> <pre class="highlight python"><code>if __name__ == "__main__": # 1. Load docs = load_pdfs("./data") # 2. Split (smaller chunk_size for short PDF pages) chunks = split_documents(docs, chunk_size=200, chunk_overlap=30) # 3. Embed & Store vector_store = build_vector_store(chunks) # 4. Retrieve retriever = get_retriever(vector_store, search_k=4) # 5. Build Chain (LCEL) rag_chain = build_rag_chain(retriever) # 6. Interactive Q&A while True: user_input = input("\nYour question (quit to exit): ").strip() if user_input.lower() in ("quit", "exit", "q"): break if user_input: query(rag_chain, user_input, retriever) </code></pre> </div> <h2> Running the Pipeline </h2> Place a PDF in <code>data/sample.pdf</code> and run: <div class="highlight js-code-highlight"> <pre class="highlight shell"><code>python rag_pipeline.py </code></pre> </div> Sample output: <div class="highlight js-code-highlight"> <pre class="highlight plaintext"><code>RAG Pipeline 启动 LLM Provider : zhipu LLM Model : glm-4-flash Embedding : openai / BAAI/bge-large-zh-v1.5 数据目录 : ./data 向量库 : ./chroma_db ================================================== 已加载 'Automotive-SPICE-PAM-v40.pdf'：153 页共加载 153 个文档片段切分为 2000 个块（chunk_size=200，overlap=30） [Embedding] Provider: openai | Model: BAAI/bge-large-zh-v1.5 | Base: https://api.siliconflow.cn/v1 已清除旧向量库：./chroma_db 向量库构建完成：2000 个向量已持久化 ================================================== RAG Pipeline 构建完成！输入问题开始问答（输入 'quit' 退出） ================================================== 你的问题：What is Automotive SPICE? ================================================== 问题：What is Automotive SPICE？ ================================================== 答案： Automotive SPICE（Automotive Software Process Improvement and Capability Determination）是一种用于评估和改进汽车软件开发过程能力的框架。它定义了软件开发生命周期中的关键过程域，并建立了过程能力的等级评估标准... 检索到的来源： [1] ./data/Automotive-SPICE-PAM-v40.pdf（第 5 页）：Automotive SPICE Process Assessment Model The Process Assessment Model (PAM) defines the processes... </code></pre> </div> Notice how the answer includes citations — we know exactly which pages the information came from. This traceability is critical for production RAG systems where users need to verify claims. <h2> What Changed From Our 100-Line Version? </h2> Let's compare the hand-written RAG from Part 1 with our LangChain pipeline: <div class="table-wrapper-paragraph"><table> <thead> <tr> <th>Aspect</th> <th>Hand-written (Part 1)</th> <th>LangChain (Part 2)</th> </tr> </thead> <tbody> <tr> <td>PDF loading</td> <td>Not supported</td> <td>One-line <code>PyPDFLoader</code> </td> </tr> <tr> <td>Text splitting</td> <td>None (whole docs)</td> <td> <code>RecursiveCharacterTextSplitter</code> with smart boundaries</td> </tr> <tr> <td>Vector persistence</td> <td>In-memory only, lost on restart</td> <td>ChromaDB persists to disk</td> </tr> <tr> <td>Embedding model swap</td> <td>Rewrite API calls</td> <td>One-line parameter change</td> </tr> <tr> <td>LLM swap</td> <td>Rewrite client code</td> <td>One-line parameter change</td> </tr> <tr> <td>Vector DB swap</td> <td>Rewrite storage + search</td> <td>Swap <code>Chroma</code> for <code>Qdrant</code> / <code>Pinecone</code> </td> </tr> <tr> <td>Prompt engineering</td> <td>Raw string formatting</td> <td>Templated prompts with <code>ChatPromptTemplate</code> </td> </tr> <tr> <td>Source citations</td> <td>Manual tracking</td> <td>Automatic metadata propagation</td> </tr> <tr> <td>Chain construction</td> <td>Manual <code>retrieve + generate</code> </td> <td>LCEL `</td> </tr> <tr> <td>Lines of pipeline code</td> <td>~80</td> <td>~25</td> </tr> </tbody> </table></div> The abstraction doesn't hide complexity — it isolates it. <blockquote> LangChain Version Compatibility: The code in this article targets {% raw %}<code>langchain 1.x</code> (current stable). LangChain 1.x performed a breaking reorganization of <code>0.3.x</code> — <code>create_retrieval_chain</code> and <code>create_stuff_documents_chain</code> were removed. We use LCEL native syntax (<code>|</code> pipe operator) instead, which is functionally equivalent and version-agnostic. When you need to debug retrieval quality, you know exactly which component to tune. When you need to swap the embedding model for a cheaper alternative, you change one line. When your data grows beyond ChromaDB's capabilities, you switch to Qdrant without touching the rest of the pipeline. </blockquote> <h2> Common Pitfalls at Each Stage </h2> <h3> Loader Pitfall: "My PDF has tables and they come out garbled" </h3> Raw PDF text extraction flattens tables into a stream of numbers. For table-heavy documents, use <code>UnstructuredPDFLoader</code> or <code>AzureAIDocumentIntelligenceLoader</code> which preserves structural relationships. <h3> Splitter Pitfall: "The answer is split across two chunks and the model only sees half" </h3> Increase <code>chunk_overlap</code> to 100-150 tokens, or reduce <code>chunk_size</code> so that key concepts fit within a single chunk. Better yet, use Parent-Document Retrieval (covered in a later article) which retrieves small chunks but returns the full parent document for context. <h3> Embedding Pitfall: "Questions and documents don't match even though they should" </h3> This is the "asymmetric retrieval" problem. A user asks "How do I reset my password?" but the document says "To reset your password, navigate to Settings → Security." The question and the answer embed to different vectors because their surface text differs. Solutions: use a model fine-tuned for Q&A retrieval (like BGE-M3), or generate hypothetical answers for retrieval (HyDE — also covered later). <h3> Retriever Pitfall: "Top-K=4 isn't enough for complex questions" </h3> If a question requires synthesizing information from five different sections of a document, <code>k=4</code> will miss one. But increasing <code>k</code> blindly adds noise. A better approach: use Multi-Query Retrieval (generate 3 variants of the question, retrieve for each, deduplicate) or Reranking (retrieve 20, then use a cross-encoder to pick the best 5). <h3> Chain Pitfall: "The model ignores the context and hallucinates" </h3> Your prompt matters. The system prompt must explicitly instruct the model to use only the provided context. Adding "If the reference content does not contain the answer, say so clearly — do not make anything up" dramatically improves faithfulness. We'll measure this quantitatively with RAGAS in the evaluation articles. <h2> Summary </h2> In this article, we took the raw RAG concept from Part 1 and wrapped it in a production-ready framework. Here's what we covered: <ol> <li> The six components of a LangChain RAG pipeline — Loader, Splitter, Embedding, Vector Store, Retriever, and Chain — and what quality risk hides in each one.</li> <li> A complete, runnable project that loads PDFs, splits them with <code>RecursiveCharacterTextSplitter</code>, embeds them with OpenAI, stores them in ChromaDB, and answers questions via a LangChain LCEL chain.</li> <li> The chunk size trade-off — in real projects, PDF pages may be very short (e.g., 200 chars). A <code>chunk_size=512</code> default can produce 0 chunks. 200 + 30 overlap is the safe default.</li> <li> Common pitfalls at each pipeline stage, from mangled PDF tables to asymmetric retrieval mismatches.</li> </ol> The code in this article is a solid foundation. It handles real PDFs, persists data, and gives you source citations. But it's still a naive RAG pipeline — one query, one retrieval pass, one answer. In the next articles, we'll add the components that separate toy demos from production systems: hybrid search, reranking, query optimization, and evaluation frameworks. <h2> References </h2> <ul> <li> <a href="https://python.langchain.com/docs/tutorials/rag/" rel="noopener noreferrer">LangChain RAG Tutorial</a> — Official LangChain RAG quickstart</li> <li> <a href="https://python.langchain.com/docs/concepts/lcel/" rel="noopener noreferrer">LangChain Expression Language (LCEL)</a> — Why and how to use LCEL for composable chains</li> <li> <a href="https://docs.trychroma.com/" rel="noopener noreferrer">ChromaDB Documentation</a> — Vector store setup, persistence, and querying</li> </ul>

此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。原文来自 — 版权归原作者所有。

推荐订阅源

DEV Community