1  """
 2  RAG Quick Start
 3  Easy to use way to get started with RAG using YOUR data
 4  
 5  For a complete application see this: https://github.com/neuml/rag
 6  
 7  TxtAI has many example notebooks covering everything the framework provides
 8  Examples: https://neuml.github.io/txtai/examples
 9  
10  Install TxtAI
11    pip install txtai[pipeline-data]
12  """
13  
14  # pylint: disable=C0103
15  import os
16  
17  from txtai import Embeddings, RAG
18  from txtai.pipeline import Textractor
19  
20  # Step 1: Collect files from local directory
21  #
22  # Defaults to "data". Set to whereever your files are.
23  path = "data"
24  files = [os.path.join(path, f) for f in os.listdir(path) if os.path.isfile(os.path.join(path, f))]
25  
26  # Step 2: Text Extraction / Chunking
27  #
28  # Using section based chunking here. More complex options available such as semantic chunking, iterative chunking etc.
29  # Documentation: https://neuml.github.io/txtai/pipeline/data/textractor
30  # Supports Chonkie chunking as well: https://docs.chonkie.ai/oss/chunkers/overview
31  textractor = Textractor(backend="docling", sections=True)
32  chunks = []
33  for f in files:
34      for chunk in textractor(f):
35          chunks.append((f, chunk))
36  
37  # Step 3: Build an embeddings database
38  #
39  # The `path` parameter sets the vector embeddings model. Supports Hugging Face models, llama.cpp, Ollama, vLLM and more.
40  # Documentation: https://neuml.github.io/txtai/embeddings/
41  embeddings = Embeddings(content=True, path="Qwen/Qwen3-Embedding-0.6B", maxlength=2048)
42  embeddings.index(chunks)
43  
44  # Step 4: Create RAG pipeline
45  #
46  # Combines an embeddings database and an LLM.
47  # Supports Hugging Face models, llama.cpp, Ollama, vLLM and more
48  # Documentation: https://neuml.github.io/txtai/pipeline/text/rag
49  
50  # User prompt template
51  template = """
52    Answer the following question using the provided context.
53  
54    Question:
55    {question}
56  
57    Context:
58    {context}
59  """
60  
61  rag = RAG(
62      embeddings,
63      "Qwen/Qwen3-0.6B",
64      system="You are a friendly assistant",
65      template=template,
66      output="flatten",
67  )
68  
69  question = "Summarize the main advancements made by BERT"
70  print(rag(question, maxlength=2048, stripthink=True))