data-weaver/docs/tutorials/04-llm-classification.yaml at main · netsirius/data-weaver · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
# Tutorial 4: LLM-powered classification
# Reads support tickets from CSV and classifies them using an LLM
#
# Requires: ANTHROPIC_API_KEY, OPENAI_API_KEY, or GEMINI_API_KEY
# Or use a local Ollama model (no API key needed):
#   Change provider to "local" and model to "llama3"
#
# Run: weaver apply docs/tutorials/04-llm-classification.yaml

name: LLMClassification
tag: tutorial
engine: local

dataSources:
  - id: tickets
    type: File
    config:
      path: docs/tutorials/data/sample_tickets.csv
      format: csv

transformations:
  - id: classified
    type: LLMTransform
    sources:
      - tickets
    config:
      provider: gemini
      model: gemini-2.0-flash
      prompt: |
        Classify this support ticket:
        Title: {title}
        Description: {description}

        Return JSON with:
        - category: bug, feature_request, question, or complaint
        - priority: low, medium, or high
        - sentiment: positive, neutral, or negative
      inputColumns: title,description
      outputSchema: category:string|priority:string|sentiment:string
      batchSize: "1"
      retryOnError: "3"

  - id: quality
    type: DataQuality
    sources:
      - classified
    checks:
      - row_count > 0
    onFail: warn

sinks:
  - id: report
    type: File
    source: quality
    config:
      path: output/classified_tickets
      format: json
      saveMode: Overwrite
      coalesce: "1"

profiles:
  local:
    engine: local
    transformations.classified.config.provider: local
    transformations.classified.config.model: llama3

tests:
  - name: "tickets classified"
    assert: report.row_count > 0