ScholarScout/config.example.yaml at main · neej4/ScholarScout · 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
# ScholarScout Configuration
# Copy this file to config.yaml and fill in your values.
# config.yaml is gitignored — your API keys stay local.

app:
  max_ideas: 10

llm:
  # Provider: gemini | groq | openrouter | openai | ollama | custom
  provider: gemini
  api_key: ''
  model: gemini-2.0-flash
  # base_url: only needed for custom/ollama providers
  # base_url: http://localhost:11434/v1/chat/completions

# Implementation Discovery (v1.6)
# GitHub token is optional — increases rate limit from 10 to 30 req/min
# You can set it here OR via GITHUB_TOKEN env var.
# Get one at: https://github.com/settings/tokens (no scopes needed for public search)
# github_token: 'ghp_...'

# Kaggle credentials — optional, enables Kaggle dataset search
# You can set them here OR via KAGGLE_USERNAME / KAGGLE_KEY env vars.
# Get from: https://www.kaggle.com/account → Create New Token
# kaggle_username: ''
# kaggle_key: ''

features:
  refine: false
  sensitivity: false
  grounding: false
  cache_expiry_days: 7

# Similarity thresholds (tune these to adjust novelty/grounding sensitivity)
thresholds:
  # Grounding verification (deep_dive.py) — badge levels
  grounding_high: 0.65
  grounding_medium: 0.40

  # Novelty checker — semantic similarity (Gemini embeddings)
  novelty_semantic_similar: 0.82
  novelty_semantic_exists: 0.92

  # Novelty checker — Jaccard fallback
  novelty_jaccard_similar: 0.40
  novelty_jaccard_exists: 0.70

  # Clusterer — minimum similarity to join existing cluster
  clusterer_min_join: 0.15