Configuration Overview

Configure Simili Bot v0.2.0

Configuration Overview

Simili Bot is configured using a YAML file. Your configuration structure depends on your chosen deployment pattern.

Configuration by pattern

Choose your setup guide to see pattern-specific configuration examples:

Pattern A

Single Repository

Pattern B

Centralized Multi-Repo

Pattern C

Distributed Multi-Repo

Configuration file location

By default, Simili Bot looks for configuration in these locations:

1. Explicit path: Specified via CLI or workflow

   simili process --config /path/to/simili.yaml
   

2. GitHub Actions: Path specified in workflow

   - uses: similigh/simili-bot@v0.2.0
     with:
      config_path: ".github/simili.yaml"
   

3. Current directory: simili.yaml in working directory

Configuration structure

A complete configuration file has these main sections:

# Search backend (qdrant | github_native | bm25)
search:
  backend: "qdrant"

# Vector database connection (required for qdrant)
qdrant:
  url: "https://your-cluster.qdrant.io:6333"
  api_key: "${QDRANT_API_KEY}"
  collection: "issues"

# AI model settings (required for qdrant)
embedding:
  provider: "gemini"
  api_key: "${GEMINI_API_KEY}"
  model: "gemini-embedding-001"
  dimensions: 3072

# LLM settings
llm:
  provider: "gemini"
  api_key: "${GEMINI_API_KEY}"

# Default behavior
defaults:
  similarity_threshold: 0.65
  max_similar_to_show: 5
  cross_repo_search: true

# Claude Code integration
claude_code:
  enabled: true
  issue_implement:
    enabled: true

# Duplicate auto-closure
auto_close:
  grace_period_hours: 72

# Issue routing rules
transfer:
  enabled: true
  strategy: "hybrid"

Section breakdown

Search Backend

Choose your similarity search implementation:

search:
  backend: "qdrant" # qdrant, github_native, or bm25
  bm25_fallback: true

Details in Search Backends.

Qdrant Configuration

Configure your vector database connection:

qdrant:
  url: "https://your-cluster.qdrant.io:6333"
  api_key: "${QDRANT_API_KEY}"
  collection: "issues"
  tls: true
  timeout: 30

Property	Type	Description	Required
url	string	Qdrant instance URL with port	Yes
api_key	string	API key (use env vars)	Yes
collection	string	Collection name for issues	Yes
tls	boolean	Use TLS for connection	No (default: true)
timeout	number	Request timeout in seconds	No (default: 30)

Embedding configuration

Configure AI embeddings:

embedding:
  provider: "gemini"
  api_key: "${GEMINI_API_KEY}"
  model: "gemini-embedding-001"
  dimensions: 3072

Property	Type	Description	Required
provider	string	Only “gemini” supported	Yes
api_key	string	Gemini API key (use env vars)	Yes
model	string	Model name	No (default: gemini-embedding-001)
dimensions	number	Vector dimensions	No (default: 768)

Defaults configuration

Set default behavior for issue processing:

defaults:
  similarity_threshold: 0.65
  max_similar_to_show: 5
  cross_repo_search: false

Property	Type	Description	Default
similarity_threshold	float (0-1)	Min score to show similar issue	0.70
max_similar_to_show	number	Max related issues to display	5
cross_repo_search	boolean	Search across all repos	false

Repositories configuration

List repositories to manage:

repositories:
  - org: "my-org"
    repo: "backend"
    enabled: true
    description: "Backend services and APIs"
    labels: ["backend", "api"]

  - org: "my-org"
    repo: "frontend"
    enabled: false  # Disable temporarily
    description: "Web UI and client applications"

Property	Type	Description	Required
org	string	GitHub organization or username	Yes
repo	string	Repository name	Yes
enabled	boolean	Enable Simili Bot for this repo	No (default: true)
description	string	Description for routing decisions	No
labels	array	Associated labels	No

Transfer configuration

Configure issue routing:

transfer:
  enabled: true
  llm_routing_enabled: true
  rules:
    - name: "Route to docs"
      priority: 10
      target: "org/docs"
      title_contains: ["documentation", "docs"]

Details in Transfer Rules.

Claude Code Configuration

Configure AI-powered workflows:

claude_code:
  enabled: true
  issue_implement:
    enabled: true
    trigger_label: "implement"

Details in Claude Code Integration.

Auto-Close Configuration

Automatically manage duplicate issues:

auto_close:
  grace_period_hours: 72
  dry_run: false

Details in Auto-Close Duplicates.

Environment Variables

Use environment variables for sensitive data:

qdrant:
  api_key: "${QDRANT_API_KEY}"

embedding:
  api_key: "${GEMINI_API_KEY}"

Set variables before running:

export QDRANT_API_KEY="your-key"
export GEMINI_API_KEY="your-key"
simili process --config simili.yaml

Or in GitHub Actions:

env:
  QDRANT_API_KEY: ${{ secrets.QDRANT_API_KEY }}
  GEMINI_API_KEY: ${{ secrets.GEMINI_API_KEY }}

Configuration inheritance

Extend configuration from a parent file:

extends: "https://raw.githubusercontent.com/my-org/configs/main/base.yaml"

# Override specific settings
repositories:
  - org: "my-org"
    repo: "my-service"
    enabled: true

Benefits:

• Organization-wide defaults
• Consistent policies across repositories
• Easy updates to shared config
• Per-repository customization

Parent configuration is fetched via HTTPS and merged.

Common configurations

Single repository

Manage one repository only:

qdrant:
  url: "https://your-cluster.qdrant.io:6333"
  api_key: "${QDRANT_API_KEY}"
  collection: "issues"

embedding:
  provider: "gemini"
  api_key: "${GEMINI_API_KEY}"

defaults:
  similarity_threshold: 0.65
  max_similar_to_show: 5
  cross_repo_search: false

repositories:
  - org: "my-username"
    repo: "my-project"
    enabled: true
    description: "My project"

Organization with multiple repos

extends: "https://raw.githubusercontent.com/my-org/policies/main/base.yaml"

repositories:
  - org: "my-org"
    repo: "backend"
    enabled: true
    description: "Backend services"

  - org: "my-org"
    repo: "frontend"
    enabled: true
    description: "Web UI"

  - org: "my-org"
    repo: "infrastructure"
    enabled: true
    description: "DevOps and infrastructure"

Development vs production

Development config (.github/simili-dev.yaml):

qdrant:
  url: "https://dev-cluster.qdrant.io:6333"
  api_key: "${QDRANT_DEV_KEY}"

defaults:
  similarity_threshold: 0.80
  max_similar_to_show: 3

Production config (.github/simili-prod.yaml):

qdrant:
  url: "https://prod-cluster.qdrant.io:6333"
  api_key: "${QDRANT_PROD_KEY}"

defaults:
  similarity_threshold: 0.65
  max_similar_to_show: 5

Validation

Simili Bot validates configuration on startup. Common errors:

Missing required fields:

Error: qdrant.url is required
Error: embedding.api_key is required

Invalid YAML syntax:

Error: Failed to parse configuration: yaml: line 5: mapping values are not allowed in this context

Invalid threshold value:

Error: similarity_threshold must be between 0.0 and 1.0

Configuration updates

To update configuration

1. Edit simili.yaml
2. Commit and push to repository
3. Next issue processed uses new configuration
4. No restart required (GitHub Action runs with latest config)

Safe testing

Test changes before deploying:

# Test with dry-run mode
simili process --issue test.json --config simili.yaml --dry-run

Or in workflow:

- uses: similigh/simili-bot@v0.2.0
  with:
    command: "process"
    config_path: ".github/simili.yaml"
    dry_run: true

Next Steps

Qdrant setup

Configure vector database

Gemini setup

Configure AI engine

Repositories

Configure multiple repos

Transfer rules

Setup issue routing

Claude Code

AI Agent workflows

Search Backends

Qdrant vs GitHub Native