Generate GitHub repository recommendations based on your starred repositories and co-stargazers. The tool analyzes your GitHub stars and identifies other repositories that users who starred the same projects also found interesting. Results are saved as JSON files in the data/recommendations/ directory.
 |
|---|
| Preview of the pusblished site of recomened repos |
 |
| Preview of the tag cloud page |
 |
| Preview of the tag overview page |
Show architecture diagram
flowchart LR
A[Start: python main.py] --> B[Load settings from config/settings.yml and env]
B --> C[Fetch starred repos from GitHub API]
C --> D{RECENT_REPOS_LIMIT set?}
D -->|Yes| E[Trim starred repos]
D -->|No| F[Use full starred list]
E --> G[Load existing data from latest.json]
F --> G
G --> H[Find new repos not already processed]
H --> I{Any new repos?}
I -->|Yes| J[Fetch source repo stars/forks from ClickHouse]
J --> K[Generate recommendations per repo via ForkEvent overlap query]
K --> L[Merge new results with existing results]
I -->|No| L
L --> M[Collect all recommended repo names]
M --> N[Fetch recommended repo stars/forks from ClickHouse]
N --> O[Fetch GitHub HTML metadata in parallel: description, topics, languages]
O --> P[Enrich source and recommendation records and compute score]
P --> Q[Write recommendations/latest.json]
Q --> R[Render index.html from templates/index.html with analytics]
R --> S[[DONE]]
All configuration is managed through config/settings.yml. You can override any setting with environment variables.
clickhouse.url: Database URL (default:https://play.clickhouse.com)clickhouse.table: Table containing GitHub events (default:github_events)clickhouse.timeout: Request timeout in seconds (default:60)
processing.recent_repos_limit: Maximum number of starred repositories to analyze (setnullfor no limit, default:null)processing.max_workers: Number of parallel workers for processing (default:4)processing.top_n: Maximum recommendations per repository (default:10)
paths.recommendations_dir: Directory for storing recommendation files (default:data/recommendations)paths.latest_json: Path to the latest recommendations file (default:data/recommendations/latest.json)
user.login: GitHub username to analyze
Run the application with:
python main.pyThe script will:
- Fetch your starred repositories
- Find repositories that share stargazers with your starred repos
- Generate recommendations based on co-stargazing patterns
- Save results to
data/recommendations/latest.json - Render a static
index.htmlusing the Jinja template intemplates/index.html
The output is grouped per starred (source) repository and contains:
-
repo - Repository name (
owner/repo) -
metadata
- description - Repository description
- languages - Language usage breakdown
- topics - GitHub topics
-
total_stars - Total GitHub stars
-
total_forks - Total GitHub forks
For each recommended repository:
-
repo - Repository name (
owner/repo) -
count - Number of overlapping stargazers
-
metadata
- description - Repository description
- languages - Language usage breakdown
- topics - GitHub topics
-
total_stars - Total GitHub stars
-
total_forks - Total GitHub forks
-
score - Overlap ratio (
count / total_stars, 0 if no stars)
Raw Json Data
{
"generated_at": "2026-01-19T05:41:26.596011+00:00",
"username": "SpreadSheets600",
"results": [
{
"repo": "ClickHouse/ClickHouse",
"recommendations": [
{
"repo": "apache/spark",
"count": 646,
"metadata": {
"description": "Apache Spark - A unified analytics engine for large-scale data processing",
"topics": [
"python",
"java",
"r",
"scala",
"sql",
"big-data",
"spark",
"jdbc"
],
"languages": {
"Scala": "67.1%",
"Python": "16.8%",
"Java": "6.5%",
"Jupyter Notebook": "4.9%",
"HiveQL": "2.1%",
"R": "1.4%",
"Other": "1.2%"
}
},
"total_stars": 48802,
"total_forks": 37386,
"score": 0.013237
},
{
"repo": "apache/flink",
"count": 640,
"metadata": {
"description": "Apache Flink",
"topics": [
"python",
"java",
"scala",
"sql",
"big-data",
"flink"
],
"languages": {
"Java": "87.3%",
"Scala": "8.3%",
"Python": "2.8%",
"Shell": "0.5%",
"TypeScript": "0.3%",
"HiveQL": "0.3%",
"Other": "0.5%"
}
},
"total_stars": 27512,
"total_forks": 16656,
"score": 0.023263
},
{
"repo": "kubernetes/kubernetes",
"count": 523,
"metadata": {
"description": "Production-Grade Container Scheduling and Management",
"topics": [
"go",
"kubernetes",
"containers",
"cncf"
],
"languages": {
"Go": "97.4%",
"Shell": "2.3%",
"PowerShell": "0.2%",
"Makefile": "0.1%",
"Dockerfile": "0.0%",
"Python": "0.0%"
}
},
"total_stars": 119934,
"total_forks": 49424,
"score": 0.004361
},
{
"repo": "elastic/elasticsearch",
"count": 468,
"metadata": {
"description": "Free and Open Source, Distributed, RESTful Search Engine",
"topics": [
"java",
"search-engine",
"elasticsearch"
],
"languages": {
"Java": "99.5%",
"Groovy": "0.2%",
"StringTemplate": "0.2%",
"Shell": "0.1%",
"ANTLR": "0.0%",
"C++": "0.0%"
}
},
"total_stars": 74985,
"total_forks": 30967,
"score": 0.006241
},
{
"repo": "pingcap/tidb",
"count": 462,
"metadata": {
"description": "TiDB - the open-source, cloud-native, distributed SQL database designed for modern applications.",
"topics": [
"mysql",
"go",
"sql",
"database",
"scale",
"serverless",
"distributed-transactions",
"distributed-database",
"cloud-native",
"tidb",
"hacktoberfest",
"htap",
"mysql-compatibility"
],
"languages": {
"Go": "94.2%",
"Starlark": "3.0%",
"Shell": "1.6%",
"Yacc": "0.8%",
"Jsonnet": "0.2%",
"TypeScript": "0.1%",
"Other": "0.1%"
}
},
"total_stars": 43052,
"total_forks": 7686,
"score": 0.010731
},
{
"repo": "tensorflow/tensorflow",
"count": 445,
"metadata": {
"description": "An Open Source Machine Learning Framework for Everyone",
"topics": [
"python",
"machine-learning",
"deep-neural-networks",
"deep-learning",
"neural-network",
"tensorflow",
"ml",
"distributed"
],
"languages": {
"C++": "55.7%",
"Python": "25.5%",
"MLIR": "6.3%",
"HTML": "4.2%",
"Starlark": "4.1%",
"Go": "1.2%",
"Other": "3.0%"
}
},
"total_stars": 228592,
"total_forks": 114876,
"score": 0.001947
},
{
"repo": "apache/kafka",
"count": 442,
"metadata": {
"description": "Mirror of Apache Kafka",
"topics": [
"scala",
"kafka"
],
"languages": {
"Java": "87.2%",
"Scala": "10.6%",
"Python": "1.9%",
"Shell": "0.2%",
"Batchfile": "0.1%",
"Dockerfile": "0.0%"
}
},
"total_stars": 33929,
"total_forks": 17900,
"score": 0.013027
},
{
"repo": "facebook/rocksdb",
"count": 420,
"metadata": {
"description": "A library that provides an embeddable, persistent key-value store for fast storage.",
"topics": [
"database",
"storage-engine"
],
"languages": {
"C++": "83.2%",
"Java": "8.8%",
"Starlark": "2.2%",
"C": "1.7%",
"Python": "1.4%",
"Perl": "0.8%",
"Other": "1.9%"
}
},
"total_stars": 33088,
"total_forks": 7862,
"score": 0.012693
},
{
"repo": "torvalds/linux",
"count": 393,
"metadata": {
"description": "Linux kernel source tree",
"topics": [],
"languages": {
"C": "98.0%",
"Assembly": "0.7%",
"Shell": "0.4%",
"Rust": "0.3%",
"Python": "0.3%",
"Makefile": "0.2%",
"Other": "0.1%"
}
},
"total_stars": 239118,
"total_forks": 79165,
"score": 0.001644
},
{
"repo": "golang/go",
"count": 363,
"metadata": {
"description": "The Go programming language",
"topics": [
"go",
"language",
"programming-language",
"golang"
],
"languages": {
"Go": "89.7%",
"Assembly": "5.4%",
"HTML": "4.5%",
"C": "0.2%",
"Shell": "0.1%",
"Perl": "0.1%"
}
},
"total_stars": 152724,
"total_forks": 26080,
"score": 0.002377
}
],
"metadata": {
"description": "ClickHouse\u00ae is a real-time analytics database management system",
"topics": [
"rust",
"embedded",
"sql",
"database",
"big-data",
"ai",
"analytics",
"cpp",
"clickhouse",
"dbms",
"self-hosted",
"distributed",
"olap",
"cloud-native",
"mpp",
"hacktoberfest",
"lakehouse"
],
"languages": {
"C++": "71.8%",
"Python": "9.2%",
"Assembly": "8.9%",
"Shell": "4.1%",
"C": "2.2%",
"Jinja": "1.6%",
"Other": "2.2%"
}
},
"total_stars": 36888,
"total_forks": 7581
}
]
}
- Last recommendations run: 2026-02-15 01:16:47 UTC
- Latest recommendations file: recommendations/latest.json