UNDERCODE COMMUNITY
2.67K subscribers
1.23K photos
31 videos
2.65K files
79.5K links
πŸ¦‘ Undercode Cyber World!
@UndercodeCommunity


1️⃣ World first platform which Collect & Analyzes every New hacking method.
+ AI Pratice
@Undercode_Testing

2️⃣ Cyber & Tech NEWS:
@Undercode_News

3️⃣ CVE @Daily_CVE

✨ Web & Services:
β†’ Undercode.help
Download Telegram
Forwarded from UNDERCODE TESTING
▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁

πŸ¦‘ Hack GraphRAG implementation:

Nano-GraphRAG is a lightweight, efficient, and portable implementation of GraphRAG. It's designed to be more user-friendly and hackable than the official version while retaining the core features.

Key Features
- Small yet powerful: ~1100 lines of core code (excluding tests/prompts).
- Highly portable: Supports integrations with Faiss, Neo4j, Ollama, etc.
- Asynchronous and fully typed: Suitable for modern Python workflows.
- Flexible backends:
- LLM: OpenAI, Amazon Bedrock, Ollama, others.
- Vector DB: Built-in nano-vectordb, HNSWlib, Milvus-lite, Faiss.
- Graph Storage: NetworkX (default), Neo4j (optional).
- Incremental and batch data handling: Easy to scale as data grows.



Installation
From Source (Recommended)
 Clone the repository
git clone https://github.com/gusye1234/nano-graphrag.git
cd nano-graphrag
pip install -e .


From PyPI
pip install nano-graphrag




Quick Start
1. Set up API Keys:
- OpenAI API:
     export OPENAI_API_KEY="sk-..."

- For Azure OpenAI or Amazon Bedrock, refer to .env.example.

2. Download a sample text:
   curl https://raw.githubusercontent.com/gusye1234/nano-graphrag/main/tests/mock_data.txt > ./book.txt


3. Basic Usage:
   from nano_graphrag import GraphRAG, QueryParam

Initialize GraphRAG
graph_func = GraphRAG(working_dir="./dickens")

Insert text data
with open("./book.txt") as f:
graph_func.insert(f.read())

Perform a global search
print(graph_func.query("What are the top themes in this story?"))

Perform a local search for scalability
print(graph_func.query("What are the top themes in this story?", param=QueryParam(mode="local")))


4. Reuse Stored Contexts:
- Reinitialize with the same working_dir to reload previous contexts.



Advanced Features
- Batch Insert:
  graph_func.insert(["TEXT1", "TEXT2", ...])


- Incremental Insert:
  graph_func.insert("New text to add")


- Async Support:
  await graph_func.ainsert("Async text")
await graph_func.aquery("What is this text about?")


- Customizable Parameters:
- Use help(GraphRAG) and help(QueryParam) to explore.



Components and Extensibility
Built-in Components:
- LLM: OpenAI, Amazon Bedrock.
- Vector DB: Nano-vectordb, HNSWlib.
- Graph Storage: NetworkX, Neo4j.

Example Extensions:
- Embedding: Sentence-transformers, custom embeddings.
- Visualization: GraphML-based visualizations.
- Chunking: Token-based or custom splitters.



Benchmarking and Contributions
- Benchmarks for performance comparisons are available under examples/benchmarks.
- Contributions for new components and features are welcomed!

Nano-GraphRAG is ideal for developers seeking a nimble and efficient RAG solution, especially for those who want the flexibility to experiment or build on top of it.

▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁
Forwarded from UNDERCODE TESTING
▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁

πŸ¦‘Google Dork : intext:"siemens" & inurl:"/portal/portal.mwsl"

locate Siemens S7 PLC (Programmable Logic Controller) web interfaces through publicly accessible search

This Google dork, intext:"siemens" & inurl:"/portal/portal.mwsl", reveals
the web interfaces of Siemens S7 series PLC controllers. These interfaces
provide access to critical control and monitoring functions of industrial
systems. Unauthorized access can lead to significant operational
disruptions and security risks in industrial environments.

Proof Of Concept (PoC):
Steps to Reproduce:
1.Open Google Search.
2.Enter the dork query: intext:"siemens" & inurl:"/portal/portal.mwsl".
3.Review the search results to find URLs of Siemens S7 PLC web interfaces.
4. Click on a search result to access the web interface of the PLC.
5.Attempt to log in using default or commonly known credentials (if login
is required).

▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁
Forwarded from UNDERCODE TESTING
▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁

πŸš€ Support & Share: t.me/undercodecommunity

This is the hub for developers and tech enthusiasts:
πŸ’» Topics We Cover:

πŸ” CVE News & Databases
πŸ“° Hacker & Tech News
πŸ›‘ Cybersecurity, Hacking, and Secret Methods
🌟 Our Mission:
Share your knowledge, collaborate, and grow together in a community designed for innovation and learning.

πŸ”— Join now: Let's build the future together!

@UndercodeCommunity
▁ β–‚ β–„ Uπ•Ÿπ”»β’Ίπ«Δ†π”¬π““β“” β–„ β–‚ ▁