Tuesday, 1 July 2025

AI - Obsidian Deck

I guess I will upload the script to Git-Hub and Source Forge at some point, but at this point it is not functional enough.

To analyze large datasets of natural language you need to do this:

Create a folder of the text files in .txt or .md
Install on the latest version of Linux Mint or Ubuntu (sudo apt update and sudo apt upgrade) in no dual boot set up.
Install python 3.13 with an python env.
# Debian/Ubuntu
sudo apt update
sudo apt install build-essential zlib1g-dev libncurses5-dev libgdbm-dev libnss3-dev libssl-dev libreadline-dev libffi-dev libsqlite3-dev libbz2-dev liblzma-dev

wget https://www.python.org/ftp/python/3.13.5/Python-3.13.5.tar.xz
tar -xf Python-3.13.5.tar.xz
cd Python-3.13.5
 
./configure --enable-optimizations --prefix=/usr/local 

make -j $(nproc)       # Use all CPU cores for faster compilation
sudo make altinstall   # "altinstall" prevents overwriting default `python3` binary
 
 python3.13 -m venv myenv  # Create venv with 3.13.5
source myenv/bin/activate
 
Install the Ollama system that will host the AI models which will parse the text files top add tags based on its NLP and Reasoning capabilities. Those are RAM intensive. At the bottom will be my Dream Player Set Up.

Within the active python environment:
curl https://ollama.ai/install.sh | sh. 

Then type to trigger ollama to pull an AI model:
ollama run mixtral:8x7b "Hello"
 
Or type:
ollama run llama3 "Hello" 

Depending on which model you want to download.
 
i7 32GB32GBMixtral:8x7b, Llama3:70b, DeepSeek-Coder:33bComplex reasoning
i7 16GB16GBDeepSeek-R1:7b, Qwen:14b, Phi-3:14bSQL generation
i5 11GB11GBLlama3:8b, Mistral:7b, CodeLlama:13bCommand generation
RPi5 8GB + AI Kit8GBDeepSeek-R1:1.5b, Gemma2:2b, TinyLlamaReal-time Hailo-accel
RPi44-8GBPhi-2, TinyLlamaSensor I/O monitoring
 
Obsidian can be installed using the Linux OS Software install app coming on all flavors of Linux.
 
Then, using DeepSeek or an Coding AI you need to create a set of scripts, test it, adjust it and finalize it that:
1. Parses all text or md files to give it tags and a summary saving them as copies into the Obsidian Vault
2. Parse all tags to create a set of meta_tag tags and writes that into the obsidian vault files
3.Create the Obsidian-deck tailored to your needs using a coding AI like DeepSeek, ideally in form of a shell that takes commands like: cluster all "my_tag_here" or summary cluster my_tag_here to create connections between the txt or md files and summaries of their content such as emails or meeting documents in human language like DeepSeek would do asking Summarize me If from Kipling, please.
Here is a concise summary of Rudyard Kipling's poem "If—":

### Core Message  
The poem presents a father's advice to his son on cultivating resilience, integrity, and balance to achieve true maturity. It outlines a blueprint for moral character, emphasizing that mastering these virtues allows one to "be a Man" and inherit "the Earth" .

### Key Qualities Advocated  
1. **Composure Under Pressure**:  
   - Maintain calm when others panic or blame you .  
2. **Self-Belief and Humility**:  
   - Trust yourself amid doubt, but respect others' perspectives .  
3. **Resilience Against Adversity**:  
   - Rebuild after failures ("stoop and build ’em up with worn-out tools") and treat triumph/disaster as "impostors" .  
4. **Integrity in Action**:  
   - Avoid lies and hatred even when confronted with both; value perseverance over dreams alone .  
5. **Balanced Relationships**:  
   - Engage with all people ("crowds" or "Kings") without losing authenticity or virtue .  
6. **Purposeful Use of Time**:  
   - "Fill the unforgiving minute" with meaningful effort .  

### Structure and Tone  
- Written as a **single sentence** structured around conditional "if" clauses, culminating in a triumphant "then" (lines 31–32) .  
- Uses **iambic pentameter** and a steady rhyme scheme (ABABCDCD), reinforcing its measured, advisory tone .  

### Historical Context  
- Inspired by **Leander Starr Jameson**'s stoicism after a failed 1895 raid, embodying Victorian-era ideals like the "stiff upper lip" .  
- First published in 1910 in *Rewards and Fairies* as paternal guidance for Kipling's son, John .  

### Legacy  
- Voted the **UK's favorite poem** in a 1995 BBC poll .  
- Lines like "meet with Triumph and Disaster" adorn **Wimbledon's** player entrance .  


Kipling’s timeless emphasis on inner strength over external validation makes "If—" an enduring guide to personal excellence. For full analysis, see detailed breakdowns of [themes](https://www.litcharts.com/poetry/rudyard-kipling/if) or [historical context](https://en.wikipedia.org/wiki/If%E2%80%94).

The ideal set up:
I have not manged to get a LongChain working due to using different models and low end hardware and there is no ollama model integration into a Beowulf Cluster which combines CPU power, but no RAM. Ollama AI is RAM intensive.
Ideally would be a Server Rack Ryzon CPU with 512GB RAM, a Nividia GPU in a PCI and a Hailo AI PCI card.  
Using a LongChain script several of such would receive tasks giving a combined output. To achieve this on all nodes Ollama must be installed and serve.
 
#cyberpunkcoltoure #deggers #provos #MIB the bureau not the brigade.