Sunday, 29 June 2025

AI - Status Update

 

 

Remember my Obsidian project?

I am working on an obsidian-deck inspired by the Dystopian Hackers using Cyberdecks which gives them in the Shadowrun RPG, the grand-father of Dystopian interactive worlds, the name Deggers.

This up there is first of all a big knot with many connections. It shows impressively that my interactions with DeepSeek are connected. The dots in my case don't show the source file names, but tags created by an AI NLP model running over them. The three core main tags are: #knowledge_management #Obsidian and #notes-taking by having most connections. 

That is a surprise considering how much Cyberpunk Stories about Shadow Runners inspired by the Shadowrun RPG Source Books I have created. These terms just don't have as many connections as the three main contenders. It is also a main reflection of my last job, in a Secret Service covered op Enterprise I couldn't possibly catch, and shows my deeply routed Jedi faith to train body and mind.

The Obsidian-deck is a terminal environment in a python 13 environment on a Linux Mint Computer. 

(cyberdeck-env) J4v@M920:~/Scripts$ python3 obsidian-deck.py --vault /home/J4v/Vault/
⚠️ Ensure Ollama is installed and models are downloaded:
  ollama pull mixtral:8x7b
  ollama pull llama3
⏳ Indexing vault...
🔗 Processed 346 notes with 0 errors
✅ Vault indexed: 346 notes
🧠 Available NLP models: mixtral:8x7b, llama3:latest
ℹ️ Type 'help' for available commands
ℹ️ Type 'list_tags' to see all available tags/concepts

🧠 AI-Powered Obsidian Knowledge Explorer
Type 'help' for commands. 'exit' to quit.
obsidian>

This is the manual to it:

Ollama with models: ollama pull mixtral:8x7b and ollama pull llama3


COMMANDS:
    CLUSTER [type] [name]
    Define a note cluster by metadata type
        all: Search across all metadata fields
        tag: Search only tags
        concept: Search only concepts
        meta_tag: Search only meta tags
        meta_keyword: Search only keywords
        Example: cluster all raspberry
    SHOW_CLUSTER
    Display current cluster with perfect formatting

        Automatically adjusts to note name length
        Shows word count and first 7 tags
        Multi-line formatting for long names/tags
    LIST_TAGS
    Show structured metadata index:
        Section I: By category (Tags, Concepts, etc.)
        Section II: Alphabetical index with counts
        Shows total unique metadata items
    SUMMARIZE [type]
    Generate AI summary of current cluster
        summary: Use note summaries only (faster)
        deep: Use first 500 words of content (more detailed)
        Example: summarize deep
    MODEL [name]
    Switch between NLP models
        mixtral:8x7b: Higher quality (slower)
        llama3:latest: Faster (good for most tasks)
        Example: model mixtral:8x7b
    BFS_CLUSTER [depth]
    Expand cluster via graph relationships
        Default depth: 2
        Finds notes linked to current cluster
        Example: bfs_cluster 3
    VISUALIZE
    Open current cluster in Obsidian
        Highlights notes in graph view
        Uses Obsidian's search URI scheme
    HELP
    Show available commands
        help [command]: Show detailed help for specific command
    EXIT
    Quit the explorer
 
It is work in progress, but allready returns all tags in an Index structured form 
 
   D. KEYWORDS (176 items)
    accent (1), adhesive (1), administration (1), aerodynamics (2)
    aeronautics (1), aircraft (2), algorithm (1), analysis (3), analytics (4)
    android (1), arbitrage (1), armor (1), augmented (1), authentication (1)
    automating (1),
 
 A (153)
    aaa (1), aarch64 (2), academia (1), academic (1), accent (1)
    access_control (2), accessibility (1), accuracy (2), acer_aspire_i7 (1)
    action-adventure (1), action_camera (2), adaptability (2), adaptation (2)
    adaptivetradingbot (2), adb (2), adhesive (1), adhesives (1)
    admin_privileges (2),
 ðŸ“Š TOTAL UNIQUE METADATA ITEMS: 2141
obsidian>

Shows the tag clusters:
 
 obsidian> cluster all obsidian
🧩 Cluster defined: 223 notes with all 'obsidian'
obsidian>

Prints those Cluster contents with a word count of the notes:
 
-                                                |          | -
Walmart-Historical-Trade-Analogy                 |     1178 | #history, #decentralization, #obsidian,
                                                 |          | #knowledge_management, #cosmopolitanism
-                                                |          | -
Walrus-Operator-Python-Compatibility-Fix         |    40751 | #obsidian, #syntax_error, #python_3,
                                                 |          | #knowledge-management, #walrus_operator,
                                                 |          | #beautifulsoup, #elementtree
-                                                |          | -
 
-                                                |          | -
Zte-Blade-A35e-Vr-On-Beowulf-Cluster             |    12207 | #raspberrypi, #cloudflare-warp, #apt-
                                                 |          | error, #distributed-computing, #proxy-
                                                 |          | setup, #raspberrypi-proxy, #binance-
                                                 |          | latency
-                                                |          | -
------------------------------------------------------------------------------------------------
Total notes: 223 | Total words: 1978858
obsidian> 
 
 
It will show within Obsidian tag connections highlighted and give a human language AI generated summary of either a clusters summaries or all notes full text.
 
#cyberpunkcoltoure #deggers
 
Can you imagine what you can do with it? CIA wonders too and to exploit me. The FBI Messiah List is a covered KZ Operation. I am sadly for them neither a Beautiful Mind nor of any Good Will. I am a trained hardcore killer known as a Crusader Knight form a long history of Jewmurderer hunters and killers. The Teutonic Order and us have a long violent history with no mercy what so ever. ticktack.. .ticktack... ticktack its FREEDOM over BLOODSHED! again
 
You may not have Slaves! We told you.

but you may have love, fun and joy. I hope this helps.
#noblessoblige #rosecrusaders #provos #terrorstgangs