restructure project from golem-llm to golem-ai

[rutikthakre]

This PR rearranges all existing packages, CI and Makefile.toml for golem-ai.

• moved all previous llm's packages to golem-llm
• updated root cargo.toml
• updated github actions
• updated Makefile.toml
• tested commands

[Nanashi-lab]

Since you are doing this -

Cargo toml in ollama

base64 = "0.21"
mime_guess = "2.0"
url = "2.4" 

Some of them use older versions of the crate, crate base64 is in 22.1(workspace already has this one), others as well, and move it out to workspace, maybe.

---

          echo "Pulling Qwen2.5:1.5b"
          docker exec ollama ollama pull qwen2.5:1.5b
          echo "Pulling Gemma2:2b"
          docker exec ollama ollama pull gemma2:2b
          echo "Verifying models are available"
          docker exec ollama ollama list | grep -q "qwen2.5:1.5b" || exit 1
          docker exec ollama ollama list | grep -q "gemma2:2b" || exit 1

Ollama test pulls the wrong model, so the actual ci test never runs, since wasm files expect different model. , Output from a successful CI

  Created new worker ollama-1
  
  Component name: test:llm
  Worker name:    ollama-1
Selected profile: local
Using generated idempotency key: ec639f37-1868-4d37-8144-c3a688993b45
Fuzzy matched pattern test:llm as test:llm
Fuzzy matched pattern test1 as test:llm-exports/test-llm-api.{test1}
Invoking worker test:llm/ollama-1/test:llm-exports/test-llm-api.{test1} 

Invocation results in WAVE format:
  - "ERROR: ErrorCode::InternalError Failed to parse error response body: missing field `status_code` at line 1 column 64 ({\"error\":\"model \\\"qwen3:1.7b\\\" not found, try pulling it first\"})"
Selected profile: local
Using generated idempotency key: 17556872-c3c6-4b0c-9c9b-ad1cdd50a32b
Fuzzy matched pattern test:llm as test:llm
Fuzzy matched pattern test2 as test:llm-exports/test-llm-api.{test2}
Invoking worker test:llm/ollama-1/test:llm-exports/test-llm-api.{test2} 

Invocation results in WAVE format:
  - "No tool request"
Selected profile: local
Using generated idempotency key: 4c0eab37-b127-41a0-ba7f-3afba32eaa4c
Fuzzy matched pattern test:llm as test:llm
Fuzzy matched pattern test3 as test:llm-exports/test-llm-api.{test3}
Invoking worker test:llm/ollama-1/test:llm-exports/test-llm-api.{test3} 

If you pull the right models, the test should run properly and should show the output displayed in CI
Also the current CI fails, because they point to wrong folder, since everything has been moved to subfolder

Also a README.md

[rutikthakre]

Thanks, @Nanashi-lab. You are right. I don't remember when I changed the modal name in CI; maybe it happened while rebasing.

[rutikthakre]

The error is related to paths, and I think this will need more feedback from the project authors. So I will fix the error ASAP and wait for the project authors to review and give me feedback, till then, marking this as a draft

[rutikthakre]

Hey I will reopen this PR later.