logseq/pages/AI Playground.md

Preparation of the development environment

- ```
  mkdir todoist-insight && cd todoist-insight
  python3 -m venv venv
  source venv/bin/activate  # oppure .\venv\Scripts\activate su Windows
  touch requirements.txt README.md
  mkdir notebooks data src reports
  mkdir data/raw data/processed
  ```
	- The activation of the virtual environment has been done with different attempts. At the end, to make it working in VSCode on Windows, I opened a Powershell terminal and executed the corresponding script
		- `.\venv\Scripts\Activate.ps1`
	- This permitted to install `ipykernel` and make it visibale in vscode. The installation was done automatically via VSCode even though it would have worked also with
		- `python -m ipykernel install --user --name=todoist-env --display-name "Python (todoist-env)`
	- Once crated the virtual environment, you have to install the required dependencies with
		- `pip install -r requirements.txt`
- `uvicorn simple_service:app --reload`

• Structure of the project

  • todoist-insight/ ├── notebooks/ │ └── 01_analysis_dashboard.ipynb ← Notebook principale │ ├── data/ │ ├── raw/ │ │ ├── completed_tasks.json ← Export JSON completati │ │ └── incomplete_tasks.json ← Export JSON incompleti │ └── processed/ │ └── completed_tasks.csv ← Dataset già trasformato │ ├── reports/ │ ├── ai_analysis.md ← Output analisi GPT su task completati │ └── ai_suggestions.md ← Suggerimenti per task aperti │ ├── src/ │ ├── todoist_api.py ← Funzioni per accedere a Todoist │ ├── visualization.py ← Grafici (matplotlib/seaborn/plotly) │ ├── gpt_analysis.py ← Analisi qualitativa via OpenAI │ └── utils.py ← Funzioni ausiliarie (date, parsing, mapping) │ ├── .env ← File per chiavi API (non pushare su GitHub) ├── requirements.txt ← Dipendenze (todoist-api, openai, pandas, ...) └── README.md ← Panoramica e guida all’uso

• Improvement of the Palyground project with CrewAI usage

  • {{renderer :mermaid_684d7ac6-4d14-464a-8f71-53a7f313e92e, 3}} collapsed:: true

    • flowchart LR
          A[Load Task Cache] --> B[AnalyzerAgent]
          B --> C[ClassifierAgent]
          C --> D[AdvisorAgent]
          D --> E[PlannerAgent]
      
      

• #ReadingNotes of
  • @Generative AI with LangChain: build production-ready LLM applications and advanced agents using Python, LangChain, and LangGraph
  • From Zero to Your First AI Agent in 25 Minutes (No Coding)
    •