Pulsar

Synapse tracks what you learn. Pulsar tracks what you create.

PULSAR is a CLI tool that extracts structured data from free-text descriptions of your projects and published content, no manual field-by-field input. Paste a sentence, get a structured record.

Part of a two-tool personal knowledge system:

• [Synapse] maps concepts and learning progress (internal connections)
• Pulsar tracks software built and content published (outward emissions)

---

How it works

data/*.json (skill config)  →  patterns compiled at runtime
You paste free text         →  regex extracts fields + skill categories
                            →  each record becomes a typed object (Software / Contenuto)
                            →  saved to data.json  →  export to CSV anytime

Input

ho creato un tool chiamato Pulsar usando python e json,
è completato, link https://github.com/sergio/pulsar

Extracted automatically

{
  "nome": "Pulsar",
  "tecnologie": ["python", "json"],
  "stato": "completato",
  "link": "https://github.com/sergio/pulsar",
  "tipo": "tool"
}

---

Architecture

Each entry is not a loose dictionary but a typed object. A small class hierarchy models the two kinds of record, sharing what they have in common and diverging where they differ:

Entry (base)        data, link                       · to_dict()
├── Software        + tipo, nome, tecnologie, stato   · to_dict() · from_dict() · __str__()
└── Contenuto       + titolo, argomento, piattaforma  · to_dict() · from_dict() · __str__()

• Inheritance: Entry holds the fields common to every record (data, link); Software and Contenuto add their own and reuse the parent through super().
• Object/dict bridge: to_dict() turns an object into a plain dict for JSON and CSV; from_dict() (a classmethod) rebuilds an object from a stored dict. Storage stays in plain JSON, the program works with real objects.
• Polymorphism: saving (save_portfolio) and reporting (read_report) call to_dict() and print(obj) without knowing the concrete type. Each object carries its own behavior, so the same line works for both kinds of entry.

---

Features

• Object-oriented design - typed entry classes with inheritance and polymorphic serialization, not raw dictionaries
• Data-driven skill extraction - skill categories loaded from external JSON files, patterns compiled at runtime, no hardcoded lists
• Auto-extraction via regex - name, technologies, date, status, link, platform
• Two entry types - software projects and published content (LinkedIn, X)
• Persistent JSON storage - entries accumulate across sessions
• CSV export - projects.csv (all entries) + skills.csv (technologies aggregated by count)
• Terminal report - readable summary of all saved entries
• Error handling - graceful messages for missing or corrupted files

---

Usage

python pulsar.py

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
          PULSAR
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
1. Aggiungi software
2. Aggiungi contenuto
3. Esporta CSV
4. Leggi report
5. Esci

Skill categories

Pulsar loads skill patterns from the data/ folder at startup. Each file defines one category:

{
  "label": "techs",
  "keywords": ["python", "sql", "bash", "git"]
}

File	Label	Purpose
techs.json	techs	languages and core tools
frameworks.json	frameworks	libraries and frameworks
cloud_providers.json	cloud	cloud platforms and services

To add a new technology, edit the relevant JSON file. To add a new category, drop a new .json file into data/. No code changes needed.

---

Input guide

Pulsar reads natural language. Include the right keywords and it extracts the rest.

Software entry

Field	Keyword pattern	Example
Name	chiamato X / si chiama X	chiamato Pulsar
Technologies	mention them directly	python, json, csv
Status	use the word directly	completato / in corso / wip
Link	any URL	https://github.com/...
Type	use the word directly	tool / progetto
Date	any standard format	05/2025 / maggio 2025

Note: a project name is read as a single token (no spaces), so chiamato Pulsar or chiamato call-outcome-tracker work; multi-word names with spaces are truncated to the first word.

Example input:

ho creato un tool chiamato Pulsar usando python e json, è completato,
link https://github.com/sergio/pulsar, maggio 2025

Content entry

Field	Keyword pattern	Example
Title	wrap in quotes "..."	"Il mio primo tool Python"
Topic	tech/skills mentioned, auto-detected	python, git
Platform	social X / su linkedin	su linkedin
Link	any URL	https://linkedin.com/...
Date	any standard format	05/2025

Example input:

ho pubblicato "Il mio primo tool Python" su linkedin,
link https://linkedin.com/post/123, maggio 2025

---

Output files

data.json

{
  "portfolio": {
    "software": [
      {
        "data": "maggio 2025",
        "tipo": "tool",
        "nome": "Pulsar",
        "tecnologie": ["python", "json"],
        "stato": "completato",
        "link": "https://github.com/sergio/pulsar"
      }
    ],
    "contenuti": [
      {
        "data": "maggio 2025",
        "titolo": "Il mio primo tool Python",
        "argomento": ["python"],
        "piattaforma": "linkedin",
        "link": "https://linkedin.com/post/123"
      }
    ]
  }
}

projects.csv

sezione	nome	tipo	piattaforma	tecnologie	argomento	stato	link	data
software	Pulsar	tool		python, json		completato	https://...	maggio 2025
contenuti	Il mio primo tool Python		linkedin		python		https://...	maggio 2025

skills.csv

tecnologia	occorrenze
python	3
json	2

---

Requirements

• Python 3.10+
• Standard library only (re, os, json, csv), no installs needed

---

Part of a larger roadmap

Pulsar was built as a practical learning project while completing CS50P by Harvard University and transitioning into AI development. It covers object-oriented design, regex, JSON, CSV, file I/O, CLI design, and error handling, all from the Python standard library.