!pip set up magika openai -q
import os, io, json, zipfile, textwrap, hashlib, tempfile, getpass
from pathlib import Path
from collections import Counter
from magika import Magika
from magika.sorts import MagikaResult, PredictionMode
from openai import OpenAI
print("🔑 Enter your OpenAI API key (enter is hidden):")
api_key = getpass.getpass("OpenAI API Key: ")
consumer = OpenAI(api_key=api_key)
strive:
consumer.fashions.listing()
print("✅ OpenAI linked successfullyn")
besides Exception as e:
elevate SystemExit(f"❌ OpenAI connection failed: {e}")
m = Magika()
print("✅ Magika loaded successfullyn")
print(f" module model : {m.get_module_version()}")
print(f" mannequin title : {m.get_model_name()}")
print(f" output sorts : {len(m.get_output_content_types())} supported labelsn")
def ask_gpt(system: str, person: str, mannequin: str = "gpt-4o", max_tokens: int = 600) -> str:
resp = consumer.chat.completions.create(
mannequin=mannequin,
max_tokens=max_tokens,
messages=[
{"role": "system", "content": system},
{"role": "user", "content": user},
],
)
return resp.selections[0].message.content material.strip()
print("=" * 60)
print("SECTION 1 — Core API + GPT Plain-Language Rationalization")
print("=" * 60)
samples = {
"Python": b'import osndef greet(title):n print(f"Whats up, {title}")n',
"JavaScript": b'const fetch = require("node-fetch");nasync operate getData() { return await fetch("/api"); }',
"CSV": b'title,age,citynAlice,30,NYCnBob,25,LAn',
"JSON": b'{"title": "Alice", "scores": [10, 20, 30], "lively": true}',
"Shell": b'#!/bin/bashnecho "Whats up"nfor i in $(seq 1 5); do echo $i; performed',
"PDF magic": b'%PDF-1.4n1 0 objn<< /Sort /Catalog >>nendobjn',
"ZIP magic": bytes([0x50, 0x4B, 0x03, 0x04]) + bytes(26),
}
print(f"n{'Label':<12} {'MIME Sort':<30} {'Rating':>6}")
print("-" * 52)
magika_labels = []
for title, uncooked in samples.gadgets():
res = m.identify_bytes(uncooked)
magika_labels.append(res.output.label)
print(f"{res.output.label:<12} {res.output.mime_type:<30} {res.rating:>5.1%}")
clarification = ask_gpt(
system="You're a concise ML engineer. Clarify in 4–5 sentences.",
person=(
f"Magika is Google's AI file-type detector. It simply recognized these sorts from uncooked bytes: "
f"{magika_labels}. Clarify how a deep-learning mannequin detects file sorts from "
"simply bytes, and why this beats counting on file extensions."
),
max_tokens=250,
)
print(f"n💬 GPT on how Magika works:n{textwrap.fill(clarification, 72)}n")
print("=" * 60)
print("SECTION 2 — Batch Identification + GPT Abstract")
print("=" * 60)
tmp_dir = Path(tempfile.mkdtemp())
file_specs = {
"code.py": b"import sysnprint(sys.model)n",
"fashion.css": b"physique { font-family: Arial; margin: 0; }n",
"knowledge.json": b'[{"id": 1, "val": "foo"}, {"id": 2, "val": "bar"}]',
"script.sh": b"#!/bin/shnecho Whats up Worldn",
"doc.html": b"Whats up
",
"config.yaml": b"server:n host: localhostn port: 8080n",
"question.sql": b"CREATE TABLE t (id INT PRIMARY KEY, title TEXT);n",
"notes.md": b"# Headingnn- merchandise onen- merchandise twon",
}
paths = []
for fname, content material in file_specs.gadgets():
p = tmp_dir / fname
p.write_bytes(content material)
paths.append(p)
outcomes = m.identify_paths(paths)
batch_summary = [
{"file": p.name, "label": r.output.label,
"group": r.output.group, "score": f"{r.score:.1%}"}
for p, r in zip(paths, results)
]
print(f"n{'File':<18} {'Label':<14} {'Group':<12} {'Rating':>6}")
print("-" * 54)
for row in batch_summary:
print(f"{row['file']:<18} {row['label']:<14} {row['group']:<12} {row['score']:>6}")
gpt_summary = ask_gpt(
system="You're a DevSecOps skilled. Be concise and sensible.",
person=(
f"A file add scanner detected these file sorts in a batch: "
f"{json.dumps(batch_summary)}. "
"In 3–4 sentences, summarise what sort of venture this seems like "
"and flag any file sorts which may warrant further scrutiny."
),
max_tokens=220,
)
print(f"n💬 GPT venture evaluation:n{textwrap.fill(gpt_summary, 72)}n")