pickle and base64 — Object Serialization and Binary-to-Text Conversion

We'll cover two modules used when moving data through files or across the network. pickle saves Python objects (dicts, class instances, etc.) in a form that can be fully restored later, while base64 converts binary data like images or audio into a string of "alphanumerics plus two symbols" so you can drop it straight into email, JSON, or URLs. Both are standard library modules for data conversion at internal boundaries or external transports.

Sorting out what each module is for

The two modules naturally split by input/output combination. Use the diagram below to grab the big picture, then we'll dig into typical uses and APIs in each section.

pickle — Convert Python Objects Straight to Bytes

pickle is a standard library module that converts Python objects into a form that can be fully restored later. The result is a bytes value (a contiguous sequence of integers from 0 to 255 — a type for handling "sequences of numbers" rather than "sequences of characters" like strings). Bytes are the basic format for situations where machines, not humans, will read the data — saving to files, sending over the network, handing data to another process.

Unlike json, pickle's strength is being able to save not just dicts / lists / strings / numbers but also custom class instances and function objects, which makes it useful for saving ML models or passing objects between Python processes.

Never loads pickle data from someone else

pickle.loads can execute arbitrary Python code, so passing in bytes from an untrusted source becomes an arbitrary-code-execution vulnerability. Limit yourself to "I made it, I read it back" scenarios and never use it on data received over the network or from someone else. For external communication, use json instead.

Convert a score dictionary to bytes with pickle and check the return type. Pickle's bytes can vary in length across environments, so we'll verify it via type.

① Import the pickle module.

② Set up the dictionary {"name": "Alice", "scores": [80, 90, 75]} in a variable named data.

③ Convert data to bytes with pickle, store it in pickled, and print the type name as Type: ◯◯ (it should be bytes).

(If you run this correctly, an explanation will appear.)

Restore the pickled you made in Practice 1 back into a Python object and verify it matches the original data exactly. The variables data and pickled from Practice 1 are still available, so you only need to write the restoration step.

④ Restore pickled with pickle.loads into a variable named restored.

⑤ Print Restored name: ◯◯ (the name field of the restored object) and Equal to original: True / False for the equality check.

base64 — Move Binary Data Through Text

base64 is a module that converts arbitrary binary data into a string of "alphanumerics plus two symbols". Since 3 bytes become 4 characters, the size grows about 1.33×, but it's used widely in situations where you need to carry binary data through text-only channels, such as email attachments, embedding images in JSON or URLs, or storing bytes in SQL.

The two basic APIs:

import base64

# bytes → Base64
binary = b"Python is fun!"
encoded = base64.b64encode(binary)
print(encoded)              # b'UHl0aG9uIGlzIGZ1biE='
print(encoded.decode())     # UHl0aG9uIGlzIGZ1biE= (converted to str)

# Base64 → bytes
decoded = base64.b64decode(encoded)
print(decoded)              # b'Python is fun!'
print(decoded == binary)    # True

For URLs, use urlsafe_b64

Standard b64encode includes + and /, so embedding it directly in URLs or JSON keys requires escaping. For URLs and JSON, choose base64.urlsafe_b64encode instead — it swaps + → - and / → _ to give you a URL-safe variant. Use base64.urlsafe_b64decode to restore.

Encode a bytes string with Base64, decode it again, and verify the result matches the original.

① Import the base64 module.

② Set up the bytes string b"Python is fun!".

③ Don't print the Base64-encoded bytes as bytes — convert it to an ASCII string first and print it as Base64 string: ◯◯.

④ Decode the Base64 back to the original binary and print whether it matches the original as Equal to original: True / False.