source: flexoentity/README.md@ ca39274

#+TITLE: flexoentity
#+SUBTITLE: A hardened entity base and deterministic identifier system for the Flex-O family
#+AUTHOR: Enno
#+DATE: 2025-10-20
#+OPTIONS: toc:3 num:nil

* Overview

`flexoentity` provides the *identity and lifecycle backbone* for all Flex-O components  
(Flex-O-Grader, Flex-O-Vault, Flex-O-Drill, …).

It defines how entities such as questions, media, catalogs, and exams are *identified, versioned, signed, and verified* — all without any external database dependencies.

At its heart lie two modules:

- =id_factory.py= – robust, cryptographically-verifiable *Flex-O ID generator*
- =versioned_entity.py= – abstract *base class for all versioned entities*

Together, they form a compact yet powerful foundation for audit-ready, reproducible data structures across offline and air-gapped deployments.

✳️ Design Goals
|----------------|--------------------------------------------------------------------------------------------------------|
| Goal           | Description                                                                                            |
|----------------|--------------------------------------------------------------------------------------------------------|
| *Determinism*  | IDs are derived from canonicalized entity content — identical input always yields identical ID prefix. |
| *Integrity*    | BLAKE2s hashing and digital signatures protect against manual tampering.                               |
| *Traceability* | Version numbers (=@001A=, =@002S=, …) track entity lifecycle transitions.                              |
| *Stability*    | Hash prefixes remain constant across state changes; only version and state suffixes evolve.            |
| *Auditability* | Every entity can be serialized, verified, and reconstructed without hidden dependencies.               |
| *Simplicity*   | Pure-Python, zero external libraries, self-contained and easy to embed.                                |
|----------------|--------------------------------------------------------------------------------------------------------|

* Flex-O ID Structure

Each entity carries a unique *Flex-O ID*, generated by =FlexOID.generate()=.

#+BEGIN_EXAMPLE
AF-Q250101-9A4C2D@003S
#+END_EXAMPLE

|-----------|----------|---------------------------------------------|
| Segment   | Example  | Meaning                                     |
|-----------|----------|---------------------------------------------|
| *Domain*  | =AF=     | Logical scope (e.g. "Air Force")            |
| *Type*    | =Q=      | Entity type (e.g. Question)                 |
| *Date*    | =250101= | UTC creation date (YYMMDD)                  |
| *Hash*    | =9A4C2D= | 6-digit BLAKE2s digest of canonical content |
| *Version* | =003=    | Sequential version counter                  |
| *State*   | =S=      | Lifecycle state (=D=, =A=, =S=, =P=, =O=)   |
|-----------|----------|---------------------------------------------|

Hash collisions within a single session are automatically disambiguated (=-1=, =-2=, …).

* Lifecycle States

|-----------------------|---------|-----------------------------|
| State                 | Abbrev. | Description                 |
|-----------------------|---------|-----------------------------|
| *DRAFT*               | =D=     | Editable, not yet validated |
| *APPROVED*            | =A=     | Reviewed and accepted       |
| *APPROVED_AND_SIGNED* | =S=     | Cryptographically signed    |
| *PUBLISHED*           | =P=     | Released to consumers       |
| *OBSOLETE*            | =O=     | Archived or replaced        |
|-----------------------|---------|-----------------------------|

Transitions follow a strict progression:
#+BEGIN_EXAMPLE
DRAFT → APPROVED → APPROVED_AND_SIGNED → PUBLISHED → OBSOLETE
#+END_EXAMPLE

Version increments occur automatically for all *stable* transitions (APPROVED, SIGNED, PUBLISHED).

* Core Classes

** =FlexOID=

A lightweight immutable class representing the full identity of an entity.

*Highlights*
- =generate(domain, etype, estate, text, version=1)= → create a new ID
- =next_version(oid)= → increment version safely
- =clone_new_base(domain, etype, estate, text)= → start a new lineage
- Deterministic prefix, state-dependent signature

** =FlexoEntity=
Abstract base class for all versioned entities (e.g., Question, Exam, Catalog).

Implements:
- ID lifecycle management (=approve()=, =sign()=, =publish()=, =obsolete()=)
- JSON serialization (=to_json()=, =from_json()=)
- Integrity verification (=verify_integrity(entity)=)
- Controlled state transitions with automatic timestamps

Subclasses define a single property:

#+BEGIN_SRC python
@property
def text_seed(self) -> str:
    """Canonical text or core content for hashing."""
#+END_SRC

* Integrity Verification

Each entity can self-verify its integrity:

#+BEGIN_SRC python
entity = Question("AF", EntityType.QUESTION, "What is Ohm’s law?")
json_str = entity.to_json()
reloaded = Question.from_json(json_str)

assert FlexoEntity.verify_integrity(reloaded)
#+END_SRC

If the file is tampered with (e.g. "Ohm’s" → "Omm’s"), verification fails:

#+BEGIN_SRC python
assert not FlexoEntity.verify_integrity(reloaded)
#+END_SRC

* Example
#+BEGIN_SRC python
from flexoentity.versioned_entity import FlexoEntity, EntityType, EntityState

class Question(FlexoEntity):
    def __init__(self, domain, text):
        self._text = text
        super().__init__(domain, EntityType.QUESTION)

    @property
    def text_seed(self):
        return self._text

* Usage
#+BEGIN_SRC python
q = Question("AF", "What is Ohm’s law?")
print(q.flexo_id)             # AF-Q250101-9A4C2D@001D
q.approve()
print(q.flexo_id)             # AF-Q250101-9A4C2D@002A
q.sign()
print(q.flexo_id)             # AF-Q250101-9A4C2D@003S
#+END_SRC

* JSON Example
#+BEGIN_SRC json
{
  "domain": "AF",
  "etype": "QUESTION",
  "text_seed": "What is Ohm’s law?",
  "state": "APPROVED_AND_SIGNED",
  "version": 3,
  "flexo_id": "AF-Q250101-9A4C2D@003S",
  "signature": "1E3A9F2A8B7C4D11",
}
#+END_SRC

* Entity Type and State Codes

|-------------|------|------------------------------|
| EntityType  | Code | Typical Use                  |
|-------------|------|------------------------------|
| QUESTION    | Q    | Exam question                |
| CATALOG     | CAT  | Question or media collection |
| EXAM        | EX   | Composed test instance       |
| DATABASE    | DB   | Persistent store             |
| CERTIFICATE | CERT | Digital or approval document |
|-------------|------|------------------------------|

|---------------------|------|-------------------|
| EntityState         | Code | Meaning           |
|---------------------|------|-------------------|
| DRAFT               | D    | Work in progress  |
| APPROVED            | A    | Reviewed          |
| APPROVED_AND_SIGNED | S    | Signed version    |
| PUBLISHED           | P    | Publicly released |
| OBSOLETE            | O    | Deprecated        |
|---------------------|------|-------------------|

* Design Notes
- *Hash Stability:* Only domain, entity type, and content text influence the hash.
  This ensures consistent prefixes across state changes.
- *State-Dependent Signatures:* Each lifecycle stage has its own signature seed.
  Modifying a file without re-signing invalidates integrity.
- *Obsolescence Threshold:* Version numbers above 900 trigger warnings;
  beyond 999 are considered obsolete.
- *Clone Lineages:* Cloning an entity resets versioning but preserves metadata lineage.

* Dependencies
- Python 3.11+
- Standard library only (=hashlib=, =json=, =datetime=, =enum=, =dataclasses=)

No external packages. Fully compatible with *Guix*, *air-gapped* deployments, and *reproducible builds*.

* Integration
`flexoentity` is imported by higher-level modules such as:

- *Flex-O-Grader* → manages question catalogs and exam bundles
- *Flex-O-Vault* → provides persistent media storage with metadata integrity
- *Flex-O-Drill* → uses versioned entities for training simulations

All share the same identity and versioning logic — ensuring that
*what was approved, signed, and published remains provably authentic.*

* License
MIT License 2025
Part of the *Flex-O family* by Flex-O-Dyne GmbH 
Designed for reproducible, audit-ready, human-centered software.