Peerpedia Core

Zero-dependency protocol layer for PeerPedia — the peer-reviewed academic publishing platform built on Git.

peerpedia-core          ← this package: protocols, types, exceptions
├── peerpedia-storage   ← SQL + Git backends   (separate package)
├── peerpedia-transport ← HTTP / P2P transport  (separate package)
├── peerpedia-compute   ← scoring & reputation  (separate package)
├── peerpedia-social    ← notifications, shares (separate package)
└── peerpedia-app       ← CLI, server, REPL     (separate package)

What this package is

peerpedia-core defines the contracts — every other PeerPedia package depends on this one. It contains:

• Types — frozen dataclasses for Article, Review, User, Scores, Version, query types, and write payloads. All storage-agnostic, no IO.
• Protocols — typing.Protocol classes for storage backends, lifecycle state machines, sync, auth, authorization, compilation, and scoring.
• Exceptions — semantic exception hierarchy with machine-readable codes.
• Crypto protocols — algorithm-agnostic SigningKey / PublicKey interfaces.
• Lifecycle dispatcher — execute() reduces an action+context through a pluggable state machine.
• Peerpedia facade — dependency-injection wrapper that wires protocols into a usable engine.

The Scores type is a generic Mapping[str, float] wrapper — the compute plugin (peerpedia-compute) defines which dimensions exist and how they are aggregated. Aggregate scores are a read-side projection over review data, not an article attribute stored in core.

What this package is NOT

• Not a storage backend. See peerpedia-storage for SQLAlchemy and GitPython implementations of the storage protocols.
• Not a scoring algorithm. See peerpedia-compute for average, reputation-weighted, and sedimentation engines.
• Not a CLI, server, or REPL. See peerpedia-app.

Quick start

from peerpedia_core import Peerpedia
from peerpedia_core.types import Article, ArticleId, User, UserId

# Peerpedia needs concrete backends — swap these with real implementations
# from peerpedia-storage in production:
#
#   from peerpedia_storage import SqlArticleStorage, SqlUserStorage, SqlLifecycle
#
from tests.conftest import MemArticleStorage, MemLifecycle, MemUserStorage

storage = MemArticleStorage()
pp = Peerpedia(
    storage=storage,
    lifecycle=MemLifecycle(storage),
    user_storage=MemUserStorage(),
)

# Create an article
aid = pp.create()

# Revise it
article = Article(id=aid, title="My Paper", status="draft", authors=("Alice",))
pp.revise(aid, content="# Hello World", article=article)

# Read metadata back
meta = pp.read_meta(aid)
print(meta.title)  # "My Paper"

Note: The Mem* backends above live in tests/conftest.py and are only available in editable/dev installs. Production code imports from peerpedia-storage (SQLAlchemy + GitPython) or peerpedia-transport instead.

Protocols

This table shows every protocol in core and which package implements it.

Exceptions

All business-logic errors inherit from PeerpediaError:

PeerpediaError            code="ERROR"
  ├── NotFoundError       code="NOT_FOUND"
  ├── NotAuthorizedError  code="NOT_AUTHORIZED"
  ├── ConflictError       code="CONFLICT"
  │   └── MergeConflictError  code="MERGE_CONFLICT"
  └── BadRequestError     code="BAD_REQUEST"

Each carries structured context (permission, resource_type, conflicting_entity, etc.) for typed error handling.

Requirements

• Python ≥ 3.11
• Zero dependencies

License

AGPL-3.0 — see LICENSE.