Merge pull request #147 from PR0M3TH3AN/codex/implement-snapshot-processing-and-retrieval

Implement Nostr snapshot backup

src/nostr/__init__.py

@@ -22,10 +22,13 @@ __all__ = [
     "KIND_DELTA",
     "Manifest",
     "ChunkMeta",
+    "prepare_snapshot",
 ]
 
 
 def __getattr__(name: str):
     if name == "NostrClient":
         return import_module(".client", __name__).NostrClient
+    if name == "prepare_snapshot":
+        return import_module(".client", __name__).prepare_snapshot
     raise AttributeError(f"module '{__name__}' has no attribute '{name}'")

src/nostr/client.py

@@ -4,9 +4,10 @@ import base64
 import json
 import logging
 import time
-from typing import List, Optional
+from typing import List, Optional, Tuple
 import hashlib
 import asyncio
+import gzip
 
 # Imports from the nostr-sdk library
 from nostr_sdk import (
@@ -22,6 +23,7 @@ from nostr_sdk import (
 from datetime import timedelta
 
 from .key_manager import KeyManager as SeedPassKeyManager
+from .backup_models import Manifest, ChunkMeta, KIND_MANIFEST, KIND_SNAPSHOT_CHUNK
 from password_manager.encryption import EncryptionManager
 from utils.file_lock import exclusive_lock
 
@@ -39,6 +41,44 @@ DEFAULT_RELAYS = [
 ]
 
 
+def prepare_snapshot(
+    encrypted_bytes: bytes, limit: int
+) -> Tuple[Manifest, list[bytes]]:
+    """Compress and split the encrypted vault into chunks.
+
+    Each chunk is hashed with SHA-256 and described in the returned
+    :class:`Manifest`.
+
+    Parameters
+    ----------
+    encrypted_bytes : bytes
+        The encrypted vault contents.
+    limit : int
+        Maximum chunk size in bytes.
+
+    Returns
+    -------
+    Tuple[Manifest, list[bytes]]
+        The manifest describing all chunks and the list of chunk bytes.
+    """
+
+    compressed = gzip.compress(encrypted_bytes)
+    chunks = [compressed[i : i + limit] for i in range(0, len(compressed), limit)]
+
+    metas: list[ChunkMeta] = []
+    for i, chunk in enumerate(chunks):
+        metas.append(
+            ChunkMeta(
+                id=f"seedpass-chunk-{i:04d}",
+                size=len(chunk),
+                hash=hashlib.sha256(chunk).hexdigest(),
+            )
+        )
+
+    manifest = Manifest(ver=1, algo="gzip", chunks=metas)
+    return manifest, chunks
+
+
 class NostrClient:
     """Interact with the Nostr network using nostr-sdk."""
 
@@ -195,6 +235,82 @@ class NostrClient:
         self.last_error = "Latest event contained no content"
         return None
 
+    async def publish_snapshot(
+        self, encrypted_bytes: bytes, limit: int = 50_000
+    ) -> Manifest:
+        """Publish a compressed snapshot split into chunks.
+
+        Parameters
+        ----------
+        encrypted_bytes : bytes
+            Vault contents already encrypted with the user's key.
+        limit : int, optional
+            Maximum chunk size in bytes. Defaults to 50 kB.
+        """
+
+        manifest, chunks = prepare_snapshot(encrypted_bytes, limit)
+        for meta, chunk in zip(manifest.chunks, chunks):
+            content = base64.b64encode(chunk).decode("utf-8")
+            builder = EventBuilder(Kind(KIND_SNAPSHOT_CHUNK), content).tags(
+                [Tag.identifier(meta.id)]
+            )
+            event = builder.build(self.keys.public_key()).sign_with_keys(self.keys)
+            await self.client.send_event(event)
+
+        manifest_json = json.dumps(
+            {
+                "ver": manifest.ver,
+                "algo": manifest.algo,
+                "chunks": [meta.__dict__ for meta in manifest.chunks],
+                "delta_since": manifest.delta_since,
+            }
+        )
+
+        manifest_event = (
+            EventBuilder(Kind(KIND_MANIFEST), manifest_json)
+            .build(self.keys.public_key())
+            .sign_with_keys(self.keys)
+        )
+        await self.client.send_event(manifest_event)
+        return manifest
+
+    async def fetch_latest_snapshot(self) -> Tuple[Manifest, list[bytes]] | None:
+        """Retrieve the latest manifest and all snapshot chunks."""
+
+        pubkey = self.keys.public_key()
+        f = Filter().author(pubkey).kind(Kind(KIND_MANIFEST)).limit(1)
+        timeout = timedelta(seconds=10)
+        events = (await self.client.fetch_events(f, timeout)).to_vec()
+        if not events:
+            return None
+        manifest_raw = events[0].content()
+        data = json.loads(manifest_raw)
+        manifest = Manifest(
+            ver=data["ver"],
+            algo=data["algo"],
+            chunks=[ChunkMeta(**c) for c in data["chunks"]],
+            delta_since=data.get("delta_since"),
+        )
+
+        chunks: list[bytes] = []
+        for meta in manifest.chunks:
+            cf = (
+                Filter()
+                .author(pubkey)
+                .kind(Kind(KIND_SNAPSHOT_CHUNK))
+                .identifier(meta.id)
+                .limit(1)
+            )
+            cev = (await self.client.fetch_events(cf, timeout)).to_vec()
+            if not cev:
+                raise ValueError(f"Missing chunk {meta.id}")
+            chunk_bytes = base64.b64decode(cev[0].content().encode("utf-8"))
+            if hashlib.sha256(chunk_bytes).hexdigest() != meta.hash:
+                raise ValueError(f"Checksum mismatch for chunk {meta.id}")
+            chunks.append(chunk_bytes)
+
+        return manifest, chunks
+
     def close_client_pool(self) -> None:
         """Disconnects the client from all relays."""
         try:

src/tests/test_nostr_snapshot.py

@@ -0,0 +1,98 @@
+import hashlib
+import json
+import gzip
+from pathlib import Path
+from tempfile import TemporaryDirectory
+from cryptography.fernet import Fernet
+import base64
+import asyncio
+from unittest.mock import patch
+
+from nostr import prepare_snapshot, NostrClient
+from password_manager.encryption import EncryptionManager
+
+
+def test_prepare_snapshot_roundtrip():
+    data = b"a" * 70000
+    manifest, chunks = prepare_snapshot(data, 50000)
+    assert len(chunks) == len(manifest.chunks)
+    joined = b"".join(chunks)
+    assert len(joined) <= len(data)
+    assert hashlib.sha256(chunks[0]).hexdigest() == manifest.chunks[0].hash
+    assert manifest.chunks[0].id == "seedpass-chunk-0000"
+    assert data == gzip.decompress(joined)
+
+
+class DummyEvent:
+    def __init__(self, content):
+        self._content = content
+
+    def content(self):
+        return self._content
+
+
+class DummyClient:
+    def __init__(self, events):
+        self.events = events
+        self.pos = 0
+
+    async def add_relays(self, relays):
+        pass
+
+    async def add_relay(self, relay):
+        pass
+
+    async def connect(self):
+        pass
+
+    async def disconnect(self):
+        pass
+
+    async def send_event(self, event):
+        pass
+
+    async def fetch_events(self, f, timeout):
+        ev = self.events[self.pos]
+        self.pos += 1
+
+        class E:
+            def __init__(self, ev):
+                self._ev = ev
+
+            def to_vec(self):
+                return [self._ev]
+
+        return E(ev)
+
+
+def test_fetch_latest_snapshot():
+    data = b"seedpass" * 1000
+    manifest, chunks = prepare_snapshot(data, 50000)
+    manifest_json = json.dumps(
+        {
+            "ver": manifest.ver,
+            "algo": manifest.algo,
+            "chunks": [c.__dict__ for c in manifest.chunks],
+            "delta_since": None,
+        }
+    )
+    events = [DummyEvent(manifest_json)] + [
+        DummyEvent(base64.b64encode(c).decode()) for c in chunks
+    ]
+
+    client = DummyClient(events)
+    with TemporaryDirectory() as tmpdir:
+        enc_mgr = EncryptionManager(Fernet.generate_key(), Path(tmpdir))
+        with patch("nostr.client.Client", lambda signer: client), patch(
+            "nostr.client.KeyManager"
+        ) as MockKM, patch.object(NostrClient, "initialize_client_pool"), patch.object(
+            enc_mgr, "decrypt_parent_seed", return_value="seed"
+        ):
+            km = MockKM.return_value
+            km.keys.private_key_hex.return_value = "1" * 64
+            km.keys.public_key_hex.return_value = "2" * 64
+            nc = NostrClient(enc_mgr, "fp")
+            result_manifest, result_chunks = asyncio.run(nc.fetch_latest_snapshot())
+
+    assert manifest == result_manifest
+    assert result_chunks == chunks