ethereum.arrow_glacier.trie
State Trie ^^^^^^^^^^
.. contents:: Table of Contents :backlinks: none :local:
Introduction
The state trie is the structure responsible for storing
.fork_types.Account
objects.
EMPTY_TRIE_ROOT
57 | EMPTY_TRIE_ROOT = Root( |
---|---|
58 | hex_to_bytes( |
59 | "56e81f171bcc55a6ff8345e692c0f86e5b48e01b996cadc001622fb5e363b421" |
60 | ) |
61 | ) |
Node
63 | Node = Union[Account, Bytes, LegacyTransaction, Receipt, Uint, U256, None] |
---|
K
64 | K = TypeVar("K", bound=Bytes) |
---|
V
65 | V = TypeVar( |
---|---|
66 | "V", |
67 | Optional[Account], |
68 | Optional[Bytes], |
69 | Bytes, |
70 | Optional[Union[LegacyTransaction, Bytes]], |
71 | Optional[Union[Receipt, Bytes]], |
72 | Uint, |
73 | U256, |
74 | ) |
LeafNode
Leaf node in the Merkle Trie
77 | @slotted_freezable |
---|
78 | @dataclass |
---|
class LeafNode:
rest_of_key
82 | rest_of_key: Bytes |
---|
value
83 | value: rlp.Extended |
---|
ExtensionNode
Extension node in the Merkle Trie
86 | @slotted_freezable |
---|
87 | @dataclass |
---|
class ExtensionNode:
key_segment
91 | key_segment: Bytes |
---|
subnode
92 | subnode: rlp.Extended |
---|
BranchNode
Branch node in the Merkle Trie
95 | @slotted_freezable |
---|
96 | @dataclass |
---|
class BranchNode:
subnodes
100 | subnodes: List[rlp.Extended] |
---|
value
101 | value: rlp.Extended |
---|
InternalNode
104 | InternalNode = Union[LeafNode, ExtensionNode, BranchNode] |
---|
encode_internal_node
Encodes a Merkle Trie node into its RLP form. The RLP will then be
serialized into a Bytes
and hashed unless it is less that 32 bytes
when serialized.
This function also accepts None
, representing the absence of a node,
which is encoded to b""
.
Parameters
node : Optional[InternalNode] The node to encode.
Returns
encoded : rlp.RLP
The node encoded as RLP.
def encode_internal_node(node: Optional[InternalNode]) -> ethereum.rlp.Extended:
108 | """ |
---|---|
109 | Encodes a Merkle Trie node into its RLP form. The RLP will then be |
110 | serialized into a `Bytes` and hashed unless it is less that 32 bytes |
111 | when serialized. |
112 |
|
113 | This function also accepts `None`, representing the absence of a node, |
114 | which is encoded to `b""`. |
115 |
|
116 | Parameters |
117 | ---------- |
118 | node : Optional[InternalNode] |
119 | The node to encode. |
120 |
|
121 | Returns |
122 | ------- |
123 | encoded : `rlp.RLP` |
124 | The node encoded as RLP. |
125 | """ |
126 | unencoded: rlp.Extended |
127 | if node is None: |
128 | unencoded = b"" |
129 | elif isinstance(node, LeafNode): |
130 | unencoded = ( |
131 | nibble_list_to_compact(node.rest_of_key, True), |
132 | node.value, |
133 | ) |
134 | elif isinstance(node, ExtensionNode): |
135 | unencoded = ( |
136 | nibble_list_to_compact(node.key_segment, False), |
137 | node.subnode, |
138 | ) |
139 | elif isinstance(node, BranchNode): |
140 | unencoded = node.subnodes + [node.value] |
141 | else: |
142 | raise AssertionError(f"Invalid internal node type {type(node)}!") |
143 | |
144 | encoded = rlp.encode(unencoded) |
145 | if len(encoded) < 32: |
146 | return unencoded |
147 | else: |
148 | return keccak256(encoded) |
encode_node
Encode a Node for storage in the Merkle Trie.
Currently mostly an unimplemented stub.
def encode_node(node: Node, storage_root: Optional[Bytes]) -> Bytes:
152 | """ |
---|---|
153 | Encode a Node for storage in the Merkle Trie. |
154 |
|
155 | Currently mostly an unimplemented stub. |
156 | """ |
157 | if isinstance(node, Account): |
158 | assert storage_root is not None |
159 | return encode_account(node, storage_root) |
160 | elif isinstance(node, (LegacyTransaction, Receipt, U256)): |
161 | return rlp.encode(node) |
162 | elif isinstance(node, Bytes): |
163 | return node |
164 | else: |
165 | return previous_trie.encode_node(node, storage_root) |
Trie
The Merkle Trie.
168 | @dataclass |
---|
class Trie:
secured
174 | secured: bool |
---|
default
175 | default: V |
---|
_data
176 | _data: Dict[K, V] = field(default_factory=dict) |
---|
copy_trie
Create a copy of trie
. Since only frozen objects may be stored in tries,
the contents are reused.
Parameters
trie: Trie
Trie to copy.
Returns
new_trie : Trie[K, V]
A copy of the trie.
def copy_trie(trie: Trie[K, V]) -> Trie[K, V]:
180 | """ |
---|---|
181 | Create a copy of `trie`. Since only frozen objects may be stored in tries, |
182 | the contents are reused. |
183 |
|
184 | Parameters |
185 | ---------- |
186 | trie: `Trie` |
187 | Trie to copy. |
188 |
|
189 | Returns |
190 | ------- |
191 | new_trie : `Trie[K, V]` |
192 | A copy of the trie. |
193 | """ |
194 | return Trie(trie.secured, trie.default, copy.copy(trie._data)) |
trie_set
Stores an item in a Merkle Trie.
This method deletes the key if value == trie.default
, because the Merkle
Trie represents the default value by omitting it from the trie.
Parameters
trie: Trie
Trie to store in.
key : Bytes
Key to lookup.
value : V
Node to insert at key
.
def trie_set(trie: Trie[K, V], key: K, value: V) -> None:
198 | """ |
---|---|
199 | Stores an item in a Merkle Trie. |
200 |
|
201 | This method deletes the key if `value == trie.default`, because the Merkle |
202 | Trie represents the default value by omitting it from the trie. |
203 |
|
204 | Parameters |
205 | ---------- |
206 | trie: `Trie` |
207 | Trie to store in. |
208 | key : `Bytes` |
209 | Key to lookup. |
210 | value : `V` |
211 | Node to insert at `key`. |
212 | """ |
213 | if value == trie.default: |
214 | if key in trie._data: |
215 | del trie._data[key] |
216 | else: |
217 | trie._data[key] = value |
trie_get
Gets an item from the Merkle Trie.
This method returns trie.default
if the key is missing.
Parameters
trie: Trie to lookup in. key : Key to lookup.
Returns
node : V
Node at key
in the trie.
def trie_get(trie: Trie[K, V], key: K) -> V:
221 | """ |
---|---|
222 | Gets an item from the Merkle Trie. |
223 |
|
224 | This method returns `trie.default` if the key is missing. |
225 |
|
226 | Parameters |
227 | ---------- |
228 | trie: |
229 | Trie to lookup in. |
230 | key : |
231 | Key to lookup. |
232 |
|
233 | Returns |
234 | ------- |
235 | node : `V` |
236 | Node at `key` in the trie. |
237 | """ |
238 | return trie._data.get(key, trie.default) |
common_prefix_length
Find the longest common prefix of two sequences.
def common_prefix_length(a: Sequence, b: Sequence) -> int:
242 | """ |
---|---|
243 | Find the longest common prefix of two sequences. |
244 | """ |
245 | for i in range(len(a)): |
246 | if i >= len(b) or a[i] != b[i]: |
247 | return i |
248 | return len(a) |
nibble_list_to_compact
Compresses nibble-list into a standard byte array with a flag.
A nibble-list is a list of byte values no greater than 15
. The flag is
encoded in high nibble of the highest byte. The flag nibble can be broken
down into two two-bit flags.
Highest nibble::
+---+---+----------+--------+
| _ | _ | is_leaf | parity |
+---+---+----------+--------+
3 2 1 0
The lowest bit of the nibble encodes the parity of the length of the
remaining nibbles -- 0
when even and 1
when odd. The second lowest bit
is used to distinguish leaf and extension nodes. The other two bits are not
used.
Parameters
x : Array of nibbles. is_leaf : True if this is part of a leaf node, or false if it is an extension node.
Returns
compressed : bytearray
Compact byte array.
def nibble_list_to_compact(x: Bytes, is_leaf: bool) -> Bytes:
252 | """ |
---|---|
253 | Compresses nibble-list into a standard byte array with a flag. |
254 |
|
255 | A nibble-list is a list of byte values no greater than `15`. The flag is |
256 | encoded in high nibble of the highest byte. The flag nibble can be broken |
257 | down into two two-bit flags. |
258 |
|
259 | Highest nibble:: |
260 |
|
261 | +---+---+----------+--------+ |
262 | | _ | _ | is_leaf | parity | |
263 | +---+---+----------+--------+ |
264 | 3 2 1 0 |
265 |
|
266 |
|
267 | The lowest bit of the nibble encodes the parity of the length of the |
268 | remaining nibbles -- `0` when even and `1` when odd. The second lowest bit |
269 | is used to distinguish leaf and extension nodes. The other two bits are not |
270 | used. |
271 |
|
272 | Parameters |
273 | ---------- |
274 | x : |
275 | Array of nibbles. |
276 | is_leaf : |
277 | True if this is part of a leaf node, or false if it is an extension |
278 | node. |
279 |
|
280 | Returns |
281 | ------- |
282 | compressed : `bytearray` |
283 | Compact byte array. |
284 | """ |
285 | compact = bytearray() |
286 | |
287 | if len(x) % 2 == 0: # ie even length |
288 | compact.append(16 * (2 * is_leaf)) |
289 | for i in range(0, len(x), 2): |
290 | compact.append(16 * x[i] + x[i + 1]) |
291 | else: |
292 | compact.append(16 * ((2 * is_leaf) + 1) + x[0]) |
293 | for i in range(1, len(x), 2): |
294 | compact.append(16 * x[i] + x[i + 1]) |
295 | |
296 | return Bytes(compact) |
bytes_to_nibble_list
Converts a Bytes
into to a sequence of nibbles (bytes with value < 16).
Parameters
bytes_:
The Bytes
to convert.
Returns
nibble_list : Bytes
The Bytes
in nibble-list format.
def bytes_to_nibble_list(bytes_: Bytes) -> Bytes:
300 | """ |
---|---|
301 | Converts a `Bytes` into to a sequence of nibbles (bytes with value < 16). |
302 |
|
303 | Parameters |
304 | ---------- |
305 | bytes_: |
306 | The `Bytes` to convert. |
307 |
|
308 | Returns |
309 | ------- |
310 | nibble_list : `Bytes` |
311 | The `Bytes` in nibble-list format. |
312 | """ |
313 | nibble_list = bytearray(2 * len(bytes_)) |
314 | for byte_index, byte in enumerate(bytes_): |
315 | nibble_list[byte_index * 2] = (byte & 0xF0) >> 4 |
316 | nibble_list[byte_index * 2 + 1] = byte & 0x0F |
317 | return Bytes(nibble_list) |
_prepare_trie
Prepares the trie for root calculation. Removes values that are empty,
hashes the keys (if secured == True
) and encodes all the nodes.
Parameters
trie :
The Trie
to prepare.
get_storage_root :
Function to get the storage root of an account. Needed to encode
Account
objects.
Returns
out : Mapping[ethereum.base_types.Bytes, Node]
Object with keys mapped to nibble-byte form.
def _prepare_trie(trie: Trie[K, V], get_storage_root: Optional[Callable[[Address], Root]]) -> Mapping[Bytes, Bytes]:
324 | """ |
---|---|
325 | Prepares the trie for root calculation. Removes values that are empty, |
326 | hashes the keys (if `secured == True`) and encodes all the nodes. |
327 |
|
328 | Parameters |
329 | ---------- |
330 | trie : |
331 | The `Trie` to prepare. |
332 | get_storage_root : |
333 | Function to get the storage root of an account. Needed to encode |
334 | `Account` objects. |
335 |
|
336 | Returns |
337 | ------- |
338 | out : `Mapping[ethereum.base_types.Bytes, Node]` |
339 | Object with keys mapped to nibble-byte form. |
340 | """ |
341 | mapped: MutableMapping[Bytes, Bytes] = {} |
342 | |
343 | for preimage, value in trie._data.items(): |
344 | if isinstance(value, Account): |
345 | assert get_storage_root is not None |
346 | address = Address(preimage) |
347 | encoded_value = encode_node(value, get_storage_root(address)) |
348 | else: |
349 | encoded_value = encode_node(value) |
350 | if encoded_value == b"": |
351 | raise AssertionError |
352 | key: Bytes |
353 | if trie.secured: |
354 | # "secure" tries hash keys once before construction |
355 | key = keccak256(preimage) |
356 | else: |
357 | key = preimage |
358 | mapped[bytes_to_nibble_list(key)] = encoded_value |
359 | |
360 | return mapped |
root
Computes the root of a modified merkle patricia trie (MPT).
Parameters
trie :
Trie
to get the root of.
get_storage_root :
Function to get the storage root of an account. Needed to encode
Account
objects.
Returns
root : .fork_types.Root
MPT root of the underlying key-value pairs.
def root(trie: Trie[K, V], get_storage_root: Optional[Callable[[Address], Root]]) -> Root:
367 | """ |
---|---|
368 | Computes the root of a modified merkle patricia trie (MPT). |
369 |
|
370 | Parameters |
371 | ---------- |
372 | trie : |
373 | `Trie` to get the root of. |
374 | get_storage_root : |
375 | Function to get the storage root of an account. Needed to encode |
376 | `Account` objects. |
377 |
|
378 |
|
379 | Returns |
380 | ------- |
381 | root : `.fork_types.Root` |
382 | MPT root of the underlying key-value pairs. |
383 | """ |
384 | obj = _prepare_trie(trie, get_storage_root) |
385 | |
386 | root_node = encode_internal_node(patricialize(obj, Uint(0))) |
387 | if len(rlp.encode(root_node)) < 32: |
388 | return keccak256(rlp.encode(root_node)) |
389 | else: |
390 | assert isinstance(root_node, Bytes) |
391 | return Root(root_node) |
patricialize
Structural composition function.
Used to recursively patricialize and merkleize a dictionary. Includes memoization of the tree structure and hashes.
Parameters
obj : Underlying trie key-value pairs, with keys in nibble-list format. level : Current trie level.
Returns
node : ethereum.base_types.Bytes
Root node of obj
.
def patricialize(obj: Mapping[Bytes, Bytes], level: Uint) -> Optional[InternalNode]:
397 | """ |
---|---|
398 | Structural composition function. |
399 |
|
400 | Used to recursively patricialize and merkleize a dictionary. Includes |
401 | memoization of the tree structure and hashes. |
402 |
|
403 | Parameters |
404 | ---------- |
405 | obj : |
406 | Underlying trie key-value pairs, with keys in nibble-list format. |
407 | level : |
408 | Current trie level. |
409 |
|
410 | Returns |
411 | ------- |
412 | node : `ethereum.base_types.Bytes` |
413 | Root node of `obj`. |
414 | """ |
415 | if len(obj) == 0: |
416 | return None |
417 | |
418 | arbitrary_key = next(iter(obj)) |
419 | |
420 | # if leaf node |
421 | if len(obj) == 1: |
422 | leaf = LeafNode(arbitrary_key[level:], obj[arbitrary_key]) |
423 | return leaf |
424 | |
425 | # prepare for extension node check by finding max j such that all keys in |
426 | # obj have the same key[i:j] |
427 | substring = arbitrary_key[level:] |
428 | prefix_length = len(substring) |
429 | for key in obj: |
430 | prefix_length = min( |
431 | prefix_length, common_prefix_length(substring, key[level:]) |
432 | ) |
433 |
|
434 | # finished searching, found another key at the current level |
435 | if prefix_length == 0: |
436 | break |
437 | |
438 | # if extension node |
439 | if prefix_length > 0: |
440 | prefix = arbitrary_key[int(level) : int(level) + prefix_length] |
441 | return ExtensionNode( |
442 | prefix, |
443 | encode_internal_node( |
444 | patricialize(obj, level + Uint(prefix_length)) |
445 | ), |
446 | ) |
447 | |
448 | branches: List[MutableMapping[Bytes, Bytes]] = [] |
449 | for _ in range(16): |
450 | branches.append({}) |
451 | value = b"" |
452 | for key in obj: |
453 | if len(key) == level: |
454 | # shouldn't ever have an account or receipt in an internal node |
455 | if isinstance(obj[key], (Account, Receipt, Uint)): |
456 | raise AssertionError |
457 | value = obj[key] |
458 | else: |
459 | branches[key[level]][key] = obj[key] |
460 | |
461 | return BranchNode( |
462 | [ |
463 | encode_internal_node(patricialize(branches[k], level + Uint(1))) |
464 | for k in range(16) |
465 | ], |
466 | value, |
467 | ) |