-
Notifications
You must be signed in to change notification settings - Fork 0
H2B: make canonical status bytes the evidence boundary #47
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,331 @@ | ||
| from __future__ import annotations | ||
|
|
||
| import hashlib | ||
| import json | ||
| import re | ||
| from collections.abc import Iterable, Mapping | ||
| from dataclasses import dataclass | ||
| from enum import Enum | ||
|
|
||
|
|
||
| STATUS_VERSION = "pert.feed_status_canonical.v1" | ||
| MAX_SAFE_JSON_INTEGER = 2**53 - 1 | ||
| MAX_ROWS_PER_FEED = 10_000 | ||
| _ROW_KEYS = ("item_id", "published_at", "source_type", "source_url", "author", "text") | ||
| _OUTCOME_KEYS = frozenset({"feed_id", "feed_url", "kind", "state", "rows", "error_code"}) | ||
| _ERROR_RE = re.compile(r"^[a-z][a-z0-9_]{0,63}$") | ||
| _DIGEST_RE = re.compile(r"^[0-9a-f]{64}$") | ||
| _EMPTY_DIGEST = hashlib.sha256(b"[]").hexdigest() | ||
| _WIRE_KEYS = frozenset( | ||
| { | ||
| "status_version", | ||
| "configured_feed_count", | ||
| "feed_count", | ||
| "successful_feed_count", | ||
| "failed_feed_count", | ||
| "quarantined_feed_count", | ||
| "accepted_row_count", | ||
| "rejected_row_count", | ||
| "publication_complete", | ||
| "eligible_for_live_publication", | ||
| "aggregate_row_digest", | ||
| "feeds", | ||
| } | ||
| ) | ||
| _FEED_KEYS = frozenset( | ||
| { | ||
| "feed_id", | ||
| "feed_url", | ||
| "kind", | ||
| "state", | ||
| "accepted_row_count", | ||
| "rejected_row_count", | ||
| "row_digest", | ||
| "error_code", | ||
| } | ||
| ) | ||
|
|
||
|
|
||
| class DecisionContractError(ValueError): | ||
| """Sanitized canonical status contract error.""" | ||
|
|
||
| def __init__(self, code: str): | ||
| super().__init__(code) | ||
| self.code = code | ||
|
|
||
|
|
||
| class DecisionKind(str, Enum): | ||
| SUCCESS = "success" | ||
| QUARANTINE = "quarantine" | ||
| HARD_FAIL = "hard_fail" | ||
|
|
||
|
|
||
| @dataclass(frozen=True) | ||
| class ProducerDecision: | ||
| kind: DecisionKind | ||
|
|
||
|
|
||
| @dataclass(frozen=True) | ||
| class CanonicalDecision: | ||
| status_bytes: bytes | ||
| decision: ProducerDecision | ||
|
|
||
|
|
||
| def _fail(code: str) -> None: | ||
| raise DecisionContractError(code) | ||
|
|
||
|
|
||
| def _mapping(value: object, keys: frozenset[str], code: str) -> dict[str, object]: | ||
| if not isinstance(value, Mapping): | ||
| _fail(code) | ||
| try: | ||
| data = dict(value) | ||
| except (AttributeError, KeyError, OverflowError, RuntimeError, TypeError, UnicodeError, ValueError): | ||
| _fail(code) | ||
| if set(data) != keys or any(type(key) is not str for key in data): | ||
| _fail(code) | ||
| return data | ||
|
|
||
|
|
||
| def _string(value: object, code: str, *, allow_empty: bool = False) -> str: | ||
| if type(value) is not str or (not allow_empty and not value) or any(ord(char) < 0x20 for char in value): | ||
| _fail(code) | ||
| return value | ||
|
|
||
|
|
||
| def _integer(value: object, code: str) -> int: | ||
| if type(value) is not int or value < 0 or value > MAX_SAFE_JSON_INTEGER: | ||
| _fail(code) | ||
| return value | ||
|
|
||
|
|
||
| def _row(value: object) -> dict[str, str]: | ||
| data = _mapping(value, frozenset(_ROW_KEYS), "row_invalid") | ||
| return {key: _string(data[key], "row_invalid", allow_empty=key == "author") for key in _ROW_KEYS} | ||
|
|
||
|
|
||
| def _digest(rows: list[dict[str, str]]) -> str: | ||
| ordered = sorted(rows, key=lambda row: tuple(row[key] for key in _ROW_KEYS)) | ||
| try: | ||
| payload = json.dumps( | ||
| ordered, | ||
| ensure_ascii=False, | ||
| sort_keys=True, | ||
| separators=(",", ":"), | ||
| allow_nan=False, | ||
| ).encode("utf-8") | ||
| except (RecursionError, TypeError, UnicodeError, ValueError): | ||
| _fail("row_digest_invalid") | ||
| return hashlib.sha256(payload).hexdigest() | ||
|
|
||
|
|
||
| def _parse_outcome(value: object) -> tuple[dict[str, object], list[dict[str, str]]]: | ||
| data = _mapping(value, _OUTCOME_KEYS, "outcome_invalid") | ||
| feed_id = _string(data["feed_id"], "feed_invalid") | ||
| feed_url = _string(data["feed_url"], "feed_invalid") | ||
| kind = data["kind"] | ||
| state = data["state"] | ||
| if type(kind) is not str or kind not in {"rss2", "atom", "unknown"}: | ||
| _fail("feed_kind_invalid") | ||
| if type(state) is not str or state not in {"accepted", "failed", "quarantined"}: | ||
| _fail("feed_state_invalid") | ||
| rows = data["rows"] | ||
| if not isinstance(rows, (list, tuple)) or len(rows) > MAX_ROWS_PER_FEED: | ||
| _fail("rows_invalid") | ||
| row_snapshot = [_row(item) for item in rows] | ||
| error = data["error_code"] | ||
| if error is not None and (type(error) is not str or not _ERROR_RE.fullmatch(error)): | ||
| _fail("error_code_invalid") | ||
| if state in {"accepted", "quarantined"} and kind not in {"rss2", "atom"}: | ||
| _fail("feed_kind_invalid") | ||
| if state == "accepted" and (not row_snapshot or error is not None): | ||
| _fail("outcome_invariant_invalid") | ||
| if state == "quarantined" and (row_snapshot or error is None): | ||
| _fail("outcome_invariant_invalid") | ||
| if state == "failed" and (row_snapshot or error is None): | ||
| _fail("outcome_invariant_invalid") | ||
| return {"feed_id": feed_id, "feed_url": feed_url, "kind": kind, "state": state, "error_code": error}, row_snapshot | ||
|
|
||
|
|
||
| def _canonical(value: object) -> bytes: | ||
| try: | ||
| return json.dumps(value, ensure_ascii=False, sort_keys=True, separators=(",", ":"), allow_nan=False).encode( | ||
| "utf-8" | ||
| ) | ||
| except (RecursionError, TypeError, UnicodeError, ValueError): | ||
| _fail("status_serialization_invalid") | ||
|
|
||
|
|
||
| def _build_wire(parsed: list[tuple[dict[str, object], list[dict[str, str]]]]) -> dict[str, object]: | ||
| parsed.sort(key=lambda pair: (pair[0]["feed_id"], pair[0]["feed_url"])) | ||
| accepted_rows = sorted( | ||
| [row for data, rows in parsed if data["state"] == "accepted" for row in rows], | ||
| key=lambda row: tuple(row[key] for key in _ROW_KEYS), | ||
| ) | ||
| feeds = [] | ||
| for data, rows in parsed: | ||
| accepted = data["state"] == "accepted" | ||
| feeds.append( | ||
| { | ||
| "feed_id": data["feed_id"], | ||
| "feed_url": data["feed_url"], | ||
| "kind": data["kind"], | ||
| "state": data["state"], | ||
| "accepted_row_count": len(rows) if accepted else 0, | ||
| "rejected_row_count": 0, | ||
| "row_digest": _digest(rows) if accepted else hashlib.sha256(b"[]").hexdigest(), | ||
| "error_code": data["error_code"], | ||
| } | ||
| ) | ||
| failed = sum(data["state"] == "failed" for data, _ in parsed) | ||
| quarantined = sum(data["state"] == "quarantined" for data, _ in parsed) | ||
| complete = failed == 0 and quarantined == 0 | ||
| return { | ||
| "status_version": STATUS_VERSION, | ||
| "configured_feed_count": len(parsed), | ||
| "feed_count": len(parsed), | ||
| "successful_feed_count": len(parsed) - failed - quarantined, | ||
| "failed_feed_count": failed, | ||
| "quarantined_feed_count": quarantined, | ||
| "accepted_row_count": len(accepted_rows), | ||
| "rejected_row_count": 0, | ||
| "publication_complete": complete, | ||
| "eligible_for_live_publication": complete, | ||
| "aggregate_row_digest": _digest(accepted_rows), | ||
| "feeds": feeds, | ||
| } | ||
|
|
||
|
|
||
| def build_decision(validated_outcomes: Iterable[Mapping[str, object]]) -> CanonicalDecision: | ||
| try: | ||
| values = list(validated_outcomes) | ||
| except (AttributeError, RuntimeError, TypeError, ValueError): | ||
| _fail("outcomes_invalid") | ||
| if not values: | ||
| _fail("feed_config_empty") | ||
| parsed = [_parse_outcome(value) for value in values] | ||
| if len({data["feed_id"] for data, _ in parsed}) != len(parsed): | ||
| _fail("feed_duplicate") | ||
| status = _build_wire(parsed) | ||
| if status["failed_feed_count"]: | ||
| decision = DecisionKind.HARD_FAIL | ||
| elif status["quarantined_feed_count"]: | ||
| decision = DecisionKind.QUARANTINE | ||
| else: | ||
| decision = DecisionKind.SUCCESS | ||
| return CanonicalDecision(_canonical(status), ProducerDecision(decision)) | ||
|
|
||
|
|
||
| def _reject_duplicate_keys(pairs: list[tuple[str, object]]) -> dict[str, object]: | ||
| result: dict[str, object] = {} | ||
| for key, value in pairs: | ||
| if key in result: | ||
| _fail("status_duplicate_key") | ||
| result[key] = value | ||
| return result | ||
|
|
||
|
|
||
| def _validate_wire(value: object) -> dict[str, object]: | ||
| data = _mapping(value, _WIRE_KEYS, "status_invalid") | ||
| if data["status_version"] != STATUS_VERSION: | ||
| _fail("status_version_invalid") | ||
| counter_keys = _WIRE_KEYS - { | ||
| "status_version", | ||
| "publication_complete", | ||
| "eligible_for_live_publication", | ||
| "aggregate_row_digest", | ||
| "feeds", | ||
| } | ||
| for key in counter_keys: | ||
| _integer(data[key], "status_counter_invalid") | ||
| for key in ("publication_complete", "eligible_for_live_publication"): | ||
| if type(data[key]) is not bool: | ||
| _fail("status_flag_invalid") | ||
| aggregate = _string(data["aggregate_row_digest"], "status_digest_invalid") | ||
| if not _DIGEST_RE.fullmatch(aggregate): | ||
| _fail("status_digest_invalid") | ||
| feeds = data["feeds"] | ||
| if not isinstance(feeds, list) or not feeds: | ||
| _fail("feed_invalid") | ||
| previous: tuple[str, str] | None = None | ||
| feed_ids: set[str] = set() | ||
| for value in feeds: | ||
| item = _mapping(value, _FEED_KEYS, "feed_invalid") | ||
| feed_id = _string(item["feed_id"], "feed_invalid") | ||
| feed_url = _string(item["feed_url"], "feed_invalid") | ||
| key = (feed_id, feed_url) | ||
| if feed_id in feed_ids: | ||
| _fail("feed_duplicate") | ||
| if previous is not None and key <= previous: | ||
| _fail("feed_order_invalid") | ||
| feed_ids.add(feed_id) | ||
| previous = key | ||
| kind = item["kind"] | ||
| state = item["state"] | ||
| if type(kind) is not str or kind not in {"rss2", "atom", "unknown"}: | ||
| _fail("feed_kind_invalid") | ||
| if type(state) is not str or state not in {"accepted", "failed", "quarantined"}: | ||
| _fail("feed_state_invalid") | ||
| if state in {"accepted", "quarantined"} and kind not in {"rss2", "atom"}: | ||
| _fail("feed_kind_invalid") | ||
| accepted_count = _integer(item["accepted_row_count"], "feed_counter_invalid") | ||
| rejected_count = _integer(item["rejected_row_count"], "feed_counter_invalid") | ||
| if accepted_count > MAX_ROWS_PER_FEED: | ||
| _fail("feed_counter_invalid") | ||
| if rejected_count != 0: | ||
| _fail("rejected_count_invalid") | ||
| if state == "accepted" and accepted_count == 0: | ||
| _fail("feed_state_invalid") | ||
| if state != "accepted" and accepted_count != 0: | ||
| _fail("feed_state_invalid") | ||
| digest = _string(item["row_digest"], "status_digest_invalid") | ||
| if not _DIGEST_RE.fullmatch(digest): | ||
| _fail("status_digest_invalid") | ||
|
Comment on lines
+281
to
+283
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
For failed or quarantined feeds there are no rows by invariant, and Useful? React with 👍 / 👎. |
||
| if state != "accepted" and digest != _EMPTY_DIGEST: | ||
| _fail("empty_digest_invalid") | ||
|
Comment on lines
+284
to
+285
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
When a feed is Useful? React with 👍 / 👎. |
||
| error = item["error_code"] | ||
| if state == "accepted" and error is not None: | ||
| _fail("feed_state_invalid") | ||
| if state != "accepted" and (type(error) is not str or not _ERROR_RE.fullmatch(error)): | ||
| _fail("feed_error_invalid") | ||
| if data["configured_feed_count"] != len(feeds) or data["feed_count"] != len(feeds): | ||
| _fail("status_counter_invalid") | ||
|
Comment on lines
+291
to
+292
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
When persisted bytes are read from the evidence boundary, a canonical status with Useful? React with 👍 / 👎. |
||
| accepted = sum(item["state"] == "accepted" for item in feeds) | ||
| failed = sum(item["state"] == "failed" for item in feeds) | ||
| quarantined = sum(item["state"] == "quarantined" for item in feeds) | ||
| accepted_rows = sum(item["accepted_row_count"] for item in feeds) | ||
| rejected_rows = sum(item["rejected_row_count"] for item in feeds) | ||
| if data["successful_feed_count"] != accepted or data["failed_feed_count"] != failed: | ||
| _fail("status_counter_invalid") | ||
| if ( | ||
| data["quarantined_feed_count"] != quarantined | ||
| or data["accepted_row_count"] != accepted_rows | ||
| or rejected_rows != 0 | ||
| or data["rejected_row_count"] != 0 | ||
| ): | ||
| _fail("status_counter_invalid") | ||
| if accepted_rows == 0 and data["aggregate_row_digest"] != _EMPTY_DIGEST: | ||
| _fail("empty_digest_invalid") | ||
|
Comment on lines
+307
to
+308
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
When Useful? React with 👍 / 👎. |
||
| complete = failed == 0 and quarantined == 0 | ||
| if data["publication_complete"] != complete or data["eligible_for_live_publication"] != complete: | ||
| _fail("status_flag_invalid") | ||
| return data | ||
|
|
||
|
|
||
| def read_status(status_bytes: bytes) -> dict[str, object]: | ||
| if type(status_bytes) is not bytes: | ||
| _fail("status_bytes_invalid") | ||
| try: | ||
| value = json.loads(status_bytes.decode("utf-8"), object_pairs_hook=_reject_duplicate_keys) | ||
| except (UnicodeError, json.JSONDecodeError, RecursionError): | ||
| _fail("status_bytes_invalid") | ||
| data = _validate_wire(value) | ||
| if _canonical(data) != status_bytes: | ||
| _fail("status_noncanonical") | ||
| return json.loads(status_bytes.decode("utf-8")) | ||
|
|
||
|
|
||
| def status_digest(status_bytes: bytes) -> str: | ||
| if type(status_bytes) is not bytes: | ||
| _fail("status_bytes_invalid") | ||
| return hashlib.sha256(status_bytes).hexdigest() | ||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
When
accepted_row_countis0(all feeds are failed/quarantined), the aggregate row set is known to be empty andbuild_decision()always emits_digest([]), but readback only checks thataggregate_row_digestis hex. A canonical status for an all-failed or all-quarantined run can therefore change this digest to any 64 hex characters and still passread_status(), which breaks consumers using the aggregate digest as the row-set evidence.Useful? React with 👍 / 👎.