fix(validation): enforce strict CDR types and storage ranges

This commit is contained in:
2026-09-19 21:30:15 +03:30
parent 2f4c99620e
commit 0869282949
2 changed files with 93 additions and 87 deletions

49
app/common/validation.py Normal file
View File

@@ -0,0 +1,49 @@
"""Shared wire schema. Semantic storage validation belongs to the worker."""
import json
import math
STRING_FIELDS = ("raw_id", "device_id", "caller_num", "svc", "lac", "ci")
INTEGER_FIELDS = ("rat", "epoch_start", "epoch_end", "up_bytes", "down_bytes")
REQUIRED_FIELDS = STRING_FIELDS + INTEGER_FIELDS
class PayloadError(ValueError):
def __init__(self, reason, detail, field=None):
super().__init__(detail)
self.reason = reason
self.field = field
def parse_payload(data):
def reject_constant(value):
raise ValueError(f"Non-finite number: {value}")
def finite_float(value):
number = float(value)
if not math.isfinite(number):
raise ValueError(f"Non-finite number: {value}")
return number
try:
payload = json.loads(data.decode("utf-8"), parse_constant=reject_constant, parse_float=finite_float)
except (UnicodeError, ValueError, RecursionError) as exc:
raise PayloadError("invalid_json", str(exc)) from exc
if not isinstance(payload, dict):
raise PayloadError("invalid_payload", "Expected a JSON object")
return payload
def validate_structure(payload):
for field in REQUIRED_FIELDS:
if payload.get(field) is None:
raise PayloadError("missing_required_field", f"Field {field} is required", field)
for field in STRING_FIELDS:
if not isinstance(payload[field], str) or not payload[field].strip():
raise PayloadError("invalid_payload", f"Field {field} must be a nonempty string", field)
try:
payload[field].encode("utf-8")
except UnicodeError as exc:
raise PayloadError("invalid_payload", f"Field {field} contains invalid Unicode", field) from exc
for field in INTEGER_FIELDS:
if type(payload[field]) is not int:
raise PayloadError("invalid_payload", f"Field {field} must be an integer", field)
if payload.get("src_ip") is not None and not isinstance(payload["src_ip"], str):
raise PayloadError("invalid_payload", "Field src_ip must be a string or null", "src_ip")

View File

@@ -1,4 +1,6 @@
import json import json
from ipaddress import IPv4Address
from app.common.validation import PayloadError, parse_payload, validate_structure
import uuid import uuid
from dataclasses import dataclass from dataclasses import dataclass
from datetime import datetime, timezone from datetime import datetime, timezone
@@ -58,9 +60,9 @@ class NormalizedRecord:
self.bytes_downloaded, self.bytes_downloaded,
self.total_bytes, self.total_bytes,
self.duration_seconds, self.duration_seconds,
datetime.strptime(self.session_start_time, "%Y-%m-%d %H:%M:%S"), datetime.strptime(self.session_start_time, "%Y-%m-%d %H:%M:%S").replace(tzinfo=timezone.utc),
datetime.strptime(self.session_end_time, "%Y-%m-%d %H:%M:%S"), datetime.strptime(self.session_end_time, "%Y-%m-%d %H:%M:%S").replace(tzinfo=timezone.utc),
self.source_ip, IPv4Address(self.source_ip),
) )
@@ -105,96 +107,51 @@ def normalize_cdr_payload(
(record, failure_reason, failed_field, raw_id) (record, failure_reason, failed_field, raw_id)
If record is None, failure_reason describes why it should be routed to DLQ. If record is None, failure_reason describes why it should be routed to DLQ.
""" """
# 1. Parse JSON
try: try:
if isinstance(raw_payload_bytes, bytes): payload = parse_payload(raw_payload_bytes)
payload_str = raw_payload_bytes.decode("utf-8") except PayloadError:
else:
payload_str = str(raw_payload_bytes)
payload: Dict[str, Any] = json.loads(payload_str)
except Exception:
return None, "malformed_json", None, None return None, "malformed_json", None, None
if not isinstance(payload, dict):
return None, "malformed_json", None, None
raw_id = payload.get("raw_id") raw_id = payload.get("raw_id")
raw_id_str = str(raw_id) if raw_id is not None else None raw_id_str = raw_id if isinstance(raw_id, str) else None
# 2. Check required fields
for field in REQUIRED_FIELDS:
if field not in payload or payload[field] is None:
return None, "missing_required_field", field, raw_id_str
# 3. Check timestamps (epoch_start <= epoch_end)
try: try:
epoch_start = int(payload["epoch_start"]) validate_structure(payload)
epoch_end = int(payload["epoch_end"]) except PayloadError as exc:
if epoch_start > epoch_end: reason = "missing_required_field" if exc.reason == "missing_required_field" else (
return None, "invalid_epoch", "epoch_start", raw_id_str "invalid_epoch" if exc.field in ("epoch_start", "epoch_end") else
except (ValueError, TypeError): "invalid_rat_value" if exc.field == "rat" else "normalization_error")
return None, reason, exc.field, raw_id_str
for field in ("epoch_start", "epoch_end"):
if not 0 <= payload[field] <= 2**32 - 1:
return None, "invalid_epoch", field, raw_id_str
epoch_start, epoch_end = payload["epoch_start"], payload["epoch_end"]
if epoch_start > epoch_end:
return None, "invalid_epoch", "epoch_start", raw_id_str return None, "invalid_epoch", "epoch_start", raw_id_str
for field in ("up_bytes", "down_bytes"):
# 4. Check bytes (must be >= 0) if not 0 <= payload[field] <= 2**64 - 1:
try: return None, "normalization_error", field, raw_id_str
up_bytes = int(payload["up_bytes"]) up_bytes, down_bytes = payload["up_bytes"], payload["down_bytes"]
down_bytes = int(payload["down_bytes"]) if up_bytes + down_bytes > 2**64 - 1:
if up_bytes < 0: return None, "normalization_error", "total_bytes", raw_id_str
return None, "normalization_error", "up_bytes", raw_id_str if payload["rat"] not in RAT_MAP:
if down_bytes < 0:
return None, "normalization_error", "down_bytes", raw_id_str
except (ValueError, TypeError):
return None, "normalization_error", "bytes", raw_id_str
# 5. Radio Access Type mapping
rat_raw = payload.get("rat")
try:
rat_int = int(rat_raw)
if rat_int not in RAT_MAP:
return None, "invalid_rat_value", "rat", raw_id_str
radio_access_type = RAT_MAP[rat_int]
except (ValueError, TypeError):
return None, "invalid_rat_value", "rat", raw_id_str return None, "invalid_rat_value", "rat", raw_id_str
msisdn = normalize_phone_number(payload["caller_num"])
# 6. E.164 phone normalization if msisdn is None:
caller_num = str(payload["caller_num"])
msisdn = normalize_phone_number(caller_num)
if not msisdn:
return None, "normalization_error", "caller_num", raw_id_str return None, "normalization_error", "caller_num", raw_id_str
try:
# 7. Field mappings and derivations source_ip = str(IPv4Address(payload["src_ip"] if payload.get("src_ip") is not None else "0.0.0.0"))
imsi = str(payload["device_id"]) except (ValueError, TypeError):
service_type = str(payload["svc"]) return None, "normalization_error", "src_ip", raw_id_str
cell_id = f"{payload['lac']}_{payload['ci']}" try:
total_bytes = up_bytes + down_bytes start_dt = datetime.fromtimestamp(epoch_start, timezone.utc).strftime("%Y-%m-%d %H:%M:%S")
duration_seconds = epoch_end - epoch_start end_dt = datetime.fromtimestamp(epoch_end, timezone.utc).strftime("%Y-%m-%d %H:%M:%S")
except (ValueError, OverflowError, OSError):
start_dt = datetime.fromtimestamp(epoch_start, tz=timezone.utc).strftime( return None, "invalid_epoch", "epoch_start", raw_id_str
"%Y-%m-%d %H:%M:%S"
)
end_dt = datetime.fromtimestamp(epoch_end, tz=timezone.utc).strftime(
"%Y-%m-%d %H:%M:%S"
)
src_ip = str(payload.get("src_ip") or "0.0.0.0")
# Deterministic UUIDv5 cdr_id from raw_id
cdr_id = str(uuid.uuid5(uuid.NAMESPACE_DNS, raw_id_str))
record = NormalizedRecord( record = NormalizedRecord(
cdr_id=cdr_id, cdr_id=str(uuid.uuid5(uuid.NAMESPACE_DNS, raw_id)), raw_id=raw_id,
raw_id=raw_id_str, msisdn=msisdn, imsi=payload["device_id"], service_type=payload["svc"],
msisdn=msisdn, radio_access_type=RAT_MAP[payload["rat"]], cell_id=f"{payload['lac']}_{payload['ci']}",
imsi=imsi, bytes_uploaded=up_bytes, bytes_downloaded=down_bytes, total_bytes=up_bytes+down_bytes,
service_type=service_type, duration_seconds=epoch_end-epoch_start, session_start_time=start_dt,
radio_access_type=radio_access_type, session_end_time=end_dt, source_ip=source_ip,
cell_id=cell_id,
bytes_uploaded=up_bytes,
bytes_downloaded=down_bytes,
total_bytes=total_bytes,
duration_seconds=duration_seconds,
session_start_time=start_dt,
session_end_time=end_dt,
source_ip=src_ip,
) )
return record, None, None, raw_id_str return record, None, None, raw_id