Spaces:
Sleeping
Sleeping
File size: 4,560 Bytes
32d42b3 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 | """
data_validator.py
-----------------
Validation and normalization utilities for extracted business data.
Used in the orchestrator pipeline between extraction and skill routing:
extract_fields() → validate_data() → route_to_skill()
The validator normalises:
- text fields (customer, item, reason) → stripped, lowercased or title-cased
- payment_type aliases (gpay → upi, paytm → upi, etc.)
- numeric fields (amount, quantity) → int or float, never negative
"""
from __future__ import annotations
import re
from typing import Any
_NUMBER_PATTERN = re.compile(r"-?\d+(?:\.\d+)?")
class DataValidator:
"""Validate and normalize extraction-agent output."""
def validate(self, intent: str, data: dict[str, Any]) -> dict[str, Any]:
"""Return a cleaned copy of extracted data for the given intent."""
cleaned = dict(data or {})
if "customer" in cleaned:
cleaned["customer"] = self._clean_text(cleaned.get("customer"), title=True)
if "item" in cleaned:
cleaned["item"] = self._clean_text(cleaned.get("item"))
if "reason" in cleaned:
cleaned["reason"] = self._clean_text(cleaned.get("reason"))
if "payment_type" in cleaned:
cleaned["payment_type"] = self._normalize_payment_type(cleaned.get("payment_type"))
if "amount" in cleaned:
cleaned["amount"] = self._to_number(cleaned.get("amount"), as_int_if_possible=True)
if "quantity" in cleaned:
cleaned["quantity"] = self._to_number(cleaned.get("quantity"), as_int_if_possible=True)
# Business rules: amounts and quantities must be positive
if intent == "payment" and cleaned.get("amount") is not None and cleaned["amount"] < 0:
cleaned["amount"] = abs(cleaned["amount"])
if (
intent in {"order", "credit", "preparation"}
and cleaned.get("quantity") is not None
and cleaned["quantity"] < 0
):
cleaned["quantity"] = abs(cleaned["quantity"])
return cleaned
@staticmethod
def _clean_text(value: Any, *, title: bool = False) -> str | None:
if value is None:
return None
text = str(value).strip()
if not text:
return None
text = re.sub(r"\s+", " ", text)
return text.title() if title else text.lower()
@staticmethod
def _normalize_payment_type(value: Any) -> str | None:
text = DataValidator._clean_text(value)
if text is None:
return None
aliases = {
"gpay": "upi",
"google pay": "upi",
"phonepe": "upi",
"phone pe": "upi",
"paytm": "upi",
"upi": "upi",
"cash": "cash",
"online": "online",
"bank transfer": "online",
"neft": "online",
"imps": "online",
"rtgs": "online",
"cheque": "cheque",
"check": "cheque",
}
return aliases.get(text, text)
@staticmethod
def _to_number(value: Any, *, as_int_if_possible: bool = False) -> int | float | None:
if value is None or value == "":
return None
if isinstance(value, (int, float)) and not isinstance(value, bool):
number = float(value)
else:
text = str(value).replace(",", "").lower()
match = _NUMBER_PATTERN.search(text)
if not match:
return None
number = float(match.group(0))
if as_int_if_possible and number.is_integer():
return int(number)
return number
# ---------------------------------------------------------------------------
# Convenience wrapper — used by the orchestrator
# ---------------------------------------------------------------------------
def validate_data(intent: str, data: dict[str, Any]) -> dict[str, Any]:
"""
Normalise and validate extracted data for the given intent.
This is the function the orchestrator imports:
# (this is the canonical location — no re-import needed)
cleaned = validate_data(intent, raw_data)
Args:
intent: Detected intent string (e.g. "payment", "order").
data: Raw extraction dict from the Extraction Agent.
Returns:
Cleaned, normalised copy of the data dict.
"""
return DataValidator().validate(intent, data) |