Spaces:
Sleeping
Sleeping
File size: 2,373 Bytes
85ba912 |
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 |
"""
Validation and gentle editing layer for generated copy.
This module:
- Applies banned-term replacements (e.g., "guaranteed" -> "aim to").
- Trims text to platform character cap.
- Returns an audit log of what changed.
"""
from typing import List, Dict, Tuple
from .platform_rules import PlatformConfig
# Soft language map: you can expand this list as you like
BANNED_MAP = {
"guaranteed": "aim to",
"guarantee": "aim to",
"no risk": "low risk",
}
def _apply_banned_terms(text: str) -> Tuple[str, List[Dict]]:
"""Replace banned phrases and record changes."""
audit: List[Dict] = []
cleaned = text
for bad, replacement in BANNED_MAP.items():
if bad.lower() in cleaned.lower():
before = cleaned
# simple case-insensitive replace
cleaned = cleaned.replace(bad, replacement)
cleaned = cleaned.replace(bad.capitalize(), replacement)
cleaned = cleaned.replace(bad.upper(), replacement.upper())
audit.append(
{
"rule": "banned_term",
"bad": bad,
"replacement": replacement,
}
)
return cleaned, audit
def _apply_length_cap(text: str, platform: PlatformConfig) -> Tuple[str, List[Dict]]:
"""Trim text to the platform's character cap if necessary."""
audit: List[Dict] = []
cap = platform.char_cap
if len(text) > cap:
before_len = len(text)
trimmed = text[:cap].rstrip()
audit.append(
{
"rule": "length_trim",
"before_len": before_len,
"after_len": len(trimmed),
"cap": cap,
}
)
return trimmed, audit
return text, audit
def validate_and_edit(
text: str,
platform: PlatformConfig,
) -> Tuple[str, List[Dict]]:
"""
Apply all validators in order and collect a combined audit log.
Returns:
final_text, audit_log
"""
audit_log: List[Dict] = []
# 1) banned terms
text, banned_audit = _apply_banned_terms(text)
audit_log.extend(banned_audit)
# 2) length trim
text, trim_audit = _apply_length_cap(text, platform)
audit_log.extend(trim_audit)
# (you can add more steps later: CTA normalization, emoji limits, etc.)
return text, audit_log
|