#!/usr/bin/env bash
# RDLC AI Slop Scanner — TEMPLATE
# Ported from states-project-research RDLC slop section (claude-rdlc-wizard v0.1.0)
#
# Run: bash scripts/slop_scan.sh
# Exit code 0 = zero hits (pass), 1 = hits found
#
# Pattern: any banned phrase = a credibility issue. One "smoking gun" or
# "game-changer" and the reader stops trusting the whole document.

set -uo pipefail

# Default scan paths — customize for your project
SCAN_PATHS=("${@:-output/ research/ evidence/}")

# Allowlist (project-specific proper nouns, direct quotes)
ALLOWLIST=".rdlc/slop-allowlist.txt"

# --- Hard fail tier (rewrite immediately) ---
HARD_FAIL_PATTERN='deep dive|game.changer|cutting.edge|elevate|delve|tapestry|holistic|robust|paradigm|groundbreaking|streamline|empower|harness|unleash|unpack|pivotal|crucial|bigger picture|smoking gun|synergy|arguably|it.s worth noting|at the end of the day|in today.s world|needless to say'

# --- Soft warn tier (check context — legitimate use OK) ---
SOFT_WARN_PATTERN='leverage|comprehensive|landscape|stakeholder|navigate|facilitate|best practices|fundamentally|inherently'

red()    { printf "\033[31m%s\033[0m\n" "$1"; }
green()  { printf "\033[32m%s\033[0m\n" "$1"; }
yellow() { printf "\033[33m%s\033[0m\n" "$1"; }

echo "=== RDLC SLOP SCAN ==="
echo ""

# Build allowlist regex (one phrase per line in the allowlist file)
ALLOW_REGEX=""
if [ -f "$ALLOWLIST" ]; then
    ALLOW_REGEX=$(grep -vE '^[[:space:]]*(#|$)' "$ALLOWLIST" 2>/dev/null | tr '\n' '|' | sed 's/|$//')
fi

# --- Hard fail scan ---
echo "Hard fail tier:"
HARD_HITS=""
for path in "${SCAN_PATHS[@]}"; do
    [ ! -e "$path" ] && continue
    if [ -n "$ALLOW_REGEX" ]; then
        path_hits=$(grep -rniE "$HARD_FAIL_PATTERN" "$path" --include='*.md' 2>/dev/null \
            | grep -viE "$ALLOW_REGEX" || true)
    else
        path_hits=$(grep -rniE "$HARD_FAIL_PATTERN" "$path" --include='*.md' 2>/dev/null || true)
    fi
    if [ -n "$path_hits" ]; then
        HARD_HITS="${HARD_HITS}${path_hits}"$'\n'
    fi
done

if [ -n "$HARD_HITS" ]; then
    red "FAIL: hard-tier slop detected"
    printf '%s' "$HARD_HITS" | head -20
    HARD_COUNT=$(printf '%s' "$HARD_HITS" | grep -c '' 2>/dev/null || echo 0)
    echo ""
    red "Hard fail hits: $HARD_COUNT — rewrite before commit."
    HARD_FAIL=1
else
    green "PASS: zero hard-tier slop"
    HARD_FAIL=0
fi

echo ""

# --- Soft warn scan ---
echo "Soft warn tier (review for context):"
SOFT_HITS=""
for path in "${SCAN_PATHS[@]}"; do
    [ ! -e "$path" ] && continue
    if [ -n "$ALLOW_REGEX" ]; then
        path_hits=$(grep -rniE "$SOFT_WARN_PATTERN" "$path" --include='*.md' 2>/dev/null \
            | grep -viE "$ALLOW_REGEX" || true)
    else
        path_hits=$(grep -rniE "$SOFT_WARN_PATTERN" "$path" --include='*.md' 2>/dev/null || true)
    fi
    if [ -n "$path_hits" ]; then
        SOFT_HITS="${SOFT_HITS}${path_hits}"$'\n'
    fi
done

if [ -n "$SOFT_HITS" ]; then
    SOFT_COUNT=$(printf '%s' "$SOFT_HITS" | grep -c '' 2>/dev/null || echo 0)
    yellow "WARN: $SOFT_COUNT soft-tier hits — review context, may be legitimate domain use"
    printf '%s' "$SOFT_HITS" | head -10
else
    green "PASS: zero soft-tier slop"
fi

echo ""
echo "======================="

if [ "$HARD_FAIL" -eq 1 ]; then
    echo "Slop scan FAILED. Rewrite hard-tier hits or add proper-noun proper exclusions to $ALLOWLIST."
    exit 1
fi

echo "Slop scan PASSED."
exit 0