#!/usr/bin/env bash
# =============================================================================
# BreakPilot Compliance — Phase H RAG Ingestion
#
# Downloads and ingests ~35 new legal sources into Qdrant:
# - 16 German laws (gesetze-im-internet.de) → bp_compliance_gesetze
# - 18 EU regulations (EUR-Lex PDFs) → bp_compliance_ce
# - 3 framework docs (NIST, HLEG) → bp_compliance_datenschutz
#
# Run on Mac Mini:
#   bash ~/Projekte/breakpilot-compliance/scripts/ingest-phase-h.sh
# =============================================================================
set -euo pipefail

WORK_DIR="${WORK_DIR:-$HOME/rag-ingestion}"
RAG_URL="${RAG_URL:-https://localhost:8097/api/v1/documents/upload}"
QDRANT_URL="${QDRANT_URL:-http://localhost:6333}"
CURL_OPTS="-sk --connect-timeout 10 --max-time 300"
CURL_OPTS_LARGE="-sk --connect-timeout 10 --max-time 900"

UPLOADED=0
FAILED=0
SKIPPED=0

log()   { echo "[$(date '+%H:%M:%S')] $*"; }
ok()    { echo "[$(date '+%H:%M:%S')] ok $*"; }
warn()  { echo "[$(date '+%H:%M:%S')] WARN $*" >&2; }
fail()  { echo "[$(date '+%H:%M:%S')] FAIL $*" >&2; }

download_pdf() {
  local url="$1"
  local target="$2"
  if [[ -f "$target" ]]; then
    log "PDF exists: $(basename "$target") (skipping download)"
    return 0
  fi
  log "Downloading: $(basename "$target")"
  curl $CURL_OPTS -L "$url" -o "$target" 2>/dev/null || {
    warn "Download failed: $url"
    rm -f "$target"
    return 0
  }
  local fsize
  fsize=$(stat -f%z "$target" 2>/dev/null || stat -c%s "$target" 2>/dev/null || echo 0)
  if [[ "$fsize" -lt 1000 ]]; then
    warn "Download too small (${fsize}B): $(basename "$target")"
    rm -f "$target"
  fi
}

download_gesetz_pdf() {
  local law_id="$1"
  local target="$2"
  if [[ -f "$target" ]]; then
    log "PDF exists: $(basename "$target") (skipping download)"
    return 0
  fi
  log "Downloading: $law_id (gesetze-im-internet.de)"
  curl $CURL_OPTS_LARGE -L "https://www.gesetze-im-internet.de/${law_id}/gesamt.pdf" -o "$target" 2>/dev/null || {
    warn "Download failed: $law_id"
    rm -f "$target"
    return 0
  }
  local fsize
  fsize=$(stat -f%z "$target" 2>/dev/null || stat -c%s "$target" 2>/dev/null || echo 0)
  if [[ "$fsize" -lt 1000 ]]; then
    warn "Download too small (${fsize}B): $law_id"
    rm -f "$target"
  else
    log "  Downloaded: $(( fsize / 1024 ))KB"
  fi
}

upload_file() {
  local file="$1"
  local collection="$2"
  local data_type="$3"
  local use_case="$4"
  local year="$5"
  local metadata_json="$6"
  local label="${7:-$(basename "$file")}"

  if [[ ! -f "$file" ]]; then
    warn "File not found: $file"
    FAILED=$((FAILED + 1))
    return 0
  fi

  # Dedup check
  local reg_id
  reg_id=$(echo "$metadata_json" | python3 -c "import sys,json; print(json.load(sys.stdin).get('regulation_id',''))" 2>/dev/null || echo "")
  if [[ -n "$reg_id" ]]; then
    local existing
    existing=$(curl -sk --max-time 5 -X POST "${QDRANT_URL}/collections/${collection}/points/scroll" \
      -H "Content-Type: application/json" \
      -d "{\"filter\":{\"must\":[{\"key\":\"regulation_id\",\"match\":{\"value\":\"$reg_id\"}}]},\"limit\":1}" \
      2>/dev/null | python3 -c "import sys,json; r=json.load(sys.stdin).get('result',{}); print(len(r.get('points',[])))" 2>/dev/null || echo "0")
    if [[ "$existing" -gt 0 ]] 2>/dev/null; then
      log "SKIP (already in Qdrant): $label [regulation_id=$reg_id]"
      SKIPPED=$((SKIPPED + 1))
      return 0
    fi
  fi

  local filesize
  filesize=$(stat -f%z "$file" 2>/dev/null || stat -c%s "$file" 2>/dev/null || echo 0)
  if [[ "$filesize" -lt 100 ]]; then
    warn "File too small (${filesize}B): $label"
    SKIPPED=$((SKIPPED + 1))
    return 0
  fi

  log "Uploading: $label -> $collection ($(( filesize / 1024 ))KB)"

  local curl_opts="$CURL_OPTS"
  [[ "$filesize" -gt 256000 ]] && curl_opts="$CURL_OPTS_LARGE"

  local response
  response=$(curl $curl_opts -X POST "$RAG_URL" \
    -F "file=@${file}" \
    -F "collection=${collection}" \
    -F "data_type=${data_type}" \
    -F "use_case=${use_case}" \
    -F "year=${year}" \
    -F "chunk_strategy=recursive" \
    -F "chunk_size=1024" \
    -F "chunk_overlap=128" \
    -F "metadata_json=${metadata_json}" \
    2>/dev/null) || true

  if echo "$response" | grep -q '"chunks_count"\|"vectors_indexed"'; then
    local chunks
    chunks=$(echo "$response" | python3 -c "import sys,json; d=json.load(sys.stdin); print(d.get('chunks_count', d.get('vectors_indexed',0)))" 2>/dev/null || echo "?")
    ok "$label -> $chunks chunks"
    UPLOADED=$((UPLOADED + 1))
  else
    fail "Upload failed: $label"
    fail "Response: ${response:0:200}"
    FAILED=$((FAILED + 1))
  fi
}

# =============================================================================
# PHASE H-1: Downloads
# =============================================================================
phase_h_download() {
  log "=========================================="
  log "PHASE H-1: Downloads"
  log "=========================================="
  mkdir -p "$WORK_DIR/pdfs" "$WORK_DIR/texts"

  # --- EU Regulations (EUR-Lex) ---
  log "--- EUR-Lex PDFs ---"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32016R0679" \
    "$WORK_DIR/pdfs/dsgvo_2016_679.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32024R2847" \
    "$WORK_DIR/pdfs/cra_2024_2847.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32024R1689" \
    "$WORK_DIR/pdfs/ai_act_2024_1689.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32022L2555" \
    "$WORK_DIR/pdfs/nis2_2022_2555.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32022R1925" \
    "$WORK_DIR/pdfs/dma_2022_1925.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32011L0083" \
    "$WORK_DIR/pdfs/consumer_rights_2011_83.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32019L0770" \
    "$WORK_DIR/pdfs/digital_content_2019_770.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32019L0771" \
    "$WORK_DIR/pdfs/sale_of_goods_2019_771.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32000L0031" \
    "$WORK_DIR/pdfs/ecommerce_2000_31.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:31993L0013" \
    "$WORK_DIR/pdfs/unfair_terms_93_13.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32005L0029" \
    "$WORK_DIR/pdfs/unfair_practices_2005_29.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:31998L0006" \
    "$WORK_DIR/pdfs/price_indication_98_6.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32019L2161" \
    "$WORK_DIR/pdfs/omnibus_2019_2161.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32023R0988" \
    "$WORK_DIR/pdfs/gpsr_2023_988.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:31985L0374" \
    "$WORK_DIR/pdfs/product_liability_85_374.pdf"

  download_pdf "https://eur-lex.europa.eu/legal-content/DE/TXT/PDF/?uri=CELEX:32023R1542" \
    "$WORK_DIR/pdfs/battery_2023_1542.pdf"

  # --- German Laws (gesetze-im-internet.de) ---
  log "--- Deutsche Gesetze (PDFs) ---"

  download_gesetz_pdf "pangv" "$WORK_DIR/pdfs/pangv.pdf"
  download_gesetz_pdf "vsbg" "$WORK_DIR/pdfs/vsbg.pdf"
  download_gesetz_pdf "prodhaftg" "$WORK_DIR/pdfs/prodhaftg.pdf"
  download_gesetz_pdf "verpackg" "$WORK_DIR/pdfs/verpackg.pdf"
  download_gesetz_pdf "elektrog" "$WORK_DIR/pdfs/elektrog.pdf"
  download_gesetz_pdf "battg_2009" "$WORK_DIR/pdfs/battdg.pdf"
  download_gesetz_pdf "bfsg" "$WORK_DIR/pdfs/bfsg.pdf"
  download_gesetz_pdf "uwg_2004" "$WORK_DIR/pdfs/uwg.pdf"
  download_gesetz_pdf "bdsg_2018" "$WORK_DIR/pdfs/bdsg.pdf"
  download_gesetz_pdf "ddg" "$WORK_DIR/pdfs/ddg.pdf"
  download_gesetz_pdf "tkg_2021" "$WORK_DIR/pdfs/tkg.pdf"
  download_gesetz_pdf "hgb" "$WORK_DIR/pdfs/hgb.pdf"
  download_gesetz_pdf "ao_1977" "$WORK_DIR/pdfs/ao.pdf"
  download_gesetz_pdf "gewo" "$WORK_DIR/pdfs/gewo.pdf"
  download_gesetz_pdf "bgb" "$WORK_DIR/pdfs/bgb_komplett.pdf"
  download_gesetz_pdf "bgbeg" "$WORK_DIR/pdfs/egbgb.pdf"

  # --- NIST & HLEG ---
  log "--- Frameworks (NIST, HLEG) ---"

  download_pdf "https://nvlpubs.nist.gov/nistpubs/CSWP/NIST.CSWP.29.pdf" \
    "$WORK_DIR/pdfs/nist_csf_2_0.pdf"

  download_pdf "https://nvlpubs.nist.gov/nistpubs/CSWP/NIST.CSWP.01162020.pdf" \
    "$WORK_DIR/pdfs/nist_privacy_framework.pdf"

  download_pdf "https://op.europa.eu/en/publication-detail/-/publication/d3988569-0434-11ea-8c1f-01aa75ed71a1" \
    "$WORK_DIR/pdfs/hleg_trustworthy_ai.pdf"

  log "Downloads abgeschlossen."
}

# =============================================================================
# PHASE H-2: Deutsche Gesetze → bp_compliance_gesetze
# =============================================================================
phase_h_gesetze() {
  log "=========================================="
  log "PHASE H-2: Deutsche Gesetze -> bp_compliance_gesetze"
  log "=========================================="

  local col="bp_compliance_gesetze"

  upload_file "$WORK_DIR/pdfs/pangv.pdf" "$col" "compliance" "legal_reference" "2022" \
    '{"regulation_id":"pangv","regulation_name_de":"Preisangabenverordnung (PAngV)","category":"verbraucherschutz","license":"public_law","source":"gesetze-im-internet.de"}' \
    "PAngV (Preisangabenverordnung)"

  upload_file "$WORK_DIR/pdfs/vsbg.pdf" "$col" "compliance" "legal_reference" "2016" \
    '{"regulation_id":"vsbg","regulation_name_de":"Verbraucherstreitbeilegungsgesetz (VSBG)","category":"verbraucherschutz","license":"public_law","source":"gesetze-im-internet.de"}' \
    "VSBG (Verbraucherstreitbeilegung)"

  upload_file "$WORK_DIR/pdfs/prodhaftg.pdf" "$col" "compliance" "legal_reference" "1989" \
    '{"regulation_id":"prodhaftg","regulation_name_de":"Produkthaftungsgesetz (ProdHaftG)","category":"produkthaftung","license":"public_law","source":"gesetze-im-internet.de"}' \
    "ProdHaftG (Produkthaftung)"

  upload_file "$WORK_DIR/pdfs/verpackg.pdf" "$col" "compliance" "legal_reference" "2017" \
    '{"regulation_id":"verpackg","regulation_name_de":"Verpackungsgesetz (VerpackG)","category":"umwelt","license":"public_law","source":"gesetze-im-internet.de"}' \
    "VerpackG (Verpackungsgesetz)"

  upload_file "$WORK_DIR/pdfs/elektrog.pdf" "$col" "compliance" "legal_reference" "2015" \
    '{"regulation_id":"elektrog","regulation_name_de":"Elektro- und Elektronikgeraetegesetz (ElektroG)","category":"umwelt","license":"public_law","source":"gesetze-im-internet.de"}' \
    "ElektroG (WEEE)"

  upload_file "$WORK_DIR/pdfs/battdg.pdf" "$col" "compliance" "legal_reference" "2009" \
    '{"regulation_id":"battdg","regulation_name_de":"Batteriegesetz (BattG)","category":"umwelt","license":"public_law","source":"gesetze-im-internet.de"}' \
    "BattG (Batteriegesetz)"

  upload_file "$WORK_DIR/pdfs/bfsg.pdf" "$col" "compliance" "legal_reference" "2021" \
    '{"regulation_id":"bfsg","regulation_name_de":"Barrierefreiheitsstaerkungsgesetz (BFSG)","category":"barrierefreiheit","license":"public_law","source":"gesetze-im-internet.de"}' \
    "BFSG (Barrierefreiheit)"

  upload_file "$WORK_DIR/pdfs/uwg.pdf" "$col" "compliance" "legal_reference" "2004" \
    '{"regulation_id":"uwg","regulation_name_de":"Gesetz gegen den unlauteren Wettbewerb (UWG)","category":"wettbewerb","license":"public_law","source":"gesetze-im-internet.de"}' \
    "UWG (Unlauterer Wettbewerb)"

  upload_file "$WORK_DIR/pdfs/bdsg.pdf" "$col" "compliance" "legal_reference" "2018" \
    '{"regulation_id":"bdsg","regulation_name_de":"Bundesdatenschutzgesetz (BDSG)","category":"datenschutz","license":"public_law","source":"gesetze-im-internet.de"}' \
    "BDSG (Bundesdatenschutzgesetz)"

  upload_file "$WORK_DIR/pdfs/ddg.pdf" "$col" "compliance" "legal_reference" "2024" \
    '{"regulation_id":"ddg","regulation_name_de":"Digitale-Dienste-Gesetz (DDG)","category":"plattformen","license":"public_law","source":"gesetze-im-internet.de"}' \
    "DDG (Digitale-Dienste-Gesetz, komplett)"

  upload_file "$WORK_DIR/pdfs/tkg.pdf" "$col" "compliance" "legal_reference" "2021" \
    '{"regulation_id":"tkg","regulation_name_de":"Telekommunikationsgesetz (TKG)","category":"telekommunikation","license":"public_law","source":"gesetze-im-internet.de"}' \
    "TKG (Telekommunikationsgesetz)"

  upload_file "$WORK_DIR/pdfs/hgb.pdf" "$col" "compliance" "legal_reference" "1897" \
    '{"regulation_id":"hgb","regulation_name_de":"Handelsgesetzbuch (HGB)","category":"handelsrecht","license":"public_law","source":"gesetze-im-internet.de"}' \
    "HGB (Handelsgesetzbuch)"

  upload_file "$WORK_DIR/pdfs/ao.pdf" "$col" "compliance" "legal_reference" "1977" \
    '{"regulation_id":"ao","regulation_name_de":"Abgabenordnung (AO)","category":"steuerrecht","license":"public_law","source":"gesetze-im-internet.de"}' \
    "AO (Abgabenordnung)"

  upload_file "$WORK_DIR/pdfs/gewo.pdf" "$col" "compliance" "legal_reference" "1999" \
    '{"regulation_id":"gewo","regulation_name_de":"Gewerbeordnung (GewO)","category":"gewerberecht","license":"public_law","source":"gesetze-im-internet.de"}' \
    "GewO (Gewerbeordnung)"

  upload_file "$WORK_DIR/pdfs/bgb_komplett.pdf" "$col" "compliance" "legal_reference" "2002" \
    '{"regulation_id":"bgb_komplett","regulation_name_de":"Buergerliches Gesetzbuch (BGB, komplett)","category":"zivilrecht","license":"public_law","source":"gesetze-im-internet.de"}' \
    "BGB (komplett als PDF)"

  upload_file "$WORK_DIR/pdfs/egbgb.pdf" "$col" "compliance" "legal_reference" "1896" \
    '{"regulation_id":"egbgb_komplett","regulation_name_de":"EGBGB (komplett)","category":"zivilrecht","license":"public_law","source":"gesetze-im-internet.de"}' \
    "EGBGB (komplett als PDF)"
}

# =============================================================================
# PHASE H-3: EU-Rechtstexte → bp_compliance_ce
# =============================================================================
phase_h_eu() {
  log "=========================================="
  log "PHASE H-3: EU-Rechtstexte -> bp_compliance_ce"
  log "=========================================="

  local col="bp_compliance_ce"

  upload_file "$WORK_DIR/pdfs/dsgvo_2016_679.pdf" "$col" "compliance_ce" "legal_reference" "2016" \
    '{"regulation_id":"eu_2016_679","regulation_name_de":"Datenschutz-Grundverordnung (DSGVO)","regulation_name_en":"General Data Protection Regulation","regulation_short":"DSGVO","category":"datenschutz","celex":"32016R0679","source":"eur-lex","license":"public_law"}' \
    "DSGVO (EU) 2016/679"

  upload_file "$WORK_DIR/pdfs/cra_2024_2847.pdf" "$col" "compliance_ce" "legal_reference" "2024" \
    '{"regulation_id":"eu_2024_2847","regulation_name_de":"Cyber Resilience Act (CRA)","regulation_name_en":"Cyber Resilience Act","regulation_short":"CRA","category":"cybersecurity","celex":"32024R2847","source":"eur-lex","license":"public_law"}' \
    "Cyber Resilience Act (EU) 2024/2847"

  upload_file "$WORK_DIR/pdfs/ai_act_2024_1689.pdf" "$col" "compliance_ce" "legal_reference" "2024" \
    '{"regulation_id":"eu_2024_1689","regulation_name_de":"KI-Verordnung (AI Act)","regulation_name_en":"Artificial Intelligence Act","regulation_short":"AI Act","category":"ki_regulierung","celex":"32024R1689","source":"eur-lex","license":"public_law"}' \
    "AI Act (EU) 2024/1689"

  upload_file "$WORK_DIR/pdfs/nis2_2022_2555.pdf" "$col" "compliance_ce" "legal_reference" "2022" \
    '{"regulation_id":"eu_2022_2555","regulation_name_de":"NIS2-Richtlinie","regulation_name_en":"NIS2 Directive","regulation_short":"NIS2","category":"cybersecurity","celex":"32022L2555","source":"eur-lex","license":"public_law"}' \
    "NIS2 Directive (EU) 2022/2555"

  upload_file "$WORK_DIR/pdfs/dma_2022_1925.pdf" "$col" "compliance_ce" "legal_reference" "2022" \
    '{"regulation_id":"eu_2022_1925","regulation_name_de":"Digital Markets Act (DMA)","regulation_name_en":"Digital Markets Act","regulation_short":"DMA","category":"plattformregulierung","celex":"32022R1925","source":"eur-lex","license":"public_law"}' \
    "Digital Markets Act (EU) 2022/1925"

  upload_file "$WORK_DIR/pdfs/consumer_rights_2011_83.pdf" "$col" "compliance_ce" "legal_reference" "2011" \
    '{"regulation_id":"eu_2011_83","regulation_name_de":"Verbraucherrechte-Richtlinie","regulation_name_en":"Consumer Rights Directive","regulation_short":"CRD","category":"verbraucherschutz","celex":"32011L0083","source":"eur-lex","license":"public_law"}' \
    "Consumer Rights Directive 2011/83/EU"

  upload_file "$WORK_DIR/pdfs/digital_content_2019_770.pdf" "$col" "compliance_ce" "legal_reference" "2019" \
    '{"regulation_id":"eu_2019_770","regulation_name_de":"Digitale-Inhalte-Richtlinie","regulation_name_en":"Digital Content Directive","regulation_short":"DCD","category":"verbraucherschutz","celex":"32019L0770","source":"eur-lex","license":"public_law"}' \
    "Digital Content Directive 2019/770"

  upload_file "$WORK_DIR/pdfs/sale_of_goods_2019_771.pdf" "$col" "compliance_ce" "legal_reference" "2019" \
    '{"regulation_id":"eu_2019_771","regulation_name_de":"Warenkauf-Richtlinie","regulation_name_en":"Sale of Goods Directive","regulation_short":"SGD","category":"verbraucherschutz","celex":"32019L0771","source":"eur-lex","license":"public_law"}' \
    "Sale of Goods Directive 2019/771"

  upload_file "$WORK_DIR/pdfs/ecommerce_2000_31.pdf" "$col" "compliance_ce" "legal_reference" "2000" \
    '{"regulation_id":"eu_2000_31","regulation_name_de":"E-Commerce-Richtlinie","regulation_name_en":"E-Commerce Directive","regulation_short":"ECD","category":"plattformregulierung","celex":"32000L0031","source":"eur-lex","license":"public_law"}' \
    "E-Commerce Directive 2000/31/EC"

  upload_file "$WORK_DIR/pdfs/unfair_terms_93_13.pdf" "$col" "compliance_ce" "legal_reference" "1993" \
    '{"regulation_id":"eu_1993_13","regulation_name_de":"Klausel-Richtlinie","regulation_name_en":"Unfair Contract Terms Directive","regulation_short":"UCTD","category":"verbraucherschutz","celex":"31993L0013","source":"eur-lex","license":"public_law"}' \
    "Unfair Contract Terms Directive 93/13/EEC"

  upload_file "$WORK_DIR/pdfs/unfair_practices_2005_29.pdf" "$col" "compliance_ce" "legal_reference" "2005" \
    '{"regulation_id":"eu_2005_29","regulation_name_de":"UGP-Richtlinie","regulation_name_en":"Unfair Commercial Practices Directive","regulation_short":"UCPD","category":"verbraucherschutz","celex":"32005L0029","source":"eur-lex","license":"public_law"}' \
    "Unfair Commercial Practices Directive 2005/29/EC"

  upload_file "$WORK_DIR/pdfs/price_indication_98_6.pdf" "$col" "compliance_ce" "legal_reference" "1998" \
    '{"regulation_id":"eu_1998_6","regulation_name_de":"Preisangaben-Richtlinie","regulation_name_en":"Price Indication Directive","regulation_short":"PID","category":"verbraucherschutz","celex":"31998L0006","source":"eur-lex","license":"public_law"}' \
    "Price Indication Directive 98/6/EC"

  upload_file "$WORK_DIR/pdfs/omnibus_2019_2161.pdf" "$col" "compliance_ce" "legal_reference" "2019" \
    '{"regulation_id":"eu_2019_2161","regulation_name_de":"Omnibus-Richtlinie","regulation_name_en":"Omnibus Directive","regulation_short":"Omnibus","category":"verbraucherschutz","celex":"32019L2161","source":"eur-lex","license":"public_law"}' \
    "Omnibus Directive 2019/2161"

  upload_file "$WORK_DIR/pdfs/gpsr_2023_988.pdf" "$col" "compliance_ce" "legal_reference" "2023" \
    '{"regulation_id":"eu_2023_988","regulation_name_de":"Allgemeine Produktsicherheitsverordnung (GPSR)","regulation_name_en":"General Product Safety Regulation","regulation_short":"GPSR","category":"produktsicherheit","celex":"32023R0988","source":"eur-lex","license":"public_law"}' \
    "GPSR (EU) 2023/988"

  upload_file "$WORK_DIR/pdfs/product_liability_85_374.pdf" "$col" "compliance_ce" "legal_reference" "1985" \
    '{"regulation_id":"eu_1985_374","regulation_name_de":"Produkthaftungsrichtlinie","regulation_name_en":"Product Liability Directive","regulation_short":"PLD","category":"produkthaftung","celex":"31985L0374","source":"eur-lex","license":"public_law"}' \
    "Product Liability Directive 85/374/EEC"

  upload_file "$WORK_DIR/pdfs/battery_2023_1542.pdf" "$col" "compliance_ce" "legal_reference" "2023" \
    '{"regulation_id":"eu_2023_1542","regulation_name_de":"Batterieverordnung","regulation_name_en":"Battery Regulation","regulation_short":"BattVO","category":"umwelt","celex":"32023R1542","source":"eur-lex","license":"public_law"}' \
    "Batterieverordnung (EU) 2023/1542"
}

# =============================================================================
# PHASE H-4: Datenschutz-Frameworks → bp_compliance_datenschutz
# =============================================================================
phase_h_datenschutz() {
  log "=========================================="
  log "PHASE H-4: Frameworks -> bp_compliance_datenschutz"
  log "=========================================="

  local col="bp_compliance_datenschutz"

  upload_file "$WORK_DIR/pdfs/nist_csf_2_0.pdf" "$col" "compliance" "framework" "2024" \
    '{"regulation_id":"nist_csf_2_0","regulation_name_de":"NIST Cybersecurity Framework 2.0","regulation_name_en":"NIST Cybersecurity Framework 2.0","regulation_short":"NIST CSF 2.0","category":"security","license":"public_domain_us","source":"nist.gov"}' \
    "NIST Cybersecurity Framework 2.0"

  upload_file "$WORK_DIR/pdfs/nist_privacy_framework.pdf" "$col" "compliance" "framework" "2020" \
    '{"regulation_id":"nist_privacy_1_0","regulation_name_de":"NIST Privacy Framework 1.0","regulation_name_en":"NIST Privacy Framework 1.0","regulation_short":"NIST PF 1.0","category":"datenschutz","license":"public_domain_us","source":"nist.gov"}' \
    "NIST Privacy Framework 1.0"

  # HLEG may need special handling - the op.europa.eu link may redirect
  if [[ -f "$WORK_DIR/pdfs/hleg_trustworthy_ai.pdf" ]]; then
    upload_file "$WORK_DIR/pdfs/hleg_trustworthy_ai.pdf" "$col" "compliance" "framework" "2019" \
      '{"regulation_id":"hleg_trustworthy_ai","regulation_name_de":"HLEG Ethik-Leitlinien Vertrauenswuerdige KI","regulation_name_en":"HLEG Ethics Guidelines for Trustworthy AI","regulation_short":"HLEG AI","category":"ki_ethik","license":"cc_by_4","source":"ec.europa.eu"}' \
      "HLEG Ethics Guidelines Trustworthy AI"
  else
    warn "HLEG PDF not available (download may have failed)"
  fi
}

# =============================================================================
# MAIN
# =============================================================================
main() {
  log "=========================================="
  log "PHASE H: RAG Ingestion — ~37 neue Dokumente"
  log "=========================================="
  log "Work dir: $WORK_DIR"
  log "RAG URL: $RAG_URL"

  phase_h_download
  phase_h_gesetze
  phase_h_eu
  phase_h_datenschutz

  log "=========================================="
  log "PHASE H ABGESCHLOSSEN"
  log "  Hochgeladen: $UPLOADED"
  log "  Uebersprungen: $SKIPPED"
  log "  Fehlgeschlagen: $FAILED"
  log "=========================================="
}

main "$@"