#!/bin/bash

set -euo pipefail  # Exit on error, undefined vars, and pipe failures

# Configuration
readonly BASE_URL="https://download.swissmedicinfo.ch"
readonly PAGE_FILE="page.html"
readonly OUTPUT_FILE="swissmedic_data.zip"
readonly LOG_FILE="curl_output.log"
readonly FALLBACK_VIEWSTATEGENERATOR="CA0B0334"

# Logging function
log() {
    echo "[$(date '+%Y-%m-%d %H:%M:%S')] $*"
}

# Error handling function
error_exit() {
    log "ERROR: $1" >&2
    exit 1
}

# Cleanup function
cleanup() {
    local exit_code=$?
    if [[ $exit_code -ne 0 ]]; then
        log "Script failed. Check $LOG_FILE and $PAGE_FILE for debugging."
    fi
    exit $exit_code
}

trap cleanup EXIT

# Function to extract form field value
extract_field() {
    local field_id="$1"
    local file="$2"
    
    # More robust extraction using grep and sed
    grep -o "id=\"$field_id\"[^>]*value=\"[^\"]*\"" "$file" 2>/dev/null | \
        sed 's/.*value="\([^"]*\)".*/\1/' || true
}

# Function to validate extracted values
validate_field() {
    local field_name="$1"
    local field_value="$2"
    
    if [[ -z "$field_value" ]]; then
        log "WARNING: $field_name not found or empty"
        return 1
    fi
    
    log "Extracted $field_name: ${field_value:0:50}..."  # Show first 50 chars
    return 0
}

# Function to resolve form action URL
resolve_form_action() {
    local page_file="$1"
    
    # Extract form action with improved regex
    local form_action
    form_action=$(grep -o 'action="[^"]*"' "$page_file" 2>/dev/null | \
                  sed 's/action="\([^"]*\)"/\1/' | head -n1)
    
    if [[ -z "$form_action" ]]; then
        log "WARNING: Form action not found, using base URL"
        echo "$BASE_URL/"
        return
    fi
    
    # Resolve relative URLs
    case "$form_action" in
        http*)
            echo "$form_action"
            ;;
        /*)
            echo "$BASE_URL$form_action"
            ;;
        *)
            echo "$BASE_URL/$form_action"
            ;;
    esac
}

# Function to check if file is a valid ZIP
validate_zip_file() {
    local file="$1"
    
    if [[ ! -s "$file" ]]; then
        return 1
    fi
    
    # Check file signature (first 4 bytes should be PK for ZIP)
    if command -v file >/dev/null 2>&1; then
        file "$file" | grep -qi "zip\|archive" && return 0
    fi
    
    # Alternative check using hexdump
    if command -v hexdump >/dev/null 2>&1; then
        local signature
        signature=$(hexdump -C "$file" | head -n1 | cut -d' ' -f2-3)
        [[ "$signature" == "50 4b" ]] && return 0
    fi
    
    return 1
}

# Main execution starts here
log "Starting Swissmedic download script"

# Step 1: Fetch the page
log "Fetching page from $BASE_URL/"
if ! curl -s -f --max-time 30 --retry 3 "$BASE_URL/" > "$PAGE_FILE"; then
    error_exit "Failed to fetch page. Check network connection and URL availability."
fi

log "Page saved to $PAGE_FILE ($(wc -c < "$PAGE_FILE") bytes)"

# Step 2: Validate page contains form
if ! grep -q '<form' "$PAGE_FILE"; then
    error_exit "No form found in downloaded page. Site may have changed or require JavaScript."
fi

# Step 3: Extract form values
log "Extracting form values..."

VIEWSTATE=$(extract_field "__VIEWSTATE" "$PAGE_FILE")
EVENTVALIDATION=$(extract_field "__EVENTVALIDATION" "$PAGE_FILE")
VIEWSTATEGENERATOR=$(extract_field "__VIEWSTATEGENERATOR" "$PAGE_FILE")

# Validate critical fields
validate_field "__VIEWSTATE" "$VIEWSTATE" || error_exit "Critical field __VIEWSTATE missing"
validate_field "__EVENTVALIDATION" "$EVENTVALIDATION" || error_exit "Critical field __EVENTVALIDATION missing"

# Handle optional VIEWSTATEGENERATOR with fallback
if ! validate_field "__VIEWSTATEGENERATOR" "$VIEWSTATEGENERATOR"; then
    VIEWSTATEGENERATOR="$FALLBACK_VIEWSTATEGENERATOR"
    log "Using fallback __VIEWSTATEGENERATOR: $VIEWSTATEGENERATOR"
fi

# Step 4: Determine form action URL
FORM_ACTION=$(resolve_form_action "$PAGE_FILE")
log "Form action URL: $FORM_ACTION"

# Step 5: Send POST request
log "Sending POST request..."

if ! curl -L -f --max-time 60 --retry 2 \
    -o "$OUTPUT_FILE" \
    -X POST \
    -H "User-Agent: Mozilla/5.0 (compatible; SwissmedicDownloader/1.0)" \
    -H "Referer: $BASE_URL/" \
    --data-urlencode "__VIEWSTATE=$VIEWSTATE" \
    --data-urlencode "__VIEWSTATEGENERATOR=$VIEWSTATEGENERATOR" \
    --data-urlencode "__EVENTVALIDATION=$EVENTVALIDATION" \
    --data-urlencode "ctl00\$MainContent\$BtnYes=Yes" \
    "$FORM_ACTION" > "$LOG_FILE" 2>&1; then
    
    log "POST request failed. Server response:"
    cat "$LOG_FILE" >&2
    error_exit "Failed to download file. Check server response above."
fi

# Step 6: Validate downloaded file
log "Validating downloaded file..."

if validate_zip_file "$OUTPUT_FILE"; then
    local file_size
    file_size=$(wc -c < "$OUTPUT_FILE")
    log "SUCCESS: Downloaded valid ZIP file ($file_size bytes) to $OUTPUT_FILE"
    
    # Optional: Show ZIP contents
    if command -v unzip >/dev/null 2>&1; then
        log "ZIP file contents:"
        unzip -l "$OUTPUT_FILE" | head -n 20
    fi
else
    log "ERROR: Downloaded file is not a valid ZIP archive"
    log "File content preview:"
    head -n 10 "$OUTPUT_FILE" >&2
    error_exit "Invalid file downloaded"
fi

log "Script completed successfully"