#!/usr/bin/env python

import argparse
import json
import os
import sys
import time
from datetime import datetime
from concurrent.futures import ThreadPoolExecutor
from dns import message
from dns import query
from dns import rdatatype
from dns import rcode
from dns import name
from dns import flags
from dns import exception

# Get list of supported record types
supported_types = sorted([rdatatype.to_text(rdtype) for rdtype in rdatatype.RdataType])

# Dictionary of public DNS servers
builtin_nameservers = {
    'google': ['8.8.8.8', '8.8.4.4'],
    'cloudflare': ['1.1.1.1', '1.0.0.1', '1.1.1.2', '1.0.0.2'],
    'quad9': ['9.9.9.9', '149.112.112.112'],
    'gte': ['4.2.2.1', '4.2.2.2', '4.2.2.3', '4.2.2.4', '4.2.2.5', '4.2.2.6'],
    'opendns': ['208.67.222.222', '208.67.220.220', '208.67.222.123', '208.67.220.123'],
    'verisign': ['64.6.64.6', '64.6.65.6'],
    'comodo': ['8.26.56.26', '8.20.247.20'],
    'level3': ['209.244.0.3', '209.244.0.4'],
}

parser = argparse.ArgumentParser()
parser.add_argument('--domain', help='Helps if all records to query are on the same domain.')
parser.add_argument('--nameservers', nargs='+', help='IPv4 or IPv6 address of nameserver(s) to be queried.')
parser.add_argument('--ns-builtin', nargs='+', help=f'Built-in public DNS servers. Options: {", ".join(builtin_nameservers.keys())}, all')
parser.add_argument('--records', nargs='*', help='Hostname(s) to query. If not provided with --domain, queries the domain itself.')
parser.add_argument('--record-type', default='A', help='DNS record type to query. Default: A')
parser.add_argument('--threads', type=int, default=None, help='Number of threads to use for queries. Default: CPU count - 1')
parser.add_argument('--timeout', type=float, default=5.0, help='DNS query timeout in seconds. Default: 5.0')
parser.add_argument('--verbose', action='store_true', help='Enable verbose debugging output.')
parser.add_argument('--list-record-types', action='store_true', help='Print list of supported record types and exit.')
args = parser.parse_args()

# Build nameserver map: IP -> provider name
nameserver_map = {}  # ip -> provider_key
builtin_providers = []  # list of provider keys used

# Resolve builtin nameservers to actual IPs
if args.ns_builtin:
    # Handle 'all' special case
    if 'all' in [p.lower() for p in args.ns_builtin]:
        builtin_providers = list(builtin_nameservers.keys())
    else:
        builtin_providers = [p.lower() for p in args.ns_builtin]
    
    for provider in builtin_providers:
        if provider not in builtin_nameservers:
            print(f"Error: Unknown DNS provider '{provider}'")
            print(f"Available providers: {', '.join(builtin_nameservers.keys())}, all")
            exit(1)
        for ip in builtin_nameservers[provider]:
            nameserver_map[ip] = provider

# Track manual nameservers separately
manual_nameservers = []
if args.nameservers:
    manual_nameservers = args.nameservers
    for ip in args.nameservers:
        nameserver_map[ip] = "manual_selection"

# Combine all nameservers for querying
all_nameservers = list(nameserver_map.keys())

# Use resolved nameservers, or error if none provided
if not all_nameservers:
    print("Error: Either --nameservers or --ns-builtin must be provided")
    exit(1)

# Handle listing record types
if args.list_record_types:
    print("Supported DNS record types:")
    print((", ").join(supported_types))
    exit(0)

# Convert record type string to dns.rdatatype constant
try:
    record_type = rdatatype.from_text(args.record_type)
except rdatatype.UnknownRdatatype:
    print(f"Error: Unknown record type '{args.record_type}'")
    print(f"\nSupported record types:")
    print((', ').join(supported_types))
    exit(1)

# Function to perform a single DNS query
def perform_query(full_domain, ns, record_type, timeout):
    """Perform a DNS query and return results"""
    try:
        qname = name.from_text(full_domain)
        q = message.make_query(qname, record_type)
        
        # Add EDNS0 extension for better compatibility (like dig does)
        q.use_edns(edns=0, ednsflags=0, payload=4096)
        
        # Try TCP first for better compatibility with some servers
        try:
            r = query.tcp(q, ns, timeout=timeout)
            if args.verbose:
                print(f"[DEBUG] TCP Response: Answer={len(r.answer)} RRsets, TC flag: {bool(r.flags & flags.TC)}", file=sys.stderr)
        except Exception as e:
            # Fall back to UDP if TCP fails
            if args.verbose:
                print(f"[DEBUG] TCP failed ({e}), trying UDP...", file=sys.stderr)
            r = query.udp(q, ns, timeout=timeout)
            
            # If UDP returns truncated, retry with TCP
            if r.flags & flags.TC:
                if args.verbose:
                    print(f"[DEBUG] UDP Response truncated (TC flag set), retrying with TCP...", file=sys.stderr)
                r = query.tcp(q, ns, timeout=timeout)
                if args.verbose:
                    print(f"[DEBUG] TCP Retry Response: Answer={len(r.answer)} RRsets, TC flag: {bool(r.flags & flags.TC)}", file=sys.stderr)
        
        # Verbose debugging
        if args.verbose:
            print(f"\n[DEBUG] Query: {full_domain} (@{ns}) Type: {rdatatype.to_text(record_type)}", file=sys.stderr)
            print(f"[DEBUG] Response code: {rcode.to_text(r.rcode())}", file=sys.stderr)
            print(f"[DEBUG] Answer section: {len(r.answer)} RRsets", file=sys.stderr)
            print(f"[DEBUG] Authority section: {len(r.authority)} RRsets", file=sys.stderr)
            print(f"[DEBUG] Additional section: {len(r.additional)} RRsets", file=sys.stderr)
            print(f"[DEBUG] Full message: {r}", file=sys.stderr)
        
        results = []
        response_rcode = r.rcode()
        
        # Helper function to extract records of the requested type from a section
        def extract_records_of_type(section, requested_type):
            section_results = []
            if section:
                for rrset in section:
                    # Only include RRsets that match the requested record type
                    if rrset.rdtype == requested_type:
                        for item in rrset:
                            section_results.append(str(item))
            return section_results
        
        # Check answer section first
        results = extract_records_of_type(r.answer, record_type)
        if results:
            return (full_domain, ns, results)
        
        # If no answer, check additional section (some servers put records there)
        results = extract_records_of_type(r.additional, record_type)
        if results:
            return (full_domain, ns, results)
        
        # If still no results, check authority section (only for SOA, NS, etc.)
        results = extract_records_of_type(r.authority, record_type)
        if results:
            return (full_domain, ns, results)
        
        # Truly no records found - check if there's a non-NOERROR response code to report
        if response_rcode != 0:  # 0 = NOERROR
            rcode_name = rcode.to_text(response_rcode)
            return (full_domain, ns, f"RCODE: {rcode_name}")
        
        # No records and NOERROR response
        return (full_domain, ns, "NO RECORDS")
    except (exception.Timeout, TimeoutError):
        return (full_domain, ns, "TIMED OUT")
    except Exception as e:
        error_msg = str(e) if str(e) else type(e).__name__
        return (full_domain, ns, f"ERROR: {error_msg}")

# Start timer
start_time = time.time()
start_time_iso = datetime.fromtimestamp(start_time).isoformat()

# Determine what to query
if args.domain:
    # If records are provided, use them as subdomains; otherwise use the domain itself
    records_to_query = args.records if args.records else [args.domain]
    
    full_domains = []
    for rec in records_to_query:
        if rec == args.domain:
            full_domain = args.domain
        else:
            full_domain = f'{rec}.{args.domain}'
        full_domains.append(full_domain)
else:
    # Query each record as a full domain
    if not args.records:
        print("Error: Either --domain or --records must be provided")
        exit(1)
    full_domains = args.records

# Build list of query tasks
query_tasks = []
for full_domain in full_domains:
    for ns in all_nameservers:
        query_tasks.append((full_domain, ns, record_type))

# Determine thread count: use provided value or default to CPU count - 1, minimum 1
thread_count = args.threads if args.threads else max(1, os.cpu_count() - 1)

# Execute queries in parallel
results_map = {}  # (full_domain, ns) -> resolved_ip
with ThreadPoolExecutor(max_workers=thread_count) as executor:
    futures = [executor.submit(perform_query, full_domain, ns, record_type, args.timeout) 
               for full_domain, ns, _ in query_tasks]
    
    for future in futures:
        full_domain, ns, resolved_ips = future.result()
        if resolved_ips is not None:
            results_map[(full_domain, ns)] = resolved_ips

# Build the output structure
output = {}

# Initialize output structure based on provider grouping
for full_domain in full_domains:
    output[full_domain] = {"records": {}}
    
    # Group by provider if using builtin
    if builtin_providers:
        for provider in builtin_providers:
            output[full_domain]["records"][provider] = {}
        if manual_nameservers:
            output[full_domain]["records"]["manual_selection"] = {}
    else:
        # Just use the flat nameserver structure if no builtin used
        pass

# Populate results
for full_domain in full_domains:
    if builtin_providers:
        for ns in all_nameservers:
            provider = nameserver_map.get(ns)
            if (full_domain, ns) in results_map:
                results = results_map[(full_domain, ns)]
                # Handle error/timeout/rcode cases (strings)
                if isinstance(results, str):
                    output[full_domain]["records"][provider][ns] = results
                else:
                    results = sorted(results)
                    # Store as single value if only one result, otherwise as array
                    output[full_domain]["records"][provider][ns] = results[0] if len(results) == 1 else results
    else:
        # Flat structure for manual nameservers only
        for ns in all_nameservers:
            if (full_domain, ns) in results_map:
                results = results_map[(full_domain, ns)]
                # Handle error/timeout/rcode cases (strings)
                if isinstance(results, str):
                    output[full_domain]["records"][ns] = results
                else:
                    results = sorted(results)
                    # Store as single value if only one result, otherwise as array
                    output[full_domain]["records"][ns] = results[0] if len(results) == 1 else results

# Calculate and add timing information
end_time = time.time()
end_time_iso = datetime.fromtimestamp(end_time).isoformat()
elapsed_time = end_time - start_time

output["start_time"] = start_time_iso
output["end_time"] = end_time_iso
output["runtime"] = f"{elapsed_time:.2f}s"

print(json.dumps(output, indent=4))