gh-christianlouis-dmarq/backend/app/services/dmarc_parser.py

import gzip
import io
import logging
import zipfile
from datetime import datetime
from typing import Any, Dict, Optional

import defusedxml.ElementTree as ET

# Set up logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

# Security constants for file upload protection
MAX_FILE_SIZE = 10 * 1024 * 1024  # 10 MB
MAX_UNCOMPRESSED_SIZE = 100 * 1024 * 1024  # 100 MB for zip bomb protection
MAX_FILES_IN_ARCHIVE = 10  # Maximum number of files in a zip archive


class DMARCParser:
    """
    Parser for DMARC Aggregate Reports (XML format)
    """

    @staticmethod
    def parse_file(file_content: bytes, filename: str) -> Dict[str, Any]:
        """
        Parse a DMARC report file (XML, zip, or gzip) into a dictionary

        Args:
            file_content: The binary content of the file
            filename: The name of the file (used to determine type)

        Returns:
            Dict containing the parsed report data

        Raises:
            ValueError: If file is invalid, too large, or potentially malicious
        """
        # Security: Check file size
        if len(file_content) > MAX_FILE_SIZE:
            raise ValueError(
                f"File too large. Maximum size is {MAX_FILE_SIZE / (1024*1024):.1f} MB"
            )

        # Determine file type and extract XML content
        xml_content = DMARCParser._extract_xml_content(file_content, filename)
        if not xml_content:
            raise ValueError("Could not extract XML content from file")

        # Security: Check uncompressed XML size
        if len(xml_content) > MAX_UNCOMPRESSED_SIZE:
            raise ValueError(
                f"Uncompressed content too large ({len(xml_content) / (1024*1024):.1f} MB). "
                f"Maximum is {MAX_UNCOMPRESSED_SIZE / (1024*1024):.1f} MB. "
                "Possible zip bomb attack detected."
            )

        # Parse the XML content
        return DMARCParser._parse_xml(xml_content)

    @staticmethod
    def _extract_from_zip(file_content: bytes) -> Optional[bytes]:
        """Extract the first XML file from a ZIP archive.

        Raises:
            ValueError: If the archive exceeds size/count security limits.
        """
        try:
            with zipfile.ZipFile(io.BytesIO(file_content)) as z:
                file_list = z.infolist()

                # Security: Check number of files in archive
                if len(file_list) > MAX_FILES_IN_ARCHIVE:
                    raise ValueError(
                        f"ZIP archive contains too many files ({len(file_list)}). "
                        f"Maximum is {MAX_FILES_IN_ARCHIVE}."
                    )

                # Security: Check for zip bomb by examining compression ratios
                total_uncompressed = sum(f.file_size for f in file_list)
                if total_uncompressed > MAX_UNCOMPRESSED_SIZE:
                    raise ValueError(
                        f"ZIP archive uncompressed size too large "
                        f"({total_uncompressed / (1024*1024):.1f} MB). "
                        f"Maximum is {MAX_UNCOMPRESSED_SIZE / (1024*1024):.1f} MB. "
                        "Possible zip bomb attack detected."
                    )

                # Find the first XML file in the archive
                for file_info in file_list:
                    if file_info.filename.lower().endswith(".xml"):
                        # Security: Double-check individual file size
                        if file_info.file_size > MAX_UNCOMPRESSED_SIZE:
                            raise ValueError(
                                f"XML file in archive too large "
                                f"({file_info.file_size / (1024*1024):.1f} MB)"
                            )
                        return z.read(file_info.filename)
        except zipfile.BadZipFile:
            pass
        return None

    @staticmethod
    def _extract_xml_content(file_content: bytes, filename: str) -> Optional[bytes]:
        """
        Extract XML content from various file formats (ZIP, GZIP, or plain XML)

        Raises:
            ValueError: If archive contains too many files or is potentially malicious
        """
        # Try to handle as ZIP file
        if filename.lower().endswith(".zip"):
            result = DMARCParser._extract_from_zip(file_content)
            if result is not None:
                return result

        # Try to handle as GZIP file
        if filename.lower().endswith(".gz") or filename.lower().endswith(".gzip"):
            try:
                return gzip.decompress(file_content)
            except gzip.BadGzipFile:
                pass

        # Assume it's plain XML
        if filename.lower().endswith(".xml"):
            return file_content

        return None

    @staticmethod
    def _strip_namespace(el) -> None:
        """Recursively remove XML namespace prefixes from element tags in-place."""
        if "}" in el.tag:
            el.tag = el.tag.split("}", 1)[1]
        for child in el:
            DMARCParser._strip_namespace(child)

    @staticmethod
    def _parse_metadata(root) -> dict:
        """Parse the report_metadata section of a DMARC XML report."""
        report: dict = {}
        metadata = root.find("report_metadata")
        if metadata is not None:
            report["report_id"] = metadata.findtext("report_id", "")
            report["org_name"] = metadata.findtext("org_name", "")
            report["email"] = metadata.findtext("email", "")

            date_range = metadata.find("date_range")
            if date_range is not None:
                begin_ts = int(date_range.findtext("begin", 0))
                end_ts = int(date_range.findtext("end", 0))
                report["begin_date"] = datetime.fromtimestamp(begin_ts).isoformat()
                report["end_date"] = datetime.fromtimestamp(end_ts).isoformat()
                report["begin_timestamp"] = begin_ts
                report["end_timestamp"] = end_ts
        return report

    @staticmethod
    def _parse_record(record_elem) -> dict:
        """Parse a single <record> element into a dictionary."""
        record: dict = {}

        row = record_elem.find("row")
        if row is not None:
            record["source_ip"] = row.findtext("source_ip", "")
            record["count"] = int(row.findtext("count", 0))
            policy_evaluated = row.find("policy_evaluated")
            if policy_evaluated is not None:
                record["disposition"] = policy_evaluated.findtext("disposition", "none")
                record["dkim_result"] = policy_evaluated.findtext("dkim", "").lower()
                record["spf_result"] = policy_evaluated.findtext("spf", "").lower()

        identifiers = record_elem.find("identifiers")
        if identifiers is not None:
            record["header_from"] = identifiers.findtext("header_from", "")

        auth_results = record_elem.find("auth_results")
        if auth_results is not None:
            spf_entries = [
                {
                    "domain": spf.findtext("domain", ""),
                    "result": spf.findtext("result", "").lower(),
                }
                for spf in auth_results.findall("spf")
            ]
            if spf_entries:
                record["spf"] = spf_entries

            dkim_entries = [
                {
                    "domain": dkim.findtext("domain", ""),
                    "result": dkim.findtext("result", "").lower(),
                    "selector": dkim.findtext("selector", ""),
                }
                for dkim in auth_results.findall("dkim")
            ]
            if dkim_entries:
                record["dkim"] = dkim_entries

        return record

    @staticmethod
    def _compute_summary(records: list) -> dict:
        """Compute aggregate pass/fail statistics for a list of records."""
        total_count = sum(r["count"] for r in records)
        passed_count = sum(
            r["count"]
            for r in records
            if r.get("spf_result") == "pass" or r.get("dkim_result") == "pass"
        )
        failed_count = total_count - passed_count
        return {
            "total_count": total_count,
            "passed_count": passed_count,
            "failed_count": failed_count,
            "pass_rate": (passed_count / total_count * 100) if total_count > 0 else 0,
        }

    @staticmethod
    def _parse_xml(xml_content: bytes) -> Dict[str, Any]:
        """
        Parse DMARC XML content according to RFC 7489
        """
        try:
            root = ET.fromstring(xml_content)
            DMARCParser._strip_namespace(root)

            report = DMARCParser._parse_metadata(root)

            # Parse policy published
            policy = root.find("policy_published")
            if policy is not None:
                report["domain"] = policy.findtext("domain", "")
                report["policy"] = {
                    "p": policy.findtext("p", "none"),
                    "sp": policy.findtext("sp", ""),
                    "pct": policy.findtext("pct", "100"),
                }

            # Parse records
            records = [DMARCParser._parse_record(elem) for elem in root.findall("record")]
            report["records"] = records
            report["summary"] = DMARCParser._compute_summary(records)

            # Log parse results for debugging
            total_count = report["summary"]["total_count"]
            logger.info("Parsed DMARC report for domain: %s", report.get("domain"))
            logger.info("Found %s record entries with %s total messages", len(records), total_count)
            logger.info(
                "Messages passed: %s, failed: %s",
                report["summary"]["passed_count"],
                report["summary"]["failed_count"],
            )
            if records:
                logger.info(
                    "Sample record - SPF: %s, DKIM: %s",
                    records[0].get("spf_result"),
                    records[0].get("dkim_result"),
                )

            return report

        except Exception as e:
            logger.error("Error parsing DMARC XML: %s", str(e))
            raise ValueError(f"Error parsing DMARC XML: {str(e)}") from e