Haibo Xu

import hashlib
import inspect
from abc import ABCMeta, abstractmethod
from dataclasses import asdict
from functools import cached_property
from datetime import datetime, timezone
import math
from typing import Counter, List, Optional, Dict, Any, Tuple
from redb import settings
from redb.models.dataclasses import Hash
from .database_exporters import DatabaseExporter, ElasticsearchExporter, ClickHouseExporter, PrintExporter

class Extractor(metaclass=ABCMeta):
    def __init__(
        self,
        filepath: str,
        log: Any,
        exporters: Optional[List[DatabaseExporter]] = None,
        index_prefix: Optional[str] = None,
        source: Optional[str] = None,
        elastic_index: Optional[str] = None,
        known_benign: bool = False,
        known_malicious: bool = False,
        precomputed_hashes: Optional[Dict[str, str]] = None,
    ):
        self.log = log
        self.log.debug(f"Creating {self.__class__.__name__}")
        self.filepath = filepath
        self.source = source
        self.exporters = exporters or []
        self.index_prefix = index_prefix if index_prefix else settings.ELASTIC_BINARIES_COLLECTION
        self.elastic_index = self.index_prefix + (elastic_index if elastic_index else "")
        self.known_benign = known_benign
        self.known_malicious = known_malicious

        # Use precomputed hashes if provided (e.g., from machofile, pefile)
        # Otherwise compute them from binary
        if precomputed_hashes:
            self.md5 = precomputed_hashes.get('md5') or precomputed_hashes.get('MD5')
            self.sha1 = precomputed_hashes.get('sha1') or precomputed_hashes.get('SHA1')
            self.sha256 = precomputed_hashes.get('sha256') or precomputed_hashes.get('SHA256')
        else:
            self.md5 = hashlib.md5(self.binary).hexdigest()
            self.sha1 = hashlib.sha1(self.binary).hexdigest()
            self.sha256 = hashlib.sha256(self.binary).hexdigest()
        self.hash = Hash(self.md5, self.sha1, self.sha256)

    @cached_property
    def binary(self):
        with open(self.filepath, "rb") as f:
            data = f.read()
        return data

    @property
    @abstractmethod
    def tag(self):
        pass

    @abstractmethod
    def extract(self):
        """
        this method defines the extracted data
        """

    @staticmethod
    def process_binary_string(s):
        # Remove \x00 padding
        s = s.rstrip(b"\x00")

        # Check if there are any non-printable characters
        has_non_printable = any(byte < 32 or byte > 126 for byte in s)

        if not has_non_printable:
            # If all characters are printable, decode the string
            return s.decode()
        else:
            # If there are non-printable characters, represent them as \xDD
            return "".join(
                [
                    f"\\x{byte:02x}" if byte < 32 or byte > 126 else chr(byte)
                    for byte in s
                ]
            )

    @staticmethod
    def remove_non_utf8(binary_string):
        decoded = b""
        for i in range(len(binary_string)):
            try:
                # Try to decode each byte
                char = binary_string[i : i + 1].decode("utf-8")
                decoded += char.encode("utf-8")
            except UnicodeDecodeError:
                # Skip this byte if it can't be decoded
                continue
        return decoded

    def calculate_entropy(self, data):
        """Calculate the entropy of a chunk of data.
        Based on pefile.SectionStructure.entropy_H.
        """
        # self.log.debug(inspect.currentframe().f_code.co_name)
        if not data:
            return 0.0

        if type(data) == str:
            counts = Counter(data)
            frequencies = ((i / len(data)) for i in counts.values())
            return - sum(f * math.log(f, 2) for f in frequencies)
        else:
            occurences = Counter(bytearray(data))
            entropy = 0
            for x in occurences.values():
                p_x = float(x) / len(data)
                entropy -= p_x * math.log(p_x, 2)
            return entropy

    @abstractmethod
    def prepare_export_data(self, exporter_type: str) -> Tuple[List[Any], List[str], List[str]]:
        """
        Prepare data for specific export type
        Returns:
            Tuple containing:
            - data: List of values to insert
            - column_names: List of column names
            - column_type_names: List of column types
        """
        pass

    def export_data(self):
        """Export data to all configured exporters

        Returns:
            True: Export succeeded
            False: Export failed (actual error)
            None: No data to export (not an error, e.g., no overlay, no signature)
        """
        self.log.debug(inspect.currentframe().f_code.co_name)
        success = True
        extracted_data = self.extract()

        if extracted_data is None:
            self.log.debug("extract() returned None, skipping export")
            return None  # No data to export, not a failure
            
        for exporter in self.exporters:
            if isinstance(exporter, PrintExporter):
                # For PrintExporter, we pass the extracted data directly
                success &= exporter.export(extracted_data)
            else:
                # Get the data prepared for this specific exporter type
                export_data = self.prepare_export_data(exporter.__class__.__name__)
                
                if export_data is None:
                    self.log.debug(f"prepare_export_data returned None for {exporter.__class__.__name__}")
                    return False
                
                if isinstance(exporter, ElasticsearchExporter):
                    success &= exporter.export(
                        export_data,
                        index=self.elastic_index,
                        tag=self.tag(),
                        hashes=asdict(self.hash),
                        known_benign=self.known_benign,
                        known_malicious=self.known_malicious
                    )
                elif isinstance(exporter, ClickHouseExporter):
                    # For ClickHouse, we need to pass the table name and the prepared data
                    success &= exporter.export(
                        export_data,
                        table=self.get_clickhouse_table(),
                        # Add these parameters explicitly
                        column_names=export_data[1] if isinstance(export_data, tuple) else None,
                        column_type_names=export_data[2] if isinstance(export_data, tuple) else None
                    )
                
        return success

    @abstractmethod
    def get_clickhouse_table(self) -> str:
        """Return the appropriate ClickHouse table name"""
        pass
    # def export_to_elastic(self, list_of_dataclasses, tag=None):
    #     self.log.debug(inspect.currentframe().f_code.co_name)

    #     if not isinstance(list_of_dataclasses, list):
    #         self.log.error("Called export_to_elastic wrongly")

    #     now_t = datetime.now(timezone.utc).strftime("%Y-%m-%d %H:%M:%S")
    #     for dataclass_ in list_of_dataclasses:
    #         if not settings.ELASTIC_CLIENT.ping():
    #             self.log.error("[CONNECTION ERROR] ping to elastic failed")
            
    #         # Convert dataclass to dict and filter out None values
    #         document = {k: v for k, v in asdict(dataclass_).items() if v is not None}
            
    #         if tag:
    #             document["tag"] = [tag, self.tag()]
    #         else:
    #             document["tag"] = self.tag()
    #         hashes = asdict(self.hash)
    #         document |= hashes
    #         document["timestamp_utc"] = now_t
    #         # document["source"] = self.source
    #         document["known_benign"] = self.known_benign
    #         document["known_malicious"] = self.known_malicious

    #         if "_id" in document:
    #             tmp_id = document.pop("_id") + document["sha256"]
    #             _id = hashlib.sha256(tmp_id.encode()).hexdigest()
    #         else:
    #             _id = document["sha256"]

    #         # self.log.debug(f"[DEBUG] about to export {type(document)} {document}")
    #         try:
    #             doc_dump = json.dumps(document)
    #         except TypeError as e:
    #             self.log.error(
    #                 f"Failed export of document. " f"full document: {document}"
    #             )
    #             raise e

    #         # body={"doc": doc_dump,
    #         #       "doc_as_upsert": True  # Create the document if it doesn't exist
    #         # }

    #         # Check if the index exists, and create it if it doesn't
    #         # if not settings.ELASTIC_CLIENT.indices.exists(index=self.elastic_index):
    #         #     settings.ELASTIC_CLIENT.indices.create(index=self.elastic_index)
    #         # pprint(doc_dump) #DEBUG
    #         # print("[DEBUG] _id: " + _id)
    #         # print("[DEBUG] index: " + self.elastic_index)
    #         settings.ELASTIC_CLIENT.index(
    #             index=self.elastic_index, id=_id, document=doc_dump
    #         )