Manabu Torii

50 papers Misc 3Journal 27Unranked 20
YearRankTypeTitle / Venue / Authors
2023 J jnl
CoRR
Manabu Torii, Ian M. Finn, Son Doan, Paul Wang, Elly W. Yang, Daniel S. Zisook
2019 J jnl
BMC Medical Informatics Decis. Mak.
Son Doan, Elly W. Yang, Sameer S. Tilak, Peter W. Li, Daniel S. Zisook, Manabu Torii
2019 J jnl
CoRR
Son Doan, Elly W. Yang, Sameer Tilak, Manabu Torii
2018 Misc conf
AMIA
Manabu Torii, Elly W. Yang, Son Doan
2018 conf
ICHI Workshops
Son Doan, Elly W. Yang, Sameer Tilak, Manabu Torii
2015 conf
ICHI
Manabu Torii, Jungwei Fan, Daniel S. Zisook
2015 J jnl
IEEE ACM Trans. Comput. Biol. Bioinform.
Manabu Torii, Cecilia N. Arighi, Gang Li, Qinghua Wang, Cathy H. Wu, K. Vijay-Shanker
2015 J jnl
J. Biomed. Informatics
Manabu Torii, Jungwei Fan, Weili Yang, Theodore Lee, Matthew T. Wiley, Daniel S. Zisook, Yang Huang
2014 J jnl
BMC Bioinform.
Yifan Peng, Manabu Torii, Cathy H. Wu, K. Vijay-Shanker
2014 J jnl
J. Biomed. Semant.
Manabu Torii, Kavishwar B. Wagholikar, Hongfang Liu
2014 J jnl
Database J. Biol. Databases Curation
Manabu Torii, Gang Li, Zhiwen Li, Rose Oughtred, Francesca Diella, Irem Çelen, Cecilia N. Arighi, Hongzhan Huang, K. Vijay-Shanker, Cathy H. Wu
2014 J jnl
CoRR
John E. Miller, Michael Bloodgood, Manabu Torii, K. Vijay-Shanker
2014 J jnl
Database J. Biol. Databases Curation
Yifan Peng, Catalina O. Tudor, Manabu Torii, Cathy H. Wu, K. Vijay-Shanker
2013 J jnl
BMC Syst. Biol.
Luis D. Lopez, Jingyi Yu, Cecilia N. Arighi, Catalina O. Tudor, Manabu Torii, Hongzhan Huang, K. Vijay-Shanker, Cathy H. Wu
2013 conf
BCB
Luis D. Lopez, Jingyi Yu, Cecilia N. Arighi, Manabu Torii, K. Vijay-Shanker, Hongzhan Huang, Cathy H. Wu
2013 J jnl
Database J. Biol. Databases Curation
Donald C. Comeau, Rezarta Islamaj Dogan, Paolo Ciccarese, Kevin Bretonnel Cohen, Martin Krallinger, Florian Leitner, Zhiyong Lu, Yifan Peng, Fabio Rinaldi, Manabu Torii, Alfonso Valencia, Karin Verspoor, Thomas C. Wiegers, Cathy H. Wu, W. John Wilbur
2013 J jnl
J. Biomed. Semant.
Kavishwar B. Wagholikar, Manabu Torii, Siddhartha Jonnalagadda, Hongfang Liu
2013 conf
BCB
Manabu Torii, Cecilia N. Arighi, Qinghua Wang, Cathy H. Wu, K. Vijay-Shanker
2012 conf
BIBM
Xia Bi, Hongzhan Huang, Sherri Matis-Mitchell, Peter B. McGarvey, Manabu Torii, Hagit Shatkay, Cathy H. Wu
2012 J jnl
J. Am. Medical Informatics Assoc.
Siddhartha Jonnalagadda, Dingcheng Li, Sunghwan Sohn, Stephen Tze-Inn Wu, Kavishwar B. Wagholikar, Manabu Torii, Hongfang Liu
2012 conf
BIBM Workshops
Carl J. Schmidt, Liang Sun, Cecilia N. Arighi, Keith Decker, K. Vijay-Shanker, Manabu Torii, Catalina O. Tudor, Cathy H. Wu, Peter D'Eustachio
2012 conf
BIBM
Yifan Peng, Catalina O. Tudor, Manabu Torii, Cathy H. Wu, K. Vijay-Shanker
2011 J jnl
Int. J. Medical Informatics
Manabu Torii, Lanlan Yin, Thang Nguyen, Chand T. Mazumdar, Hongfang Liu, David M. Hartley, Noele P. Nelson
2011 conf
BCB
Kavishwar B. Wagholikar, Manabu Torii, Hongfang Liu
2011 conf
BIBM Workshops
Yang Chen, Manabu Torii, Chang-Tien Lu, Hongfang Liu
2011 J jnl
BMC Bioinform.
Zhiyong Lu, Hung-Yu Kao, Chih-Hsuan Wei, Minlie Huang, Jingchen Liu, Cheng-Ju Kuo, Chun-Nan Hsu, Richard Tzong-Han Tsai, Hong-Jie Dai, Naoaki Okazaki, Han-Cheol Cho, Martin Gerner, Illés Solt, Shashank Agarwal, Feifan Liu, Dina Vishnyakova, Patrick Ruch, Martin Romacker, Fabio Rinaldi, Sanmitra Bhattacharya, Padmini Srinivasan, Hongfang Liu, Manabu Torii, Sérgio Matos, David Campos, Karin Verspoor, Kevin M. Livingston, W. John Wilbur
2011 J jnl
J. Am. Medical Informatics Assoc.
Manabu Torii, Kavishwar B. Wagholikar, Hongfang Liu
2011 J jnl
BMC Bioinform.
Jinlian Wang, Manabu Torii, Hongfang Liu, Gerald W. Hart, Zhang-Zhi Hu
2010 conf
IHI
Manabu Torii, Burt-Ujin Bayarsaikhan, Hongfang Liu, Thang Nguyen, Kevin Jones, Noele P. Nelson, David M. Hartley
2010 J jnl
Int. J. Comput. Model. Algorithms Medicine
Hongfang Liu, Manabu Torii, Guixian Xu, Johannes Goll
2010 J jnl
Artif. Intell. Medicine
Lanlan Yin, Guixian Xu, Manabu Torii, Zhendong Niu, José M. Maisog, Cathy H. Wu, Zhang-Zhi Hu, Hongfang Liu
2010 conf
Semantic Mining in Biomedicine
Chong Min Lee, Manabu Torii, Zhang-Zhi Hu, Yi-Ting Tsai, Jinesh Shah, Hongfang Liu
2009 J jnl
J. Am. Medical Informatics Assoc.
Manabu Torii, Zhang-Zhi Hu, Cathy H. Wu, Hongfang Liu
2009 Misc conf
AMIA
Manabu Torii, Hongfang Liu, Zhang-Zhi Hu
2008 conf
BIBM
Guixian Xu, Lanlan Yin, Manabu Torii, Zhendong Niu, Cathy H. Wu, Zhang-Zhi Hu, Hongfang Liu
2008 conf
BioLINK@ISMB/ECCB
Hongfang Liu, Manabu Torii, Guixian Xu, Zhang-Zhi Hu, Johannes Goll
2007 J jnl
BMC Bioinform.
Manabu Torii, Zhang-Zhi Hu, Min Song, Cathy H. Wu, Hongfang Liu
2007 conf
BioNLP@ACL
John E. Miller, Manabu Torii, K. Vijay-Shanker
2007 conf
EMNLP-CoNLL
John E. Miller, Manabu Torii, K. Vijay-Shanker
2007 conf
LBM (Short Papers)
Manabu Torii, Hongfang Liu
2007 Misc conf
AMIA
Manabu Torii, Hongfang Liu
2007 conf
ICDM Workshops
Hongfang Liu, Manabu Torii, Zhang-Zhi Hu, Cathy H. Wu
2007 J jnl
Comput. Intell.
Manabu Torii, K. Vijay-Shanker
2006 J jnl
Bioinform.
X. Yuan, Zhang-Zhi Hu, H. T. Wu, Manabu Torii, Meenakshi Narayanaswamy, K. E. Ravikumar, K. Vijay-Shanker, Cathy H. Wu
2006 conf
KDLL
Manabu Torii, Hongfang Liu
2006 conf
BioNLP@NAACL-HLT
John E. Miller, Michael Bloodgood, Manabu Torii, K. Vijay-Shanker
2006 J jnl
J. Am. Medical Informatics Assoc.
Hongfang Liu, Zhang-Zhi Hu, Manabu Torii, Cathy H. Wu, Carol Friedman
2004 J jnl
J. Biomed. Informatics
Manabu Torii, Sachin Kamboj, K. Vijay-Shanker
2003 conf
BioNLP@ACL
Manabu Torii, Sachin Kamboj, K. Vijay-Shanker
2002 J jnl
IEEE Trans. Neural Networks
Manabu Torii, Martin T. Hagan
redb/extractors/hashes.py
← Index redb/extractors/hashes.py python
from dataclasses import asdict
from typing import Any
from datetime import datetime, timezone

import hashlib
import inspect
from struct import pack

from magika import Magika
from signify.fingerprinter import AuthenticodeFingerprinter
import ppdeep
import tlsh

import pefile
from elftools.elf.elffile import ELFFile
from elftools.common.exceptions import ELFError

from redb.extractors.enum import Tag
from redb.models.dataclasses import Hashes
from redb.extractors.extractor import Extractor


class HashExtractor(Extractor):

    def __init__(
        self,
        filepath,
        log,
        exporters=None,
        index_prefix=None,
        elastic_index=None,
        known_benign=False,
        known_malicious=False,
        macho=None,
    ):
        super().__init__(
            filepath, log, exporters, index_prefix, elastic_index, known_benign, known_malicious
        )
        self.hashes = None
        self.elastic_index = self.index_prefix + "-hashes"
        self.filetype = Magika().identify_bytes(self.binary).output.label
        self.pe = None
        self.elf = None
        self.macho = macho
        if self.filetype == "pebin":  # else None
            try:
                self.pe = pefile.PE(self.filepath)
            except Exception as e:
                self.log.error(f"Failed to initialize PE file object: {e}")
                self.pe = None
        elif self.filetype == "elf":
            # ELF file will be created when needed in _extract_elf_hashes
            pass

    def _extract_hashes(self):
        self.log.debug(inspect.currentframe().f_code.co_name)
        
        # Initialize hash values with None
        md5 = None
        sha1 = None
        sha256 = None
        ssdeep_hash = None
        tlsh_hash = None
        
        # Calculate basic hashes with error handling
        try:
            md5 = hashlib.md5(self.binary).hexdigest()
        except Exception as e:
            self.log.error(f"Failed to calculate MD5 hash: {e}")
            
        try:
            sha1 = hashlib.sha1(self.binary).hexdigest()
        except Exception as e:
            self.log.error(f"Failed to calculate SHA1 hash: {e}")
            
        try:
            sha256 = hashlib.sha256(self.binary).hexdigest()
        except Exception as e:
            self.log.error(f"Failed to calculate SHA256 hash: {e}")
            
        try:
            ssdeep_hash = ppdeep.hash_from_file(self.filepath)
        except Exception as e:
            self.log.error(f"Failed to calculate ssdeep hash: {e}")
            
        try:
            tlsh_hash = tlsh.hash(self.binary)
        except Exception as e:
            self.log.error(f"Failed to calculate TLSH hash: {e}")
        
        # Create Hashes object with available values
        self.hashes = Hashes(
            md5 or "",
            sha1 or "",
            sha256 or "",
            ssdeep_hash or "",
            tlsh_hash or "",
        )

        if self.filetype == "pebin" and self.pe is not None:
            self.log.debug("Computing PEBIN related hash values")
            
            # Authentihash
            try:
                with open(self.filepath, 'rb') as f:
                    fingerprinter = AuthenticodeFingerprinter(f)
                    fingerprinter.add_authenticode_hashers(hashlib.sha256)
                    self.hashes.authentihash = fingerprinter.hash()['sha256'].hex()
            except Exception as e:
                self.log.error(f"Failed to calculate Authentihash: {e}")
                self.hashes.authentihash = None
                
            # Imphash
            try:
                self.hashes.imphash = self.pe.get_imphash()
            except Exception as e:
                self.log.error(f"Failed to calculate Imphash: {e}")
                self.hashes.imphash = None
                
            # Rich header hashes
            try:
                if self.pe.parse_rich_header():
                    self.log.debug("Computing RichHeader related hash values")
                    
                    richhash = self._compute_richhash()
                    if richhash:
                        self.hashes.richhash = richhash
                        
                    richpe_hash = self._compute_richpe_hash()
                    if richpe_hash:
                        self.hashes.richpe_hash = richpe_hash
                        
                    richpv_result = self._compute_richpv_hash()
                    if richpv_result:
                        self.hashes.richpv_hash, self.hashes.richpv_hash_sorted = richpv_result
            except Exception as e:
                self.log.error(f"Failed to calculate Rich header hashes: {e}")
        elif self.filetype == "pebin" and self.pe is None:
            self.log.warning("PE file type detected but PE object initialization failed - skipping PE-specific hashes")

        elif self.filetype == "elf":
            self.log.debug("Computing ELF related hash values")
            self._extract_elf_hashes()
        elif self.filetype == "macho":
            self.log.debug("Computing Mach-O related hash values")
            self._extract_macho_hashes()
        elif self.filetype == "apk":
            self.log.debug("Computing APK related hash values")
            self._extract_apk_hashes()
        else:
            pass
        self.log.debug(f"Hashes dump: {asdict(self.hashes)}")
        return self.hashes

    def _compute_richhash(self):
        self.log.debug(inspect.currentframe().f_code.co_name)
        try:
            rich_header = self.pe.parse_rich_header()
            if not rich_header:
                return ""
            data = rich_header["clear_data"]
            return hashlib.md5(data).hexdigest().lower()
        except:
            return ""

    def _compute_richpe_hash(self):
        """
        Computes the RichPE hash given a file path or PE object.

        RichPE hash is includes RichHeader CompID and count, as well as
        fields from IMAGE_FILE_HEADER and IMAGE_OPTIONAL_HEADER

        Parameters:
        input: it can be either a file path or a PE object

        Returns:
        richpe_hash: md5 hash of the RichPE value
        None: if no Rich Header present
        """
        try:
            # Attempt to parse Rich header
            self.log.debug(inspect.currentframe().f_code.co_name)
            rich_header = self.pe.parse_rich_header()
            if rich_header is None:
                return None

            # Get list of @Comp.IDs and counts from Rich header
            # Elements in rich_fields at even indices are @Comp.IDs
            # Elements in rich_fields at odd indices are counts
            rich_fields = rich_header.get("values", None)
            if not rich_fields or len(rich_fields) % 2 != 0:
                return None

            md5 = hashlib.md5()

            # Update hash using @Comp.IDs and masked counts from Rich header
            while len(rich_fields):
                compid = rich_fields.pop(0)
                count = rich_fields.pop(0)
                mask = 2 ** (count.bit_length() // 2 + 1) - 1
                count |= mask
                md5.update(pack("<L", compid))
                md5.update(pack("<L", count))

            # Update hash using metadata from the PE header
            md5.update(pack("<L", self.pe.FILE_HEADER.Machine))
            md5.update(pack("<L", self.pe.FILE_HEADER.Characteristics))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.Subsystem))
            md5.update(pack("<B", self.pe.OPTIONAL_HEADER.MajorLinkerVersion))
            md5.update(pack("<B", self.pe.OPTIONAL_HEADER.MinorLinkerVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MajorOperatingSystemVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MinorOperatingSystemVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MajorImageVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MinorImageVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MajorSubsystemVersion))
            md5.update(pack("<L", self.pe.OPTIONAL_HEADER.MinorSubsystemVersion))

            return md5.hexdigest()
        except Exception as e:
            self.log.error(f"Failed to compute RichPE hash: {e}")
            return None

    def _compute_richpv_hash(self):
        """
        Compute the RichPV hash values, sorted and unsorted.
        RichPV excludes the most volatile Rich Header field from the MD5 input data,
        the Product Count (pC) field.

        Returns:
        richpv_hash_unsorted: md5 hash of the RichPV value unsorted
        richpv_hash_sorted: md5 hash of the RichPV value sorted
        None: if no Rich Header present
        """
        try:
            self.log.debug(inspect.currentframe().f_code.co_name)
            # Attempt to parse Rich header
            rich_header = self.pe.parse_rich_header()
            if rich_header is None:
                return None

            # Get list of @Comp.IDs and counts from Rich header
            # Elements in rich_fields at even indices are @Comp.IDs
            # Elements in rich_fields at odd indices are counts
            rich_fields = rich_header.get("values", None)
            if not rich_fields or len(rich_fields) % 2 != 0:
                return None

            md5 = hashlib.md5()
            md5_sorted = hashlib.md5()
            sorted_vector = []

            # Update hash using @Comp.IDs only
            for i in range(0, len(rich_fields), 2):
                compid = rich_fields[i]
                md5.update(pack("<L", compid))
                sorted_vector.append(compid)

            sorted_vector.sort()
            for compid in sorted_vector:
                md5_sorted.update(pack("<L", compid))

            return md5.hexdigest(), md5_sorted.hexdigest()
        except Exception as e:
            self.log.error(f"Failed to compute RichPV hash: {e}")
            return None

    def _extract_elf_hashes(self):
        """Extract ELF specific similarity hashes."""
        self.log.debug(inspect.currentframe().f_code.co_name)

        try:
            with open(self.filepath, 'rb') as f:
                elf = ELFFile(f)
                if not elf:
                    return

                # Generate all similarity hashes
                import_hash = self._generate_elf_import_hash(elf)
                export_hash = self._generate_elf_export_hash(elf)
                section_hash = self._generate_elf_section_hash(elf)
                symbol_hash = self._generate_elf_symbol_hash(elf)
                dynamic_hash = self._generate_elf_dynamic_hash(elf)

                # Only set hashes if they have meaningful values
                if import_hash:
                    self.hashes.import_hash = import_hash
                if export_hash:
                    self.hashes.export_hash = export_hash
                if section_hash:
                    self.hashes.section_hash = section_hash
                if symbol_hash:
                    self.hashes.symhash = symbol_hash
                if dynamic_hash:
                    self.hashes.dynamic_hash = dynamic_hash

        except Exception as e:
            self.log.error(f"Failed to extract ELF hashes: {e}")

    def _generate_elf_import_hash(self, elf) -> str:
        """Generate MD5 hash of sorted, deduplicated imported symbol names."""
        try:
            imported_symbols = set()

            # Get dynamic symbol table
            dynsym_section = elf.get_section_by_name('.dynsym')
            if dynsym_section and hasattr(dynsym_section, 'iter_symbols'):
                for symbol in dynsym_section.iter_symbols():
                    # Look for undefined symbols (imports)
                    if (symbol.entry.get('st_shndx', 0) == 'SHN_UNDEF' and
                        symbol.name and
                        symbol.entry.get('st_info', {}).get('bind') in ['STB_GLOBAL', 'STB_WEAK']):
                        imported_symbols.add(symbol.name)

            # Also check relocations for additional imports
            for section in elf.iter_sections():
                if hasattr(section, 'iter_relocations'):
                    try:
                        for relocation in section.iter_relocations():
                            if hasattr(relocation, 'symbol') and relocation.symbol and relocation.symbol.name:
                                imported_symbols.add(relocation.symbol.name)
                    except:
                        pass

            # Sort and concatenate
            sorted_imports = sorted(list(imported_symbols))

            # Return None if no imports found
            if not sorted_imports:
                return None

            imports_string = '|'.join(sorted_imports)

            # Generate MD5 hash
            return hashlib.md5(imports_string.encode('utf-8')).hexdigest()

        except Exception as e:
            self.log.error(f"Error generating ELF import hash: {e}")
            return None

    def _generate_elf_export_hash(self, elf) -> str:
        """Generate MD5 hash of sorted, deduplicated exported symbol names."""
        try:
            exported_symbols = set()

            # Check both static and dynamic symbol tables
            symbol_sections = ['.symtab', '.dynsym']

            for section_name in symbol_sections:
                section = elf.get_section_by_name(section_name)
                if not section or not hasattr(section, 'iter_symbols'):
                    continue

                for symbol in section.iter_symbols():
                    # Check if symbol is exported (defined and globally visible)
                    if (symbol.name and
                        symbol.entry.get('st_shndx', 0) != 'SHN_UNDEF' and
                        symbol.entry.get('st_info', {}).get('bind') in ['STB_GLOBAL', 'STB_WEAK'] and
                        symbol.entry.get('st_info', {}).get('type') in ['STT_FUNC', 'STT_OBJECT']):
                        exported_symbols.add(symbol.name)

            # Sort and concatenate
            sorted_exports = sorted(list(exported_symbols))

            # Return None if no exports found
            if not sorted_exports:
                return None

            exports_string = '|'.join(sorted_exports)

            # Generate MD5 hash
            return hashlib.md5(exports_string.encode('utf-8')).hexdigest()

        except Exception as e:
            self.log.error(f"Error generating ELF export hash: {e}")
            return None

    def _generate_elf_section_hash(self, elf) -> str:
        """Generate MD5 hash of section layout (names, types, flags sequence)."""
        try:
            section_info = []

            for section in elf.iter_sections():
                header = section.header
                section_name = section.name or "<unnamed>"
                section_type = header.get('sh_type', 'SHT_NULL')
                section_flags = header.get('sh_flags', 0)

                # Create a consistent representation
                section_repr = f"{section_name}:{section_type}:{section_flags}"
                section_info.append(section_repr)

            # Return None if no meaningful sections found
            if not section_info:
                return None

            # Join all section info
            sections_string = '|'.join(section_info)

            # Generate MD5 hash
            return hashlib.md5(sections_string.encode('utf-8')).hexdigest()

        except Exception as e:
            self.log.error(f"Error generating ELF section hash: {e}")
            return None

    def _generate_elf_symbol_hash(self, elf) -> str:
        """Generate MD5 hash of sorted symbol names and types."""
        try:
            symbol_info = set()

            # Check both static and dynamic symbol tables
            symbol_sections = ['.symtab', '.dynsym']

            for section_name in symbol_sections:
                section = elf.get_section_by_name(section_name)
                if not section or not hasattr(section, 'iter_symbols'):
                    continue

                for symbol in section.iter_symbols():
                    if symbol.name:
                        symbol_type = symbol.entry.get('st_info', {}).get('type', 'STT_NOTYPE')
                        symbol_bind = symbol.entry.get('st_info', {}).get('bind', 'STB_LOCAL')

                        # Create a consistent representation
                        symbol_repr = f"{symbol.name}:{symbol_type}:{symbol_bind}"
                        symbol_info.add(symbol_repr)

            # Sort and concatenate
            sorted_symbols = sorted(list(symbol_info))

            # Return None if no symbols found
            if not sorted_symbols:
                return None

            symbols_string = '|'.join(sorted_symbols)

            # Generate MD5 hash
            return hashlib.md5(symbols_string.encode('utf-8')).hexdigest()

        except Exception as e:
            self.log.error(f"Error generating ELF symbol hash: {e}")
            return None

    def _generate_elf_dynamic_hash(self, elf) -> str:
        """Generate MD5 hash of dynamic section entries (DT_* tags)."""
        try:
            dynamic_info = []

            # Get the dynamic section
            dynamic_section = elf.get_section_by_name('.dynamic')
            if not dynamic_section:
                return None

            # Extract dynamic tags and their values
            for tag in dynamic_section.iter_tags():
                dt_tag = tag.entry.d_tag

                # Create a representation based on tag type
                if dt_tag == 'DT_NEEDED':
                    dynamic_info.append(f"DT_NEEDED:{tag.needed}")
                elif dt_tag == 'DT_SONAME':
                    dynamic_info.append(f"DT_SONAME:{tag.soname}")
                elif dt_tag == 'DT_RPATH':
                    dynamic_info.append(f"DT_RPATH:{tag.rpath}")
                elif dt_tag == 'DT_RUNPATH':
                    dynamic_info.append(f"DT_RUNPATH:{tag.runpath}")
                else:
                    # For other tags, use the tag name and value
                    dt_value = tag.entry.d_val if hasattr(tag.entry, 'd_val') else 0
                    dynamic_info.append(f"{dt_tag}:{dt_value}")

            # Sort to ensure consistent ordering
            dynamic_info.sort()
            dynamics_string = '|'.join(dynamic_info)

            # Generate MD5 hash
            return hashlib.md5(dynamics_string.encode('utf-8')).hexdigest()

        except Exception as e:
            self.log.error(f"Error generating ELF dynamic hash: {e}")
            return None

    def _extract_macho_hashes(self):
        """Extract Mach-O specific similarity hashes using machofile API.

        For FAT binaries, this extracts hashes for the current file being processed
        (either the FAT container or an individual slice). The machofile library
        handles the architecture-specific extraction when an arch parameter is provided.

        For slices: We parse the slice file directly since it's a standalone Mach-O.
        For FAT container: We use the provided macho object with combined/fat hashes.
        """
        self.log.debug(inspect.currentframe().f_code.co_name)

        try:
            # If we have a pre-parsed macho object (FAT container or single-arch with passed object)
            if self.macho:
                architectures = self.macho.get_architectures()
                is_fat = len(architectures) > 1

                if is_fat:
                    # For FAT container, get combined hashes (key may be 'fat' or 'combined')
                    all_hashes = self.macho.get_similarity_hashes()
                    if all_hashes:
                        # Try 'fat' first, then 'combined' for backwards compatibility
                        similarity_hashes = all_hashes.get('fat', all_hashes.get('combined', {}))
                    else:
                        similarity_hashes = {}
                else:
                    # Single-arch with pre-parsed object
                    similarity_hashes = self.macho.get_similarity_hashes(arch=architectures[0]) if architectures else {}
            else:
                # No pre-parsed object - parse the file (slice case)
                import machofile
                macho = machofile.UniversalMachO(self.filepath)
                macho.parse()

                architectures = macho.get_architectures()
                if architectures:
                    # For a slice, there's only one architecture
                    similarity_hashes = macho.get_similarity_hashes(arch=architectures[0]) or {}
                else:
                    similarity_hashes = {}

            # Set the hash values on the Hashes object
            if similarity_hashes:
                if similarity_hashes.get('dylib_hash'):
                    self.hashes.macho_dylib_hash = similarity_hashes['dylib_hash']
                if similarity_hashes.get('import_hash'):
                    self.hashes.macho_import_hash = similarity_hashes['import_hash']
                if similarity_hashes.get('export_hash'):
                    self.hashes.macho_export_hash = similarity_hashes['export_hash']
                if similarity_hashes.get('entitlement_hash'):
                    self.hashes.macho_entitlement_hash = similarity_hashes['entitlement_hash']
                if similarity_hashes.get('symhash'):
                    self.hashes.macho_symhash = similarity_hashes['symhash']

        except Exception as e:
            self.log.error(f"Failed to extract Mach-O hashes: {e}")

    def _extract_apk_hashes(self):
        """Extract APK specific similarity hashes (permhash)."""
        self.log.debug(inspect.currentframe().f_code.co_name)
        try:
            from permhash.functions import permhash_apk
            ph = permhash_apk(self.filepath)
            if ph:
                self.hashes.permhash = ph
        except Exception as e:
            self.log.error(f"Failed to extract APK hashes: {e}")

    def extract(self):
        try:
            self.log.debug(inspect.currentframe().f_code.co_name)
            self._extract_hashes()
            return self.hashes
        except Exception as e:
            self.log.error(f"Error extracting hashes: {e}")
            return None

    def prepare_export_data(self, exporter_type: str) -> Any:
        if exporter_type == "ElasticsearchExporter":
            return self.hashes
        elif exporter_type == "ClickHouseExporter":
            # Safely get hash values with fallbacks for None values
            sha256 = getattr(self.hashes, 'sha256', None) or ""
            md5 = getattr(self.hashes, 'md5', None) or ""
            sha1 = getattr(self.hashes, 'sha1', None) or ""
            ssdeep_hash = getattr(self.hashes, 'ssdeep_hash', None) or ""
            tlsh_hash = getattr(self.hashes, 'tlsh_hash', None) or ""
            authentihash = getattr(self.hashes, 'authentihash', None)
            imphash = getattr(self.hashes, 'imphash', None)
            impfuzzy = getattr(self.hashes, 'impfuzzy', None)
            typerefhash = getattr(self.hashes, 'typerefhash', None)
            richhash = getattr(self.hashes, 'richhash', None)
            richpe_hash = getattr(self.hashes, 'richpe_hash', None)
            richpv_hash = getattr(self.hashes, 'richpv_hash', None)
            richpv_hash_sorted = getattr(self.hashes, 'richpv_hash_sorted', None)
            # ELF hashes
            import_hash = getattr(self.hashes, 'import_hash', None)
            export_hash = getattr(self.hashes, 'export_hash', None)
            section_hash = getattr(self.hashes, 'section_hash', None)
            symbol_hash = getattr(self.hashes, 'symhash', None)
            dynamic_hash = getattr(self.hashes, 'dynamic_hash', None)
            # Mach-O hashes
            macho_dylib_hash = getattr(self.hashes, 'macho_dylib_hash', None)
            macho_import_hash = getattr(self.hashes, 'macho_import_hash', None)
            macho_export_hash = getattr(self.hashes, 'macho_export_hash', None)
            macho_entitlement_hash = getattr(self.hashes, 'macho_entitlement_hash', None)
            macho_symhash = getattr(self.hashes, 'macho_symhash', None)
            # APK hashes
            permhash = getattr(self.hashes, 'permhash', None)

            data = [[
                sha256,
                md5,
                sha1,
                ssdeep_hash,
                tlsh_hash,
                authentihash,
                imphash,
                impfuzzy,
                typerefhash,
                richhash,
                richpe_hash,
                richpv_hash,
                richpv_hash_sorted,
                import_hash,
                export_hash,
                section_hash,
                symbol_hash,
                dynamic_hash,
                macho_dylib_hash,
                macho_import_hash,
                macho_export_hash,
                macho_entitlement_hash,
                macho_symhash,
                permhash,
                datetime.now(timezone.utc)
            ]]

            column_names = [
                'sha256', 'md5', 'sha1', 'ssdeep_hash', 'tlsh_hash',
                'authentihash', 'imphash', 'impfuzzy', 'typerefhash',
                'richhash', 'richpe_hash', 'richpv_hash', 'richpv_hash_sorted',
                'import_hash', 'export_hash', 'section_hash', 'symbol_hash', 'dynamic_hash',
                'macho_dylib_hash', 'macho_import_hash', 'macho_export_hash',
                'macho_entitlement_hash', 'macho_symhash',
                'permhash',
                'analysis_date'
            ]

            column_type_names = [
                # sha256, md5, sha1
                'String', 'String', 'String',
                # ssdeep_hash, tlsh_hash
                'Nullable(String)', 'Nullable(String)',
                # authentihash, imphash, impfuzzy, typerefhash
                'Nullable(String)', 'Nullable(String)',
                'Nullable(String)', 'Nullable(String)',
                # richhash, richpe_hash, richpv_hash, richpv_hash_sorted
                'Nullable(String)', 'Nullable(String)',
                'Nullable(String)', 'Nullable(String)',
                # import_hash, export_hash, section_hash, symbol_hash, dynamic_hash
                'Nullable(FixedString(32))', 'Nullable(FixedString(32))',
                'Nullable(FixedString(32))', 'Nullable(FixedString(32))',
                'Nullable(FixedString(32))',
                # macho_dylib_hash, macho_import_hash, macho_export_hash, macho_entitlement_hash, macho_symhash
                'Nullable(FixedString(32))', 'Nullable(FixedString(32))',
                'Nullable(FixedString(32))', 'Nullable(FixedString(32))',
                'Nullable(FixedString(32))',
                # permhash (APK)
                'Nullable(FixedString(64))',
                # analysis_date
                'DateTime64(3, \'UTC\')'
            ]

            return (data, column_names, column_type_names)

    def get_clickhouse_table(self) -> str:
        return "redb_hashes"

    def tag(self):
        return Tag.HASH.value