R. Mukundan

27 papers Misc 2Journal 18Unranked 7
YearRankTypeTitle / Venue / Authors
2019 conf
IEEM
R. Mukundan, Chandrashekhar Chaudhari, Vishwas Dohale, Priya Ambilkar
2012 J jnl
IET Signal Process.
Wissam A. Jassim, P. Raveendran, R. Mukundan
2012 Misc conf
IVCNZ
Philip Buchanan, Michael C. Doggett, R. Mukundan
2009 J jnl
Int. J. Comput. Appl. Technol.
R. Mukundan
2008 conf
TPCG
R. Mukundan, C. Han
2008 J jnl
Int. J. Pattern Recognit. Artif. Intell.
R. Mukundan
2008 J jnl
Pattern Recognit.
Chong-Yaw Wee, Paramesran Raveendran, R. Mukundan
2007 Misc conf
ICIAP
Chong-Yaw Wee, Paramesran Raveendran, R. Mukundan
2006 conf
Advances in Computer Entertainment Technology
Jung Shin, Michael Haller, R. Mukundan
2006 conf
CGIV
Oliver Batchelor, R. Mukundan, Richard D. Green
2004 J jnl
IEEE Trans. Image Process.
R. Mukundan
2004 J jnl
Pattern Recognit.
Chee-Way Chong, P. Raveendran, R. Mukundan
2003 J jnl
Pattern Recognit.
Chee-Way Chong, P. Raveendran, R. Mukundan
2003 J jnl
Int. J. Pattern Recognit. Artif. Intell.
Chee-Way Chong, P. Raveendran, R. Mukundan
2003 conf
CISST
R. Mukundan
2003 J jnl
Pattern Anal. Appl.
Chee-Way Chong, P. Raveendran, R. Mukundan
2003 J jnl
Pattern Recognit.
Chee-Way Chong, P. Raveendran, R. Mukundan
2002 conf
JCIS
Chee-Way Chong, R. Mukundan, P. Raveendran
2001 J jnl
IEEE Trans. Image Process.
R. Mukundan, S. H. Ong, Poh Aun Lee
1999 conf
ITiCSE-WGR
R. Mukundan
1996 J jnl
Pattern Recognit. Lett.
R. Mukundan, K. R. Ramakrishnan
1995 J jnl
Pattern Recognit.
R. Mukundan, K. R. Ramakrishnan
1993 J jnl
Pattern Recognit. Lett.
R. Mukundan, N. K. Malik, K. R. Ramakrishnan
1992 J jnl
CVGIP Graph. Model. Image Process.
R. Mukundan
1990 J jnl
Simul.
R. Mukundan, S. S. Swamy
1987 J jnl
Simul.
R. Mukundan
1986 J jnl
Comput. Graph.
R. Mukundan
redb/extractor_registry.py
← Index redb/extractor_registry.py python
"""
Extractor registry with lazy loading by file type.

Groups extractors by file type (pe, elf, macho, apk) and only imports
the relevant group when that file type is first encountered. This avoids
loading heavy dependencies (pefile, lief, androguard, etc.) into workers
that don't need them.
"""
import importlib
import logging

logger = logging.getLogger(__name__)

# Registry: group name -> list of (module_path, class_names)
_REGISTRY = {
    "pe": [
        ("redb.extractors.pe_extractors", [
            "PEFeaturesExtractor",
            "PEImportExtractor",
            "PEResourceExtractor",
            "PEOverlayExtractor",
            "PESectionExtractor",
            "PESignatureExtractor",
            "PEExtraFindings",
            "PEInconstistencyTestsExtractor",
            "PEDotNetExtractor",
        ]),
    ],
    "elf": [
        ("redb.extractors.elf_extractors", [
            "ELFFeaturesExtractor",
            "ELFSegmentExtractor",
            "ELFSectionExtractor",
            "ELFDependencyExtractor",
            "ELFSymbolExtractor",
            "ELFImportExtractor",
            "ELFExportExtractor",
            "ELFRelocationExtractor",
            "ELFNotesExtractor",
        ]),
    ],
    "macho": [
        ("redb.extractors.macho_extractors", [
            "MachOFeaturesExtractor",
            "MachOSegmentExtractor",
            "MachOImportExtractor",
            "MachOExportExtractor",
            "MachODylibExtractor",
            "MachOSignatureExtractor",
        ]),
    ],
    "apk": [
        ("redb.extractors.apk_extractors", [
            "APKFeaturesExtractor",
            "APKManifestExtractor",
            "APKPermissionsExtractor",
            "APKSignatureExtractor",
            "APKDexExtractor",
            "APKResourceExtractor",
            "APKNativeLibExtractor",
            "APKInconsistencyTestsExtractor",
        ]),
        ("redb.extractors.decompiler.DecompileAPK", [
            "DecompileAPK",
        ]),
    ],
    "js": [
        ("redb.extractors.js_extractors", [
            "JSFeaturesExtractor",
            "JSSuspiciousAPIsExtractor",
            "JSStringsExtractor",
            "JSDeobfuscationExtractor",
            "JSContentExtractor",
        ]),
    ],
}

# Map filetype labels (from Magika) to registry group names
FILETYPE_TO_GROUP = {
    "pebin": "pe",
    "elf": "elf",
    "macho": "macho",
    "apk": "apk",
    "javascript": "js",
}

# Cache: group name -> {class_name: class}
_group_cache = {}


def _load_group(group_name):
    """Import all extractors for a group. Cached after first call."""
    if group_name in _group_cache:
        return _group_cache[group_name]

    logger.debug(f"Loading extractor group: {group_name}")

    # Suppress androguard logging before importing APK extractors
    if group_name == "apk":
        from loguru import logger as loguru_logger
        loguru_logger.disable("androguard")

    classes = {}
    for module_path, class_names in _REGISTRY[group_name]:
        mod = importlib.import_module(module_path)
        for name in class_names:
            classes[name] = getattr(mod, name)

    _group_cache[group_name] = classes
    return classes


def get_filetype_modules(filetype):
    """Get the list of format-specific extractor classes for a filetype.

    Returns only analysis extractors (excludes decompilers like DecompileAPK).
    """
    group = FILETYPE_TO_GROUP.get(filetype)
    if not group:
        return []
    classes = _load_group(group)
    return [cls for name, cls in classes.items() if not name.startswith("Decompile")]


def get_extractor_class(name):
    """Look up a single extractor class by name across all groups.

    Checks cached groups first, then loads groups on demand.
    """
    # Check already-loaded groups first
    for group_classes in _group_cache.values():
        if name in group_classes:
            return group_classes[name]

    # Search all groups (triggers lazy loading)
    for group_name in _REGISTRY:
        classes = _load_group(group_name)
        if name in classes:
            return classes[name]

    return None