Hang Min

18 papers Journal 8Unranked 8
YearRankTypeTitle / Venue / Authors
2025 J jnl
CoRR
Zhen Xuen Brandon Low, Rory Zhang, Hang Min, William Pham, Lucy Vivash, Jasmine Moses, Miranda Lynch, Karina Dorfman, Cassandra Marotta, Shaun Koh, Jacob Bunyamin, Ella Rowsthorn, Alex Jarema, Himashi Peiris, Zhaolin Chen, Sandy R. Shultz, David K. Wright, Dexiao Kong, Sharon L. Naismith, Terence J. O'Brien, Ying Xia, Meng Law, Benjamin Sinclair
2025 J jnl
CoRR
Yilei Wu, Yichi Zhang, Zijian Dong, Fang Ji, An Sen Tan, Gifford Tan, Sizhao Tang, Huijuan Chen, Zijiao Chen, Eric Kwun Kei Ng, José Bernal, Hang Min, Ying Xia, Ines Vati, Liz Cooper, Xiaoyu Hu, Yuchen Pei, Yutao Ma, Victor Nozais, Ami Tsuchida, Pierre-Yves Hervé, Philippe Boutinaud, Marc Joliot, Junghwa Kang, Wooseung Kim, Dayeon Bak, Rachika E. Hamadache, Valeriia Abramova, Xavier Lladó, Yuntao Zhu, Zhenyu Gong, Xin Chen, John McFadden, Pek Lan Khong, Roberto Duarte Coello, Hongwei Bran Li, Woon Puay Koh, Christopher Chen, Joanna M. Wardlaw, Maria del C. Valdés Hernández, Juan Helen Zhou
2024 J jnl
CoRR
Hang Min, Gorane Santamaria Hormaechea, Prabhakar Ramachandran, Jason Dowling
2024 J jnl
Comput. Medical Imaging Graph.
Phillip Chlap, Hang Min, Jason Dowling, Matthew Field, Kirrily Cloak, Trevor Leong, Mark Lee, Julie Chu, Jennifer Tan, Phillip Tran, Tomas Kron, Mark Sidhom, Kirsty Wiltshire, Sarah Keats, Andrew Kneebone, Annette Haworth, Martin A. Ebert, Shalini K. Vinod, Lois Holloway
2023 conf
MedInfo
Parnesh Raniga, Hang Min, Hugo Leroux
2023 J jnl
Neural Comput. Appl.
Guangcun Wei, Yanna Zhang, Hang Min, Yunfei Xu
2022 conf
ICCIR
Chunxiao Zhang, Guangcun Wei, Yunfei Xu, Hang Min
2022 conf
CLEF (Working Notes)
Bowen Xin, Hang Min, Ashley G. Gillman, Bevan Koopman, Jason Dowling, Aaron Nicolson
2022 conf
ICBBT
Hang Min, Hao Zhou, Duo Zheng, Yang Deng, Fujian Wan, Shijie Ding, Jingjiao Zhou, Jun Zhou
2022 conf
ICBBT
Hao Zhou, Hang Min, Changxia Mao, Yang Deng, Xinyue Wei, Jun Zhou
2022 conf
DMBD (2)
Yanna Zhang, Guangcun Wei, Hang Min, Yunfei Xu
2020 J jnl
CoRR
Hang Min, Darryl McClymont, Shekhar S. Chandra, Stuart Crozier, Andrew P. Bradley
2020
Hang Min
2020 conf
ISBI
Hang Min, Devin Wilson, Yinhuang Huang, Siyu Liu, Stuart Crozier, Andrew P. Bradley, Shekhar S. Chandra
2019 J jnl
CoRR
Hang Min, Devin Wilson, Yinhuang Huang, Samuel Kelly, Stuart Crozier, Andrew P. Bradley, Shekhar S. Chandra
2018 ed.
RAMBO+BIA+TIA@MICCAI
Danail Stoyanov, Zeike Taylor, Bernhard Kainz, Gabriel Maicas, Reinhard R. Beichel, Anne L. Martel, Lena Maier-Hein, Kanwal K. Bhatia, Tom Vercauteren, Ozan Oktay, Gustavo Carneiro, Andrew P. Bradley, Jacinto C. Nascimento, Hang Min, Matthew S. Brown, Colin Jacobs, Bianca Lassen-Schmidt, Kensaku Mori, Jens Petersen, Raúl San José Estépar, Alexander Schmidt-Richberg, Catarina Veiga
2017 J jnl
Int. J. Appl. Decis. Sci.
Songqing Yu, Junhai Ma, Hang Min, Xueli Zhan
2017 conf
ISBI
Hang Min, Shekhar S. Chandra, Neeraj Dhungel, Stuart Crozier, Andrew P. Bradley
redb/extractors/js_extractors/js_patterns.py
← Index redb/extractors/js_extractors/js_patterns.py python
"""Canonical, compiled JavaScript regex patterns shared across JS extractors.

All suspicious-API patterns and the few feature-only patterns live here so each
expression is compiled exactly once per Python process and so any pattern that
was previously duplicated across `js_features.py` and `js_suspicious_apis.py`
now resolves to a single shared compiled object.

JavaScript is case-sensitive at runtime, but every suspicious-API pattern matches
either a literal-case identifier (`\\beval\\s*\\(`, `String\\.fromCharCode`, etc.)
or a string-quoted token (`"powershell"`). Compiling them with `re.IGNORECASE`
matches the historical behaviour of `JSSuspiciousAPIsExtractor` and is safe for
the patterns that historically came from `JSFeaturesExtractor` — those literals
are spelled in real-world JS exactly as written.

`scan_source()` is the entry point used by extractors: it walks the source once
per pattern using the pre-compiled regexes and returns a flat
`{name: {"count": N, "lines": [unique_line_numbers_sorted]}}` dict. Both
`JSFeaturesExtractor` and `JSSuspiciousAPIsExtractor` consume the same dict so
the per-pattern × per-line loops they used to run independently collapse to a
single shared scan.
"""

import bisect
import re
from typing import Dict, Iterable, List, Mapping

_FLAGS = re.IGNORECASE

# Canonical compiled patterns, keyed by their human-readable name. The name is
# also the value emitted into `redb_js_suspicious_apis.api_name`.
PATTERNS = {
    # ---- code execution ----
    "eval": re.compile(r"\beval\s*\(", _FLAGS),
    "Function constructor": re.compile(r"\bnew\s+Function\s*\(", _FLAGS),
    "execScript": re.compile(r"\bexecScript\s*\(", _FLAGS),
    "document.write": re.compile(r"\bdocument\.write(?:ln)?\s*\(", _FLAGS),
    "innerHTML assignment": re.compile(r"\.innerHTML\s*=", _FLAGS),
    "outerHTML assignment": re.compile(r"\.outerHTML\s*=", _FLAGS),
    "insertAdjacentHTML": re.compile(r"\.insertAdjacentHTML\s*\(", _FLAGS),
    # ---- network ----
    "XMLHttpRequest": re.compile(r"\bnew\s+XMLHttpRequest\b", _FLAGS),
    "fetch": re.compile(r"\bfetch\s*\(", _FLAGS),
    "WebSocket": re.compile(r"\bnew\s+WebSocket\s*\(", _FLAGS),
    "navigator.sendBeacon": re.compile(r"\bnavigator\.sendBeacon\s*\(", _FLAGS),
    "ActiveXObject XMLHTTP": re.compile(
        r"ActiveXObject\s*\(\s*[\"\'](?:MSXML2\.XMLHTTP|Microsoft\.XMLHTTP)", _FLAGS
    ),
    "require network module": re.compile(
        r"require\s*\(\s*[\"\'](?:http|https|net|dgram)[\"\']", _FLAGS
    ),
    "axios": re.compile(r"\baxios\b", _FLAGS),
    # ---- filesystem ----
    "require fs": re.compile(r"require\s*\(\s*[\"\']fs[\"\']", _FLAGS),
    "require path": re.compile(r"require\s*\(\s*[\"\']path[\"\']", _FLAGS),
    "FileSystemObject": re.compile(r"Scripting\.FileSystemObject", _FLAGS),
    "ADODB.Stream": re.compile(r"ADODB\.Stream", _FLAGS),
    "Shell.Application": re.compile(r"Shell\.Application", _FLAGS),
    "WScript.CreateObject": re.compile(r"WScript\.CreateObject", _FLAGS),
    # ---- process ----
    "require child_process": re.compile(r"require\s*\(\s*[\"\']child_process[\"\']", _FLAGS),
    "child_process exec": re.compile(r"child_process\.(?:exec|spawn|execFile|fork)\s*\(", _FLAGS),
    "WScript.Shell": re.compile(r"WScript\.Shell", _FLAGS),
    "WScript.Shell.Run": re.compile(r"\.Run\s*\(", _FLAGS),
    "WScript.Shell.Exec": re.compile(r"\.Exec\s*\(", _FLAGS),
    "ShellExecute": re.compile(r"\bShellExecute\b", _FLAGS),
    "PowerShell reference": re.compile(r"[\"\']powershell[\"\']", _FLAGS),
    "cmd.exe reference": re.compile(r"[\"\']cmd\.exe[\"\']", _FLAGS),
    "require os": re.compile(r"require\s*\(\s*[\"\']os[\"\']", _FLAGS),
    # ---- registry ----
    "RegRead": re.compile(r"\.RegRead\s*\(", _FLAGS),
    "RegWrite": re.compile(r"\.RegWrite\s*\(", _FLAGS),
    "RegDelete": re.compile(r"\.RegDelete\s*\(", _FLAGS),
    "StdRegProv": re.compile(r"StdRegProv", _FLAGS),
    # ---- crypto / encoding ----
    "atob": re.compile(r"\batob\s*\(", _FLAGS),
    "btoa": re.compile(r"\bbtoa\s*\(", _FLAGS),
    "String.fromCharCode": re.compile(r"String\.fromCharCode\s*\(", _FLAGS),
    "unescape": re.compile(r"\bunescape\s*\(", _FLAGS),
    "decodeURIComponent": re.compile(r"\bdecodeURIComponent\s*\(", _FLAGS),
    "Buffer.from": re.compile(r"Buffer\.from\s*\(", _FLAGS),
    "crypto module": re.compile(r"crypto\.create(?:Cipher|Decipher|Hash|Hmac)", _FLAGS),
    # ---- DOM manipulation ----
    "document.forms": re.compile(r"document\.forms", _FLAGS),
    "document.cookie": re.compile(r"document\.cookie", _FLAGS),
    "querySelector sensitive input": re.compile(
        r"document\.querySelector\s*\([^)]*(?:password|credit|card|cvv|ssn)", _FLAGS
    ),
    "submit event listener": re.compile(r"addEventListener\s*\(\s*[\"\']submit", _FLAGS),
    "createElement script/iframe": re.compile(
        r"\.createElement\s*\(\s*[\"\'](?:script|iframe)", _FLAGS
    ),
    "dynamic script src": re.compile(r"\.src\s*=\s*[\"\'](?:https?://|//)", _FLAGS),
}

# Pattern name -> category (one of code_execution / network / filesystem /
# process / registry / crypto_encoding / dom_manipulation).
CATEGORIES = {
    "eval": "code_execution",
    "Function constructor": "code_execution",
    "execScript": "code_execution",
    "document.write": "code_execution",
    "innerHTML assignment": "code_execution",
    "outerHTML assignment": "code_execution",
    "insertAdjacentHTML": "code_execution",
    "XMLHttpRequest": "network",
    "fetch": "network",
    "WebSocket": "network",
    "navigator.sendBeacon": "network",
    "ActiveXObject XMLHTTP": "network",
    "require network module": "network",
    "axios": "network",
    "require fs": "filesystem",
    "require path": "filesystem",
    "FileSystemObject": "filesystem",
    "ADODB.Stream": "filesystem",
    "Shell.Application": "filesystem",
    "WScript.CreateObject": "filesystem",
    "require child_process": "process",
    "child_process exec": "process",
    "WScript.Shell": "process",
    "WScript.Shell.Run": "process",
    "WScript.Shell.Exec": "process",
    "ShellExecute": "process",
    "PowerShell reference": "process",
    "cmd.exe reference": "process",
    "require os": "process",
    "RegRead": "registry",
    "RegWrite": "registry",
    "RegDelete": "registry",
    "StdRegProv": "registry",
    "atob": "crypto_encoding",
    "btoa": "crypto_encoding",
    "String.fromCharCode": "crypto_encoding",
    "unescape": "crypto_encoding",
    "decodeURIComponent": "crypto_encoding",
    "Buffer.from": "crypto_encoding",
    "crypto module": "crypto_encoding",
    "document.forms": "dom_manipulation",
    "document.cookie": "dom_manipulation",
    "querySelector sensitive input": "dom_manipulation",
    "submit event listener": "dom_manipulation",
    "createElement script/iframe": "dom_manipulation",
    "dynamic script src": "dom_manipulation",
}

# Patterns consumed only by JSFeaturesExtractor (no category, never surfaced as
# a suspicious-API row). Kept here so every JS regex is compiled in one place.
FEATURE_PATTERNS = {
    "hex_escape": re.compile(r"\\x[0-9a-fA-F]{2}"),
    "unicode_escape": re.compile(r"\\u[0-9a-fA-F]{4}"),
    "base64_string": re.compile(r"[A-Za-z0-9+/]{40,}={0,2}"),
    # decodeURI matches BOTH decodeURI and decodeURIComponent. The latter is also
    # a suspicious-API pattern in PATTERNS; this broader form is what the
    # `decodeuri_count` feature column has historically counted.
    "decodeURI": re.compile(r"\b(?:decodeURI|decodeURIComponent)\s*\(", _FLAGS),
    "settimeout_setinterval": re.compile(r"\b(?:setTimeout|setInterval)\s*\(", _FLAGS),
    "function_decl": re.compile(r"\bfunction\s+\w+\s*\(|\bfunction\s*\("),
    "var_decl": re.compile(r"\b(?:var|let|const)\s+"),
    "string_concat": re.compile(r"[\"\'][\s]*\+[\s]*[\"\']"),
    "comment": re.compile(r"//.*?$|/\*[\s\S]*?\*/", re.MULTILINE),
    "long_string": re.compile(r"[\"\']([^\"\']{256,})[\"\']"),
    "array_function_call": re.compile(r"\[(?:0x[0-9a-f]+|[\d]+)\]\s*\(", _FLAGS),
}

# Patterns consumed only by JSStringsExtractor for encoded-string discovery.
# Scoped to *hidden* strings only — patterns whose decoded form is not visible
# to a substring search over the raw text. Plain long literals are not
# extracted here because they're already preserved in code_text_content and
# scraped by the IOC pipeline over text_raw / text_normalized.
#
# Distinct from FEATURE_PATTERNS even where the names rhyme:
#   FEATURE_PATTERNS["hex_escape"] / ["unicode_escape"]   -> single escape
#   STRING_PATTERNS["hex_escape_seq"] / ["unicode_escape_seq"] -> 4+ / 3+ in a row
#   FEATURE_PATTERNS["base64_string"]                     -> bare base64 token
#   STRING_PATTERNS["base64_quoted"]                      -> base64 inside JS quotes
# These do not share match objects with the suspicious-API or feature scans, so
# they are not folded into JSContext.scan; the strings extractor walks them
# itself (one finditer per pattern, with shared line-offset bisect in #4b).
STRING_PATTERNS = {
    "hex_escape_seq": re.compile(r"(?:\\x[0-9a-fA-F]{2}){4,}"),
    "unicode_escape_seq": re.compile(r"(?:\\u[0-9a-fA-F]{4}){3,}"),
    "charcode_call": re.compile(r"String\.fromCharCode\s*\(\s*([\d,\s]+)\s*\)"),
    "base64_quoted": re.compile(r"[\"\']([A-Za-z0-9+/]{40,}={0,2})[\"\']"),
    "concat_chain": re.compile(r"(?:[\"\'][^\"\']+[\"\']\s*\+\s*){3,}[\"\'][^\"\']+[\"\']"),
}


def line_offsets(source: str) -> List[int]:
    """Sorted list of byte offsets for every newline in `source`, plus a final
    sentinel of len(source). Used to translate match offsets into 1-indexed
    line numbers via bisect.
    """
    offsets = [-1]  # so that bisect_right of offset 0 returns line 1
    push = offsets.append
    idx = source.find("\n")
    while idx != -1:
        push(idx)
        idx = source.find("\n", idx + 1)
    return offsets


def _scan_one(
    pattern: "re.Pattern[str]", source: str, offsets: List[int]
) -> Dict[str, object]:
    """Run a single compiled pattern over `source` and return count + unique lines."""
    count = 0
    seen_lines: "set[int]" = set()
    for m in pattern.finditer(source):
        count += 1
        seen_lines.add(bisect.bisect_right(offsets, m.start()))
    if not count:
        return None  # type: ignore[return-value]
    return {"count": count, "lines": sorted(seen_lines)}


def scan_source(
    source: str,
    patterns: Iterable[Mapping[str, "re.Pattern[str]"]] = (PATTERNS, FEATURE_PATTERNS),
) -> Dict[str, Dict[str, object]]:
    """Scan `source` against every compiled pattern in `patterns`.

    Returns a dict keyed by pattern name. Each entry has:
        "count": total number of matches in the source
        "lines": sorted list of unique 1-indexed line numbers where the pattern
                 matched (deduplicated — multiple matches on the same line
                 collapse to one entry, preserving the historical
                 line-set semantics of JSSuspiciousAPIsExtractor)
    Patterns with zero matches are absent from the dict; callers should default
    to {"count": 0, "lines": []}.
    """
    if not source:
        return {}
    offsets = line_offsets(source)
    results: Dict[str, Dict[str, object]] = {}
    for table in patterns:
        for name, pat in table.items():
            entry = _scan_one(pat, source, offsets)
            if entry is not None:
                results[name] = entry
    return results