Ian Magnusson

23 papers A* 4Journal 14Unranked 5
YearRankTypeTitle / Venue / Authors
2025 A* conf
ICML
Ian Magnusson, Nguyen Tai, Ben Bogin, David Heineman, Jena D. Hwang, Luca Soldaini, Akshita Bhagia, Jiacheng Liu, Dirk Groeneveld, Oyvind Tafjord, Noah A. Smith, Pang Wei Koh, Jesse Dodge
2025 J jnl
CoRR
Ian Magnusson, Nguyen Tai, Ben Bogin, David Heineman, Jena D. Hwang, Luca Soldaini, Akshita Bhagia, Jiacheng Liu, Dirk Groeneveld, Oyvind Tafjord, Noah A. Smith, Pang Wei Koh, Jesse Dodge
2025 J jnl
CoRR
Valentin Hofmann, David Heineman, Ian Magnusson, Kyle Lo, Jesse Dodge, Maarten Sap, Pang Wei Koh, Chun Wang, Hannaneh Hajishirzi, Noah A. Smith
2025 J jnl
CoRR
Allyson Ettinger, Amanda Bertsch, Bailey Kuehl, David Graham, David Heineman, Dirk Groeneveld, Faeze Brahman, Finbarr Timbers, Hamish Ivison, Jacob Morrison, Jake Poznanski, Kyle Lo, Luca Soldaini, Matt Jordan, Mayee F. Chen, Michael Noukhovitch, Nathan Lambert, Pete Walsh, Pradeep Dasigi, Robert Berry, Saumya Malik, Saurabh Shah, Scott Geng, Shane Arora, Shashank Gupta, Taira Anderson, Teng Xiao, Tyler Murray, Tyler Romero, Victoria Graf, Akari Asai, Akshita Bhagia, Alexander Wettig, Alisa Liu, Aman Rangapur, Chloe Anastasiades, Costa Huang, Dustin Schwenk, Harsh Trivedi, Ian Magnusson, Jaron Lochner, Jiacheng Liu, Lester James V. Miranda, Maarten Sap, Malia Morgan, Michael Schmitz, Michal Guerquin, Michael Wilson, Regan Huff, Ronan Le Bras, Rui Xin, Rulin Shao, Sam Skjonsberg, Shannon Zejiang Shen, Shuyue Stella Li, Tucker Wilde, Valentina Pyatkin, William Merrill, Yapei Chang, Yuling Gu, Zhiyuan Zeng, Ashish Sabharwal, Luke Zettlemoyer, Pang Wei Koh, Ali Farhadi, Noah A. Smith, Hannaneh Hajishirzi
2025 J jnl
CoRR
David Heineman, Valentin Hofmann, Ian Magnusson, Yuling Gu, Noah A. Smith, Hannaneh Hajishirzi, Kyle Lo, Jesse Dodge
2024 conf
ACL (1)
Luca Soldaini, Rodney Kinney, Akshita Bhagia, Dustin Schwenk, David Atkinson, Russell Authur, Ben Bogin, Khyathi Raghavi Chandu, Jennifer Dumas, Yanai Elazar, Valentin Hofmann, Ananya Harsh Jha, Sachin Kumar, Li Lucy, Xinxi Lyu, Nathan Lambert, Ian Magnusson, Jacob Morrison, Niklas Muennighoff, Aakanksha Naik, Crystal Nam, Matthew E. Peters, Abhilasha Ravichander, Kyle Richardson, Zejiang Shen, Emma Strubell, Nishant Subramani, Oyvind Tafjord, Pete Walsh, Luke Zettlemoyer, Noah A. Smith, Hannaneh Hajishirzi, Iz Beltagy, Dirk Groeneveld, Jesse Dodge, Kyle Lo
2024 J jnl
CoRR
Luca Soldaini, Rodney Kinney, Akshita Bhagia, Dustin Schwenk, David Atkinson, Russell Authur, Ben Bogin, Khyathi Raghavi Chandu, Jennifer Dumas, Yanai Elazar, Valentin Hofmann, Ananya Harsh Jha, Sachin Kumar, Li Lucy, Xinxi Lyu, Nathan Lambert, Ian Magnusson, Jacob Morrison, Niklas Muennighoff, Aakanksha Naik, Crystal Nam, Matthew E. Peters, Abhilasha Ravichander, Kyle Richardson, Zejiang Shen, Emma Strubell, Nishant Subramani, Oyvind Tafjord, Pete Walsh, Luke Zettlemoyer, Noah A. Smith, Hannaneh Hajishirzi, Iz Beltagy, Dirk Groeneveld, Jesse Dodge, Kyle Lo
2024 conf
ACL (1)
Dirk Groeneveld, Iz Beltagy, Evan Pete Walsh, Akshita Bhagia, Rodney Kinney, Oyvind Tafjord, Ananya Harsh Jha, Hamish Ivison, Ian Magnusson, Yizhong Wang, Shane Arora, David Atkinson, Russell Authur, Khyathi Raghavi Chandu, Arman Cohan, Jennifer Dumas, Yanai Elazar, Yuling Gu, Jack Hessel, Tushar Khot, William Merrill, Jacob Morrison, Niklas Muennighoff, Aakanksha Naik, Crystal Nam, Matthew E. Peters, Valentina Pyatkin, Abhilasha Ravichander, Dustin Schwenk, Saurabh Shah, Will Smith, Emma Strubell, Nishant Subramani, Mitchell Wortsman, Pradeep Dasigi, Nathan Lambert, Kyle Richardson, Luke Zettlemoyer, Jesse Dodge, Kyle Lo, Luca Soldaini, Noah A. Smith, Hannaneh Hajishirzi
2024 J jnl
CoRR
Dirk Groeneveld, Iz Beltagy, Pete Walsh, Akshita Bhagia, Rodney Kinney, Oyvind Tafjord, Ananya Harsh Jha, Hamish Ivison, Ian Magnusson, Yizhong Wang, Shane Arora, David Atkinson, Russell Authur, Khyathi Raghavi Chandu, Arman Cohan, Jennifer Dumas, Yanai Elazar, Yuling Gu, Jack Hessel, Tushar Khot, William Merrill, Jacob Morrison, Niklas Muennighoff, Aakanksha Naik, Crystal Nam, Matthew E. Peters, Valentina Pyatkin, Abhilasha Ravichander, Dustin Schwenk, Saurabh Shah, Will Smith, Emma Strubell, Nishant Subramani, Mitchell Wortsman, Pradeep Dasigi, Nathan Lambert, Kyle Richardson, Luke Zettlemoyer, Jesse Dodge, Kyle Lo, Luca Soldaini, Noah A. Smith, Hannaneh Hajishirzi
2024 A* conf
NeurIPS
Ian Magnusson, Akshita Bhagia, Valentin Hofmann, Luca Soldaini, Ananya Harsh Jha, Oyvind Tafjord, Dustin Schwenk, Evan Pete Walsh, Yanai Elazar, Kyle Lo, Dirk Groeneveld, Iz Beltagy, Hanna Hajishirzi, Noah A. Smith, Kyle Richardson, Jesse Dodge
2024 A* conf
EMNLP
Clara Na, Ian Magnusson, Ananya Harsh Jha, Tom Sherborne, Emma Strubell, Jesse Dodge, Pradeep Dasigi
2024 J jnl
CoRR
Clara Na, Ian Magnusson, Ananya Harsh Jha, Tom Sherborne, Emma Strubell, Jesse Dodge, Pradeep Dasigi
2024 A* conf
ICLR
Yanai Elazar, Akshita Bhagia, Ian Magnusson, Abhilasha Ravichander, Dustin Schwenk, Alane Suhr, Evan Pete Walsh, Dirk Groeneveld, Luca Soldaini, Sameer Singh, Hannaneh Hajishirzi, Noah A. Smith, Jesse Dodge
2023 J jnl
CoRR
Dirk Groeneveld, Anas Awadalla, Iz Beltagy, Akshita Bhagia, Ian Magnusson, Hao Peng, Oyvind Tafjord, Pete Walsh, Kyle Richardson, Jesse Dodge
2023 J jnl
CoRR
Ian Magnusson, Akshita Bhagia, Valentin Hofmann, Luca Soldaini, Ananya Harsh Jha, Oyvind Tafjord, Dustin Schwenk, Evan Pete Walsh, Yanai Elazar, Kyle Lo, Dirk Groeneveld, Iz Beltagy, Hannaneh Hajishirzi, Noah A. Smith, Kyle Richardson, Jesse Dodge
2023 conf
ACL (Findings)
Ian Magnusson, Noah A. Smith, Jesse Dodge
2023 J jnl
CoRR
Ian Magnusson, Noah A. Smith, Jesse Dodge
2023 J jnl
CoRR
Yanai Elazar, Akshita Bhagia, Ian Magnusson, Abhilasha Ravichander, Dustin Schwenk, Alane Suhr, Pete Walsh, Dirk Groeneveld, Luca Soldaini, Sameer Singh, Hanna Hajishirzi, Noah A. Smith, Jesse Dodge
2022 conf
EMNLP (Findings)
Anas Awadalla, Mitchell Wortsman, Gabriel Ilharco, Sewon Min, Ian Magnusson, Hannaneh Hajishirzi, Ludwig Schmidt
2022 J jnl
CoRR
Anas Awadalla, Mitchell Wortsman, Gabriel Ilharco, Sewon Min, Ian Magnusson, Hannaneh Hajishirzi, Ludwig Schmidt
2022 J jnl
CoRR
Yuling Gu, Yao Fu, Valentina Pyatkin, Ian Magnusson, Bhavana Dalvi Mishra, Peter Clark
2021 conf
ADGD @ ACM Multimedia
Ian Magnusson, Aruna Sankaranarayanan, Andrew Lippman
2021 J jnl
CoRR
Ian Magnusson, Aruna Sankaranarayanan, Andrew Lippman
redb/extractors/js_extractors/js_xray.py
← Index redb/extractors/js_extractors/js_xray.py python
"""Subprocess wrapper for the bundled js-x-ray Node bridge.

Mirrors `js_deobfuscator.py`: shell out to a Node script with a per-sample
timeout, kill the process group on hang, demote `FileNotFoundError` to debug
(missing tool is routine — the host either has Node + the bundled package
installed or it doesn't), and return a structured result on success.

The bridge lives at `redb/extractors/js_extractors/scripts/js-xray-runner.js`.
Operators install the JS dependency once with `npm install` in that directory
(or override the path with `JS_XRAY_RUNNER_PATH`).

Configuration (env vars):
    JS_XRAY_RUNNER_PATH   Path to the Node bridge script (default: bundled).
    JS_XRAY_TIMEOUT       Seconds before the subprocess is killed. Default: 30.

`run(source, log)` returns `XRayResult(obfuscator, warnings)` on a successful
analysis, or `XRayResult(None, [])` for any non-success path (binary missing,
timeout, parse failure, etc.). The two unsuccessful states are
indistinguishable to the caller on purpose — they all collapse to "no
js-x-ray verdict, fall back to heuristic".
"""

from __future__ import annotations

import json
import os
import signal
import subprocess
import tempfile
from dataclasses import dataclass, field
from typing import List, Optional

# Bundled bridge: redb/extractors/js_extractors/scripts/js-xray-runner.js
_DEFAULT_RUNNER = os.path.join(
    os.path.dirname(__file__), "scripts", "js-xray-runner.js"
)
_DEFAULT_NODE = "node"
_DEFAULT_TIMEOUT_SECS = 30


@dataclass
class XRayResult:
    """Parsed js-x-ray output. `obfuscator` is the recognised family name
    (e.g. "jsfuck", "obfuscator.io") or None when js-x-ray did not flag the
    code. `warnings` carries every {kind, value} pair the analyser produced;
    the heuristic uses it as a corroborating signal. `avg_identifier_length`
    is js-x-ray's own AST-derived figure — used as a fallback for the
    heuristic's `avg_identifier_length<2` strong signal when pyjsparser
    can't parse the source (anything ES2015+ trips it)."""

    obfuscator: Optional[str] = None
    warnings: List[dict] = field(default_factory=list)
    avg_identifier_length: Optional[float] = None

    @property
    def flagged(self) -> bool:
        return self.obfuscator is not None


def _empty() -> XRayResult:
    return XRayResult(obfuscator=None, warnings=[])


def run(source: str, log) -> XRayResult:
    if not source:
        return _empty()

    runner = os.getenv("JS_XRAY_RUNNER_PATH", _DEFAULT_RUNNER)
    node_bin = os.getenv("JS_XRAY_NODE_BIN", _DEFAULT_NODE)
    timeout = int(os.getenv("JS_XRAY_TIMEOUT", str(_DEFAULT_TIMEOUT_SECS)))

    if not os.path.exists(runner):
        log.debug(f"js-x-ray runner not found at {runner}")
        return _empty()

    # Skip the subprocess entirely when the JS dependency isn't installed.
    # Without this, every call to a host that has `node` but never ran
    # `npm install` next to the runner would still fork node, get a require
    # error, and exit nonzero — wasted ~50–200ms per JS sample (and per test).
    runner_dir = os.path.dirname(runner)
    if not os.path.isdir(os.path.join(runner_dir, "node_modules", "@nodesecure", "js-x-ray")):
        log.debug(f"@nodesecure/js-x-ray not installed in {runner_dir}")
        return _empty()

    tmp_path = None
    try:
        with tempfile.NamedTemporaryFile(
            suffix=".js", mode="w", delete=False, encoding="utf-8"
        ) as tmp:
            tmp.write(source)
            tmp_path = tmp.name

        try:
            process = subprocess.Popen(
                [node_bin, runner, tmp_path],
                stdout=subprocess.PIPE,
                stderr=subprocess.PIPE,
                preexec_fn=os.setsid,
            )
            try:
                stdout, stderr = process.communicate(timeout=timeout)
            except subprocess.TimeoutExpired:
                # Kill the whole process group so any node helpers die too.
                try:
                    os.killpg(os.getpgid(process.pid), signal.SIGTERM)
                    process.wait(timeout=5)
                except Exception:
                    try:
                        os.killpg(os.getpgid(process.pid), signal.SIGKILL)
                    except Exception:
                        pass
                log.warning(f"js-x-ray timed out after {timeout}s")
                return _empty()

            if process.returncode != 0:
                err = stderr.decode("utf-8", errors="replace").strip()
                log.debug(f"js-x-ray exited {process.returncode}: {err}")
                return _empty()

            text = stdout.decode("utf-8", errors="replace").strip()
            if not text:
                return _empty()

            try:
                payload = json.loads(text)
            except json.JSONDecodeError as e:
                log.warning(f"js-x-ray emitted non-JSON output: {e}")
                return _empty()

            obfuscator = payload.get("obfuscator")
            warnings = payload.get("warnings") or []
            if not isinstance(warnings, list):
                warnings = []

            ids_avg = payload.get("idsLengthAvg")
            if not isinstance(ids_avg, (int, float)):
                ids_avg = None

            return XRayResult(
                obfuscator=obfuscator,
                warnings=warnings,
                avg_identifier_length=ids_avg,
            )
        finally:
            if tmp_path:
                try:
                    os.unlink(tmp_path)
                except Exception:
                    pass
    except FileNotFoundError:
        log.debug(f"node binary not found at {node_bin}")
        return _empty()
    except Exception as e:
        log.error(f"js-x-ray subprocess error: {e}")
        return _empty()