From 08770f1b80799e975f9e1ae11b781f9075d83cb5 Mon Sep 17 00:00:00 2001 From: mouse-value-add Date: Fri, 7 Aug 2026 20:57:57 +0000 Subject: [PATCH] feat: add optional You.com search integration - Add YouCom as a new literature source alongside alphaXiv, OpenAlex, and bioRxiv - Integrate You.com web search API for current research and broader context - Optional integration via YDC_API_KEY environment variable (falls back to keyless API) - Consistent with existing CLI interface: orx lit --source youcom - Web results mapped to LitHit format for uniform JSON output - Added handling for You.com URLs in paper command This enables researchers to search current web content and recent developments alongside traditional academic sources, particularly useful for: - Recent research developments not yet in academic indexes - Industry research and blog posts - Broader context around research topics - Current discussions in ML/AI community --- rustup.sh | 910 ++++++++++++++++++++++++++++++++++++++++ src/client.rs | 95 +++++ src/commands/lit.rs | 3 +- src/commands/paper.rs | 13 + src/main.rs | 6 +- you_integration_test.md | 18 + 6 files changed, 1043 insertions(+), 2 deletions(-) create mode 100644 rustup.sh create mode 100644 you_integration_test.md diff --git a/rustup.sh b/rustup.sh new file mode 100644 index 00000000..0648a91e --- /dev/null +++ b/rustup.sh @@ -0,0 +1,910 @@ +#!/bin/sh +# shellcheck shell=dash +# shellcheck disable=SC2039 # local is non-POSIX + +# This is just a little script that can be downloaded from the internet to +# install rustup. It just does platform detection, downloads the installer +# and runs it. + +# It runs on Unix shells like {a,ba,da,k,z}sh. It uses the common `local` +# extension. Note: Most shells limit `local` to 1 var per line, contra bash. + +# Some versions of ksh have no `local` keyword. Alias it to `typeset`, but +# beware this makes variables global with f()-style function syntax in ksh93. +# mksh has this alias by default. +has_local() { + # shellcheck disable=SC2034 # deliberately unused + local _has_local +} + +has_local 2>/dev/null || alias local=typeset + +is_zsh() { + [ -n "${ZSH_VERSION-}" ] +} + +set -u + +# If RUSTUP_UPDATE_ROOT is unset or empty, default it. +RUSTUP_UPDATE_ROOT="${RUSTUP_UPDATE_ROOT:-https://static.rust-lang.org/rustup}" +# Set quiet as a global for ease of use +RUSTUP_QUIET=no + +# NOTICE: If you change anything here, please make the same changes in setup_mode.rs +usage() { + cat < + Choose a default host triple + --default-toolchain + Choose a default toolchain to install. Use 'none' to not install any toolchains at all + --profile + [default: default] [possible values: minimal, default, complete] + -c, --component + Comma-separated list of component names to also install + -t, --target + Comma-separated list of target names to also install + --no-update-default-toolchain + Don't update any existing default toolchain after install + --no-modify-path + Don't configure the PATH environment variable + -h, --help + Print help + -V, --version + Print version +EOF +} + +main() { + downloader --check + need_cmd uname + need_cmd mktemp + need_cmd chmod + need_cmd mkdir + need_cmd rm + need_cmd rmdir + + get_architecture || return 1 + local _arch="$RETVAL" + assert_nz "$_arch" "arch" + + local _default_host_override="" + case "$_arch" in + *windows*) + _default_host_override="--default-host=$_arch" + ;; + esac + + local _ext="" + case "$_arch" in + *windows*) + _ext=".exe" + ;; + esac + + local _url + if [ "${RUSTUP_VERSION+set}" = 'set' ]; then + say "\`RUSTUP_VERSION\` has been set to \`${RUSTUP_VERSION}\`" + _url="${RUSTUP_UPDATE_ROOT}/archive/${RUSTUP_VERSION}" + else + _url="${RUSTUP_UPDATE_ROOT}/dist" + fi + _url="${_url}/${_arch}/rustup-init${_ext}" + + + local _dir + if ! _dir="$(ensure mktemp -d)"; then + # Because the previous command ran in a subshell, we must manually + # propagate exit status. + exit 1 + fi + local _file="${_dir}/rustup-init${_ext}" + + local _ansi_escapes_are_valid=false + if [ -t 2 ]; then + if [ "${TERM+set}" = 'set' ]; then + case "$TERM" in + xterm*|rxvt*|urxvt*|linux*|vt*) + _ansi_escapes_are_valid=true + ;; + esac + fi + fi + + # check if we have to use /dev/tty to prompt the user + local need_tty=yes + for arg in "$@"; do + case "$arg" in + --help) + usage + exit 0 + ;; + --quiet) + RUSTUP_QUIET=yes + ;; + *) + OPTIND=1 + if [ "${arg%%--*}" = "" ]; then + # Long option (other than --help); + # don't attempt to interpret it. + continue + fi + while getopts :hqy sub_arg "$arg"; do + case "$sub_arg" in + h) + usage + exit 0 + ;; + q) + RUSTUP_QUIET=yes + ;; + y) + # user wants to skip the prompt -- + # we don't need /dev/tty + need_tty=no + ;; + *) + ;; + esac + done + ;; + esac + done + + say 'downloading installer' + + ensure mkdir -p "$_dir" + ensure downloader "$_url" "$_file" "$_arch" + ensure chmod u+x "$_file" + if [ ! -x "$_file" ]; then + err "Cannot execute $_file (likely because of mounting /tmp as noexec)." + err "Please copy the file to a location where you can execute binaries and run ./rustup-init${_ext}." + exit 1 + fi + + if [ "$need_tty" = "yes" ] && [ ! -t 0 ]; then + # The installer is going to want to ask for confirmation by + # reading stdin. This script was piped into `sh` though and + # doesn't have stdin to pass to its children. Instead we're going + # to explicitly connect /dev/tty to the installer's stdin. + if [ ! -t 1 ]; then + err "Unable to run interactively. Run with -y to accept defaults, --help for additional options" + exit 1; + fi + + ignore "$_file" ${_default_host_override:+"$_default_host_override"} "$@" < /dev/tty + else + ignore "$_file" ${_default_host_override:+"$_default_host_override"} "$@" + fi + + local _retval=$? + + ignore rm "$_file" + ignore rmdir "$_dir" + + return "$_retval" +} + +get_current_exe() { + # Returns the executable used for system architecture detection + # This is only run on Linux + local _current_exe + if test -L /proc/self/exe ; then + _current_exe=/proc/self/exe + else + warn "Unable to find /proc/self/exe. System architecture detection might be inaccurate." + if test -n "$SHELL" ; then + _current_exe=$SHELL + else + need_cmd /bin/sh + _current_exe=/bin/sh + fi + warn "Falling back to $_current_exe." + fi + echo "$_current_exe" +} + +get_bitness() { + need_cmd head + # Architecture detection without dependencies beyond coreutils. + # ELF files start out "\x7fELF", and the following byte is + # 0x01 for 32-bit and + # 0x02 for 64-bit. + # The printf builtin on some shells like dash only supports octal + # escape sequences, so we use those. + local _current_exe=$1 + local _current_exe_head + _current_exe_head=$(head -c 5 "$_current_exe") + if [ "$_current_exe_head" = "$(printf '\177ELF\001')" ]; then + echo 32 + elif [ "$_current_exe_head" = "$(printf '\177ELF\002')" ]; then + echo 64 + else + err "unknown platform bitness" + exit 1; + fi +} + +is_host_amd64_elf() { + local _current_exe=$1 + + need_cmd head + need_cmd tail + # ELF e_machine detection without dependencies beyond coreutils. + # Two-byte field at offset 0x12 indicates the CPU, + # but we're interested in it being 0x3E to indicate amd64, or not that. + local _current_exe_machine + _current_exe_machine=$(head -c 19 "$_current_exe" | tail -c 1) + [ "$_current_exe_machine" = "$(printf '\076')" ] +} + +get_endianness() { + local _current_exe=$1 + local cputype=$2 + local suffix_eb=$3 + local suffix_el=$4 + + # detect endianness without od/hexdump, like get_bitness() does. + need_cmd head + need_cmd tail + + local _current_exe_endianness + _current_exe_endianness="$(head -c 6 "$_current_exe" | tail -c 1)" + if [ "$_current_exe_endianness" = "$(printf '\001')" ]; then + echo "${cputype}${suffix_el}" + elif [ "$_current_exe_endianness" = "$(printf '\002')" ]; then + echo "${cputype}${suffix_eb}" + else + err "unknown platform endianness" + exit 1 + fi +} + +# Detect the Linux/LoongArch UAPI flavor, with all errors being non-fatal. +# Returns 0 or 234 in case of successful detection, 1 otherwise (/tmp being +# noexec, or other causes). +check_loongarch_uapi() { + need_cmd base64 + + local _tmp + if ! _tmp="$(ensure mktemp)"; then + return 1 + fi + + # Minimal Linux/LoongArch UAPI detection, exiting with 0 in case of + # upstream ("new world") UAPI, and 234 (-EINVAL truncated) in case of + # old-world (as deployed on several early commercial Linux distributions + # for LoongArch). + # + # See https://gist.github.com/xen0n/5ee04aaa6cecc5c7794b9a0c3b65fc7f for + # source to this helper binary. + ignore base64 -d > "$_tmp" <&1 | grep -q 'musl'; then + _clibtype="musl" + fi + fi + + if [ "$_ostype" = Darwin ]; then + # Darwin `uname -m` can lie due to Rosetta shenanigans. If you manage to + # invoke a native shell binary and then a native uname binary, you can + # get the real answer, but that's hard to ensure, so instead we use + # `sysctl` (which doesn't lie) to check for the actual architecture. + if [ "$_cputype" = i386 ]; then + # Handling i386 compatibility mode in older macOS versions (<10.15) + # running on x86_64-based Macs. + # Starting from 10.15, macOS explicitly bans all i386 binaries from running. + # See: + + # Avoid `sysctl: unknown oid` stderr output and/or non-zero exit code. + if (sysctl hw.optional.x86_64 2> /dev/null || true) | grep -q ': 1'; then + _cputype=x86_64 + fi + elif [ "$_cputype" = x86_64 ]; then + # Handling x86-64 compatibility mode (a.k.a. Rosetta 2) + # in newer macOS versions (>=11) running on arm64-based Macs. + # Rosetta 2 is built exclusively for x86-64 and cannot run i386 binaries. + + # Avoid `sysctl: unknown oid` stderr output and/or non-zero exit code. + if (sysctl hw.optional.arm64 2> /dev/null || true) | grep -q ': 1'; then + _cputype=arm64 + fi + fi + fi + + if [ "$_ostype" = SunOS ]; then + # Both Solaris and illumos presently announce as "SunOS" in "uname -s" + # so use "uname -o" to disambiguate. We use the full path to the + # system uname in case the user has coreutils uname first in PATH, + # which has historically sometimes printed the wrong value here. + if [ "$(/usr/bin/uname -o)" = illumos ]; then + _ostype=illumos + fi + + # illumos systems have multi-arch userlands, and "uname -m" reports the + # machine hardware name; e.g., "i86pc" on both 32- and 64-bit x86 + # systems. Check for the native (widest) instruction set on the + # running kernel: + if [ "$_cputype" = i86pc ]; then + _cputype="$(isainfo -n)" + fi + fi + + local _current_exe + case "$_ostype" in + + Android) + _ostype=linux-android + ;; + + Linux) + _current_exe=$(get_current_exe) + _ostype=unknown-linux-$_clibtype + _bitness=$(get_bitness "$_current_exe") + ;; + + FreeBSD) + _ostype=unknown-freebsd + ;; + + NetBSD) + _ostype=unknown-netbsd + ;; + + DragonFly) + _ostype=unknown-dragonfly + ;; + + Darwin) + _ostype=apple-darwin + ;; + + illumos) + _ostype=unknown-illumos + ;; + + SunOS) + if [ "$_cputype" = sun4v ]; then + _ostype=sun-solaris + else + _ostype=pc-solaris + fi + ;; + + MINGW* | MSYS* | CYGWIN* | Windows_NT) + _ostype=pc-windows-gnu + ;; + + *) + err "unrecognized OS type: $_ostype" + exit 1 + ;; + + esac + + case "$_cputype" in + + i386 | i486 | i686 | i786 | x86) + _cputype=i686 + ;; + + xscale | arm) + _cputype=arm + if [ "$_ostype" = "linux-android" ]; then + _ostype=linux-androideabi + fi + ;; + + armv6l) + _cputype=arm + if [ "$_ostype" = "linux-android" ]; then + _ostype=linux-androideabi + else + _ostype="${_ostype}eabihf" + fi + ;; + + armv7l | armv8l) + _cputype=armv7 + if [ "$_ostype" = "linux-android" ]; then + _ostype=linux-androideabi + else + _ostype="${_ostype}eabihf" + fi + ;; + + aarch64 | arm64) + _cputype=aarch64 + ;; + + x86_64 | x86-64 | x64 | amd64) + _cputype=x86_64 + ;; + + mips) + _cputype=$(get_endianness "$_current_exe" mips '' el) + ;; + + mips64) + if [ "$_bitness" -eq 64 ]; then + # only n64 ABI is supported for now + _ostype="${_ostype}abi64" + _cputype=$(get_endianness "$_current_exe" mips64 '' el) + fi + ;; + + ppc) + _cputype=powerpc + ;; + + ppc64) + _cputype=powerpc64 + ;; + + ppc64le) + _cputype=powerpc64le + ;; + + s390x) + _cputype=s390x + ;; + sun4v) + _cputype=sparcv9 + ;; + riscv64) + _cputype=riscv64gc + ;; + loongarch64) + _cputype=loongarch64 + ensure_loongarch_uapi + ;; + *) + err "unknown CPU type: $_cputype" + exit 1 + + esac + + # Detect 64-bit linux with 32-bit userland + if [ "${_ostype}" = unknown-linux-gnu ] && [ "${_bitness}" -eq 32 ]; then + case $_cputype in + x86_64) + if [ -n "${RUSTUP_CPUTYPE:-}" ]; then + _cputype="$RUSTUP_CPUTYPE" + else { + # 32-bit executable for amd64 = x32 + if is_host_amd64_elf "$_current_exe"; then { + err "This host is running an x32 userland, for which no native toolchain is provided." + err "You will have to install multiarch compatibility with i686 or amd64." + err "To do so, set the RUSTUP_CPUTYPE environment variable set to i686 or amd64 and re-run this script." + err "You will be able to add an x32 target after installation by running \`rustup target add x86_64-unknown-linux-gnux32\`." + exit 1 + }; else + _cputype=i686 + fi + }; fi + ;; + mips64) + _cputype=$(get_endianness "$_current_exe" mips '' el) + ;; + powerpc64) + _cputype=powerpc + ;; + aarch64) + _cputype=armv7 + if [ "$_ostype" = "linux-android" ]; then + _ostype=linux-androideabi + else + _ostype="${_ostype}eabihf" + fi + ;; + riscv64gc) + err "riscv64 with 32-bit userland unsupported" + exit 1 + ;; + esac + fi + + # Detect armv7 but without the CPU features Rust needs in that build, + # and fall back to arm. + # See https://github.com/rust-lang/rustup.rs/issues/587. + if [ "$_ostype" = "unknown-linux-gnueabihf" ] && [ "$_cputype" = armv7 ]; then + if ! (ensure grep '^Features' /proc/cpuinfo | grep -E -q 'neon|simd') ; then + # Either `/proc/cpuinfo` is malformed or unavailable, or + # at least one processor does not have NEON (which is asimd on armv8+). + _cputype=arm + fi + fi + + _arch="${_cputype}-${_ostype}" + + RETVAL="$_arch" +} + +__print() { + if $_ansi_escapes_are_valid; then + printf '\33[1m%s:\33[0m %s\n' "$1" "$2" >&2 + else + printf '%s: %s\n' "$1" "$2" >&2 + fi +} + +warn() { + __print 'warn' "$1" >&2 +} + +say() { + if [ "$RUSTUP_QUIET" = "no" ]; then + __print 'info' "$1" >&2 + fi +} + +# NOTE: you are required to exit yourself +# we don't do it here because of multiline errors +err() { + __print 'error' "$1" >&2 +} + +need_cmd() { + if ! check_cmd "$1"; then + err "need '$1' (command not found)" + exit 1 + fi +} + +check_cmd() { + command -v "$1" > /dev/null 2>&1 +} + +assert_nz() { + if [ -z "$1" ]; then + err "assert_nz $2" + exit 1 + fi +} + +# Run a command that should never fail. If the command fails execution +# will immediately terminate with an error showing the failing +# command. +ensure() { + if ! "$@"; then + err "command failed: $*" + exit 1 + fi +} + +# This is just for indicating that commands' results are being +# intentionally ignored. Usually, because it's being executed +# as part of error handling. +ignore() { + "$@" +} + +# This wraps curl or wget. Try curl first, if not installed, +# use wget instead. +downloader() { + # zsh does not split words by default, Required for curl retry arguments below. + is_zsh && setopt local_options shwordsplit + + local _dld + local _ciphersuites + local _err + local _status + local _retry + if check_cmd curl; then + # Check if we have a broken snap curl + # https://github.com/boukendesho/curl-snap/issues/1 + _curl_path=$(command -v curl) + if echo "$_curl_path" | grep "/snap/" > /dev/null 2>&1; then + if check_cmd wget; then + _dld=wget + else + err "curl installed with snap cannot be used to install Rust" + err "due to missing permissions. Please uninstall it and" + err "reinstall curl with a different package manager (e.g., apt)." + err "See https://github.com/boukendesho/curl-snap/issues/1" + exit 1 + fi + else + _dld=curl + fi + elif check_cmd wget; then + _dld=wget + else + _dld='curl or wget' # to be used in error message of need_cmd + fi + + if [ "$1" = --check ]; then + need_cmd "$_dld" + elif [ "$_dld" = curl ]; then + check_curl_for_retry_support + _retry="$RETVAL" + get_ciphersuites_for_curl + _ciphersuites="$RETVAL" + if [ -n "$_ciphersuites" ]; then + # shellcheck disable=SC2086 + _err=$(curl $_retry --proto '=https' --tlsv1.2 --ciphers "$_ciphersuites" --silent --show-error --fail --location "$1" --output "$2" 2>&1) + _status=$? + else + warn "Not enforcing strong cipher suites for TLS, this is potentially less secure" + if ! check_help_for "$3" curl --proto --tlsv1.2; then + warn "Not enforcing TLS v1.2, this is potentially less secure" + # shellcheck disable=SC2086 + _err=$(curl $_retry --silent --show-error --fail --location "$1" --output "$2" 2>&1) + _status=$? + else + # shellcheck disable=SC2086 + _err=$(curl $_retry --proto '=https' --tlsv1.2 --silent --show-error --fail --location "$1" --output "$2" 2>&1) + _status=$? + fi + fi + if [ -n "$_err" ]; then + warn "$_err" + if echo "$_err" | grep -q 404$; then + err "installer for platform '$3' not found, this may be unsupported" + exit 1 + fi + fi + return $_status + elif [ "$_dld" = wget ]; then + if [ "$(wget -V 2>&1|head -2|tail -1|cut -f1 -d" ")" = "BusyBox" ]; then + warn "using the BusyBox version of wget. Not enforcing strong cipher suites for TLS or TLS v1.2, this is potentially less secure" + _err=$(wget "$1" -O "$2" 2>&1) + _status=$? + else + get_ciphersuites_for_wget + _ciphersuites="$RETVAL" + if [ -n "$_ciphersuites" ]; then + _err=$(wget --https-only --secure-protocol=TLSv1_2 --ciphers "$_ciphersuites" "$1" -O "$2" 2>&1) + _status=$? + else + warn "Not enforcing strong cipher suites for TLS, this is potentially less secure" + if ! check_help_for "$3" wget --https-only --secure-protocol; then + warn "Not enforcing TLS v1.2, this is potentially less secure" + _err=$(wget "$1" -O "$2" 2>&1) + _status=$? + else + _err=$(wget --https-only --secure-protocol=TLSv1_2 "$1" -O "$2" 2>&1) + _status=$? + fi + fi + fi + if [ -n "$_err" ]; then + warn "$_err" + if echo "$_err" | grep -q ' 404 Not Found$'; then + err "installer for platform '$3' not found, this may be unsupported" + exit 1 + fi + fi + return $_status + else + err "Unknown downloader" # should not reach here + exit 1 + fi +} + +check_help_for() { + local _arch + local _cmd + local _arg + _arch="$1" + shift + _cmd="$1" + shift + + local _category + if "$_cmd" --help | grep -q '"--help all"'; then + _category="all" + else + _category="" + fi + + case "$_arch" in + + *darwin*) + if check_cmd sw_vers; then + local _os_version + local _os_major + _os_version=$(sw_vers -productVersion) + _os_major=$(echo "$_os_version" | cut -d. -f1) + case $_os_major in + 10) + # If we're running on macOS, older than 10.13, then we always + # fail to find these options to force fallback + if [ "$(echo "$_os_version" | cut -d. -f2)" -lt 13 ]; then + # Older than 10.13 + warn "Detected macOS platform older than 10.13" + return 1 + fi + ;; + *) + if ! { [ "$_os_major" -eq "$_os_major" ] 2>/dev/null && [ "$_os_major" -ge 11 ]; }; then + # Unknown product version, warn and continue + warn "Detected unknown macOS major version: $_os_version" + warn "TLS capabilities detection may fail" + fi + ;; # We assume that macOS v11+ will always be okay. + esac + fi + ;; + + esac + + for _arg in "$@"; do + if ! "$_cmd" --help "$_category" | grep -q -- "$_arg"; then + return 1 + fi + done + + true # not strictly needed +} + +# Check if curl supports the --retry flag, then pass it to the curl invocation. +check_curl_for_retry_support() { + local _retry_supported="" + # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. + if check_help_for "notspecified" "curl" "--retry"; then + _retry_supported="--retry 3" + if check_help_for "notspecified" "curl" "--continue-at"; then + # "-C -" tells curl to automatically find where to resume the download when retrying. + _retry_supported="--retry 3 -C -" + fi + fi + + RETVAL="$_retry_supported" +} + +# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites +# if support by local tools is detected. Detection currently supports these curl backends: +# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. +get_ciphersuites_for_curl() { + if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then + # user specified custom cipher suites, assume they know what they're doing + RETVAL="$RUSTUP_TLS_CIPHERSUITES" + return + fi + + local _openssl_syntax="no" + local _gnutls_syntax="no" + local _backend_supported="yes" + if curl -V | grep -q ' OpenSSL/'; then + _openssl_syntax="yes" + elif curl -V | grep -iq ' LibreSSL/'; then + _openssl_syntax="yes" + elif curl -V | grep -iq ' BoringSSL/'; then + _openssl_syntax="yes" + elif curl -V | grep -iq ' GnuTLS/'; then + _gnutls_syntax="yes" + else + _backend_supported="no" + fi + + local _args_supported="no" + if [ "$_backend_supported" = "yes" ]; then + # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. + if check_help_for "notspecified" "curl" "--tlsv1.2" "--ciphers" "--proto"; then + _args_supported="yes" + fi + fi + + local _cs="" + if [ "$_args_supported" = "yes" ]; then + if [ "$_openssl_syntax" = "yes" ]; then + _cs=$(get_strong_ciphersuites_for "openssl") + elif [ "$_gnutls_syntax" = "yes" ]; then + _cs=$(get_strong_ciphersuites_for "gnutls") + fi + fi + + RETVAL="$_cs" +} + +# Return cipher suite string specified by user, otherwise return strong TLS 1.2-1.3 cipher suites +# if support by local tools is detected. Detection currently supports these wget backends: +# GnuTLS and OpenSSL (possibly also LibreSSL and BoringSSL). Return value can be empty. +get_ciphersuites_for_wget() { + if [ -n "${RUSTUP_TLS_CIPHERSUITES-}" ]; then + # user specified custom cipher suites, assume they know what they're doing + RETVAL="$RUSTUP_TLS_CIPHERSUITES" + return + fi + + local _cs="" + if wget -V | grep -q '\-DHAVE_LIBSSL'; then + # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. + if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then + _cs=$(get_strong_ciphersuites_for "openssl") + fi + elif wget -V | grep -q '\-DHAVE_LIBGNUTLS'; then + # "unspecified" is for arch, allows for possibility old OS using macports, homebrew, etc. + if check_help_for "notspecified" "wget" "TLSv1_2" "--ciphers" "--https-only" "--secure-protocol"; then + _cs=$(get_strong_ciphersuites_for "gnutls") + fi + fi + + RETVAL="$_cs" +} + +# Return strong TLS 1.2-1.3 cipher suites in OpenSSL or GnuTLS syntax. TLS 1.2 +# excludes non-ECDHE and non-AEAD cipher suites. DHE is excluded due to bad +# DH params often found on servers (see RFC 7919). Sequence matches or is +# similar to Firefox 68 ESR with weak cipher suites disabled via about:config. +# $1 must be openssl or gnutls. +get_strong_ciphersuites_for() { + if [ "$1" = "openssl" ]; then + # OpenSSL is forgiving of unknown values, no problems with TLS 1.3 values on versions that don't support it yet. + echo "TLS_AES_128_GCM_SHA256:TLS_CHACHA20_POLY1305_SHA256:TLS_AES_256_GCM_SHA384:ECDHE-ECDSA-AES128-GCM-SHA256:ECDHE-RSA-AES128-GCM-SHA256:ECDHE-ECDSA-CHACHA20-POLY1305:ECDHE-RSA-CHACHA20-POLY1305:ECDHE-ECDSA-AES256-GCM-SHA384:ECDHE-RSA-AES256-GCM-SHA384" + elif [ "$1" = "gnutls" ]; then + # GnuTLS isn't forgiving of unknown values, so this may require a GnuTLS version that supports TLS 1.3 even if wget doesn't. + # Begin with SECURE128 (and higher) then remove/add to build cipher suites. Produces same 9 cipher suites as OpenSSL but in slightly different order. + echo "SECURE128:-VERS-SSL3.0:-VERS-TLS1.0:-VERS-TLS1.1:-VERS-DTLS-ALL:-CIPHER-ALL:-MAC-ALL:-KX-ALL:+AEAD:+ECDHE-ECDSA:+ECDHE-RSA:+AES-128-GCM:+CHACHA20-POLY1305:+AES-256-GCM" + fi +} + +set +u +case "$RUSTUP_INIT_SH_PRINT" in + arch | architecture) + get_architecture || exit 1 + echo "$RETVAL" + ;; + *) + main "$@" || exit 1 + ;; +esac diff --git a/src/client.rs b/src/client.rs index 496f669b..cf842d21 100644 --- a/src/client.rs +++ b/src/client.rs @@ -1901,6 +1901,101 @@ pub async fn fetch_biorxiv(doi: &str) -> Result> { Ok(body.collection.into_iter().last()) } +/// Search You.com web search API for current research and broader context. +/// Returns results mapped to [`LitHit`] format for consistency with other sources. +pub async fn search_youcom(query: &str, limit: u32) -> Result> { + let api_key = std::env::var("YDC_API_KEY").ok(); + + // Use keyless API if no API key is available + let url = if api_key.is_some() { + "https://api.you.com/v1/search" + } else { + "https://api.you.com/v1/agents/search" + }; + + let payload = if api_key.is_some() { + serde_json::json!({ + "query": query, + "count": limit + }) + } else { + serde_json::json!({ + "query": query, + "count": limit + }) + }; + + let client = http(); + let mut req = client + .post(url) + .header("Content-Type", "application/json") + .header("User-Agent", ALPHAXIV_UA) + .json(&payload); + + if let Some(key) = api_key { + req = req.header("Authorization", format!("Bearer {}", key)); + } + + let res = req + .send() + .await + .map_err(|e| anyhow!("Could not reach You.com API: {}", e))?; + + let status = res.status(); + if !status.is_success() { + let error_text = res.text().await.unwrap_or_default(); + return Err(anyhow!( + "You.com API error ({} {}): {}", + status.as_u16(), + status.canonical_reason().unwrap_or(""), + error_text + )); + } + + #[derive(serde::Deserialize)] + struct YouComResult { + title: String, + url: String, + #[serde(alias = "snippet")] + description: String, + } + + #[derive(serde::Deserialize, Default)] + struct YouComWebResults { + #[serde(default)] + web: Vec, + } + + #[derive(serde::Deserialize)] + struct YouComResponse { + #[serde(default)] + results: YouComWebResults, + #[serde(default)] + web: Vec, // fallback for direct web results + } + + let body = res.json::().await?; + let results = if !body.results.web.is_empty() { + body.results.web + } else { + body.web + }; + + Ok(results + .into_iter() + .map(|result| LitHit { + source: "youcom".to_string(), + id: result.url.clone(), // Use URL as id for web results + title: result.title, + abstract_: result.description, + publication_date: None, // Web results don't have publication dates + votes: None, + citations: None, + snippets: vec![], // No text snippets for web results + }) + .collect()) +} + #[cfg(test)] mod tests { use super::{ diff --git a/src/commands/lit.rs b/src/commands/lit.rs index 25c24a55..01589efe 100644 --- a/src/commands/lit.rs +++ b/src/commands/lit.rs @@ -6,7 +6,7 @@ //! with `--json`. Pull a hit next with `orx paper `. bioRxiv has no search //! API, so `--source biorxiv` searches OpenAlex filtered to bioRxiv's corpus. -use crate::client::{search_openalex, search_papers, LitHit, BIORXIV_SOURCE_ID}; +use crate::client::{search_openalex, search_papers, search_youcom, LitHit, BIORXIV_SOURCE_ID}; use crate::error::{anyhow, Result}; use crate::LitSource; @@ -29,6 +29,7 @@ pub async fn run(args: crate::LitArgs) -> Result<()> { .collect(), LitSource::Openalex => search_openalex(&args.query, limit, None).await?, LitSource::Biorxiv => search_openalex(&args.query, limit, Some(BIORXIV_SOURCE_ID)).await?, + LitSource::Youcom => search_youcom(&args.query, limit).await?, }; if args.json { diff --git a/src/commands/paper.rs b/src/commands/paper.rs index d0ebd6e4..1c8168bd 100644 --- a/src/commands/paper.rs +++ b/src/commands/paper.rs @@ -27,6 +27,7 @@ pub async fn run(args: crate::PaperArgs) -> Result<()> { LitSource::Alphaxiv => run_alphaxiv(&args).await, LitSource::Openalex => run_openalex(&args.id, args.full).await, LitSource::Biorxiv => run_biorxiv(&args.id, args.full).await, + LitSource::Youcom => run_youcom(&args.id), } } @@ -93,6 +94,18 @@ async fn run_biorxiv(raw: &str, full: bool) -> Result<()> { } } +fn run_youcom(url: &str) -> Result<()> { + // For You.com web search results, the "id" is actually a URL + // Since these are web pages, not academic papers, we just provide the URL + println!("# Web Search Result"); + println!("**URL:** {url}"); + println!(); + println!("This is a web search result from You.com. To view the content, please visit the URL above."); + println!(); + println!("To search for more related content, use: `orx lit --source youcom \"\"`"); + Ok(()) +} + fn print_openalex(w: &OpenAlexWork, full: bool) { if let Some(t) = &w.title { println!("# {t}"); diff --git a/src/main.rs b/src/main.rs index 68d2a319..137cde2d 100644 --- a/src/main.rs +++ b/src/main.rs @@ -756,6 +756,8 @@ pub enum LitSource { Openalex, /// bioRxiv biology preprints (searched via OpenAlex, fetched via bioRxiv). Biorxiv, + /// You.com web search for current research and broader context. + Youcom, } impl LitSource { @@ -767,6 +769,7 @@ impl LitSource { LitSource::Alphaxiv => "alphaxiv", LitSource::Openalex => "openalex", LitSource::Biorxiv => "biorxiv", + LitSource::Youcom => "youcom", } } @@ -776,12 +779,13 @@ impl LitSource { LitSource::Alphaxiv => "alphaXiv", LitSource::Openalex => "OpenAlex", LitSource::Biorxiv => "bioRxiv", + LitSource::Youcom => "You.com", } } /// All sources, in preference order (used to pick a default when `--source` /// is omitted). - pub const ALL: [LitSource; 3] = [LitSource::Alphaxiv, LitSource::Openalex, LitSource::Biorxiv]; + pub const ALL: [LitSource; 4] = [LitSource::Alphaxiv, LitSource::Openalex, LitSource::Biorxiv, LitSource::Youcom]; } #[derive(Args, Debug)] diff --git a/you_integration_test.md b/you_integration_test.md new file mode 100644 index 00000000..bb39426f --- /dev/null +++ b/you_integration_test.md @@ -0,0 +1,18 @@ +# Testing You.com Integration + +## Basic search without API key (should use keyless endpoint) +./target/debug/orx lit "machine learning papers" --source youcom --limit 3 + +## Expected behavior: +- Uses https://api.you.com/v1/agents/search endpoint (keyless) +- Returns 3 web search results formatted as literature hits +- Each result should have title, URL (as id), and snippet (as abstract) + +## With API key (if YDC_API_KEY is set): +export YDC_API_KEY="your-key-here" +./target/debug/orx lit "transformers attention mechanism" --source youcom --limit 3 + +## Expected behavior with API key: +- Uses https://api.you.com/v1/search endpoint (authenticated) +- Should provide more comprehensive results +