Use ChatGPT Codex for launcher questions
This commit is contained in:
parent
af95242b60
commit
4c6d610263
2 changed files with 40 additions and 94 deletions
|
|
@ -69,45 +69,20 @@
|
||||||
# Instant-answer backend for the quickshell launcher on the desktops:
|
# Instant-answer backend for the quickshell launcher on the desktops:
|
||||||
# `printf '%s' "question" | ssh mediaserver qs-ask`.
|
# `printf '%s' "question" | ssh mediaserver qs-ask`.
|
||||||
#
|
#
|
||||||
# Runs through Claude Code so it draws on fred's existing subscription —
|
# Runs through Codex CLI using fred's ChatGPT subscription. Sign in on
|
||||||
# no API key, no separate billing. The trade is latency: Claude Code boots
|
# the server as fred with `codex login` once; no API key is needed.
|
||||||
# a Node process per question, so this takes ~6s against the raw API's
|
# The launcher shows Wikipedia first and replaces it when this answers.
|
||||||
# ~1s. The launcher therefore shows its Wikipedia answer first and lets
|
|
||||||
# this one supersede it when it lands.
|
|
||||||
#
|
#
|
||||||
# Reads the question from stdin — passing it as an ssh argv element would
|
# Reads the question from stdin — passing it as an ssh argv element would
|
||||||
# send it through the remote shell for a second round of word splitting.
|
# send it through the remote shell for a second round of word splitting.
|
||||||
#
|
#
|
||||||
# Every flag here is either isolation or trimming the invocation:
|
# Luna at low reasoning effort is enough for a short fact. The CLI runs
|
||||||
# --safe-mode no CLAUDE.md, skills, hooks, plugins, MCP or
|
# read-only, without web search, shell access, user config, exec rules,
|
||||||
# custom agents. Critical for correctness, not
|
# or session persistence.
|
||||||
# just speed: fred's ~/.claude/CLAUDE.md tells
|
|
||||||
# Claude to answer in caveman mode, which would
|
|
||||||
# otherwise leak into launcher answers. Auth
|
|
||||||
# keeps working normally, which is what makes
|
|
||||||
# the subscription usable here.
|
|
||||||
# --no-session-persistence writes no transcript, so launcher questions
|
|
||||||
# never appear in /resume or any project history
|
|
||||||
# --tools "" no tools at all: nothing to load, and it
|
|
||||||
# cannot touch the filesystem or network
|
|
||||||
# --disable-slash-commands no skill resolution
|
|
||||||
# --strict-mcp-config ignore every MCP config (none supplied)
|
|
||||||
# --system-prompt replaces the full Claude Code system prompt
|
|
||||||
# with one line, rather than appending to it
|
|
||||||
# --permission-mode dontAsk never block waiting on a prompt nobody sees
|
|
||||||
# --model claude-haiku-4-5 smallest model — a one-line fact needs no more
|
|
||||||
#
|
|
||||||
# NOT --bare, which looks like the right flag and is a trap: it skips even
|
|
||||||
# more, but its auth is "strictly ANTHROPIC_API_KEY or apiKeyHelper (OAuth
|
|
||||||
# and keychain are never read)" — i.e. it cannot use the subscription.
|
|
||||||
#
|
#
|
||||||
# STDOUT is the launcher's channel: the answer, or nothing at all if the
|
# STDOUT is the launcher's channel: the answer, or nothing at all if the
|
||||||
# call fails or the model doesn't know — silence is what the launcher
|
# call fails or the model doesn't know — silence means "keep Wikipedia".
|
||||||
# reads as "keep the Wikipedia answer". Every reason for that silence goes
|
# Reasons go to STDERR and the journal (`journalctl -t qs-ask`).
|
||||||
# to STDERR instead, which the launcher discards and a human running
|
|
||||||
# `qs-ask` by hand can read — and to the journal, since over ssh nobody
|
|
||||||
# ever sees that STDERR. `journalctl -t qs-ask` is the only way to tell a
|
|
||||||
# blank card caused by UNKNOWN from one caused by a failed call.
|
|
||||||
(pkgs.writeShellScriptBin "qs-ask" ''
|
(pkgs.writeShellScriptBin "qs-ask" ''
|
||||||
# Every line carries this script's pid, because the launcher fires a
|
# Every line carries this script's pid, because the launcher fires a
|
||||||
# fresh call per debounce and concurrent runs otherwise interleave with
|
# fresh call per debounce and concurrent runs otherwise interleave with
|
||||||
|
|
@ -124,45 +99,25 @@
|
||||||
[ -z "$q" ] && { log "no question on stdin"; exit 0; }
|
[ -z "$q" ] && { log "no question on stdin"; exit 0; }
|
||||||
log "ask: $q"
|
log "ask: $q"
|
||||||
|
|
||||||
# Claude Code derives its project state (transcripts, auto-memory) from
|
# Keep repository instructions and project files out of this invocation.
|
||||||
# the working directory, so run from a dedicated empty one. Combined with
|
|
||||||
# --no-session-persistence and --safe-mode, a launcher question leaves no
|
|
||||||
# trace in the real projects' history or memory.
|
|
||||||
work="$HOME/.cache/qs-ask"
|
work="$HOME/.cache/qs-ask"
|
||||||
${pkgs.coreutils}/bin/mkdir -p "$work" || { log "cannot create $work"; exit 0; }
|
${pkgs.coreutils}/bin/mkdir -p "$work" || { log "cannot create $work"; exit 0; }
|
||||||
cd "$work" || { log "cannot enter $work"; exit 0; }
|
cd "$work" || { log "cannot enter $work"; exit 0; }
|
||||||
|
|
||||||
# MAX_THINKING_TOKENS=0 is the single biggest win here, worth ~2.4x.
|
ans=$(printf 'Answer the question in one or two short sentences, under 240 characters. Lead with the specific fact asked for, including units. The query is often bare keywords rather than a question ("shadow hunter knife classic wow"); treat that as a request for the key facts about that thing. No preamble, caveats, markdown, or follow-up offers. Do not use tools. Reply with exactly UNKNOWN only when you genuinely do not know, or the answer depends on live data you do not have.\n\nQuestion: %s\n' "$q" | \
|
||||||
# Measured: Haiku defaults to extended thinking, and it spent 225 output
|
${pkgs.coreutils}/bin/timeout 30s ${pkgs.codex}/bin/codex exec \
|
||||||
# tokens and 1.9s deliberating before emitting an 80-character answer
|
--model gpt-5.6-luna \
|
||||||
# about the height of a tower. With thinking off: 43 output tokens,
|
--config model_reasoning_effort=low \
|
||||||
# time-to-first-text 2657ms -> 1122ms, wall clock 5.7s -> 2.4s, same
|
--config web_search=disabled \
|
||||||
# answer. A one-line lookup has nothing to reason about.
|
--disable shell_tool \
|
||||||
#
|
--sandbox read-only \
|
||||||
# --name is not cosmetic: without it Claude Code fires a second,
|
--skip-git-repo-check \
|
||||||
# concurrent `source=generate_session_title` request to invent a title
|
--ephemeral \
|
||||||
# for a session --no-session-persistence then throws away. It costs no
|
--ignore-user-config \
|
||||||
# wall time (both requests dispatch ~1ms apart), but it spent 507 input
|
--ignore-rules \
|
||||||
# tokens against 153 for the real question. Naming the session up front
|
- \
|
||||||
# removes it: one request instead of two, 660 input tokens -> 153.
|
|
||||||
#
|
|
||||||
# For the record, two things that are NOT the bottleneck, both measured:
|
|
||||||
# process startup (0.12s boot + 28ms to fire the request, so keeping a
|
|
||||||
# warm process saves nothing) and prompt size (184 input tokens).
|
|
||||||
ans=$(printf '%s' "$q" | ${pkgs.coreutils}/bin/env MAX_THINKING_TOKENS=0 \
|
|
||||||
${pkgs.claude-code}/bin/claude -p \
|
|
||||||
--safe-mode \
|
|
||||||
--no-session-persistence \
|
|
||||||
--tools "" \
|
|
||||||
--disable-slash-commands \
|
|
||||||
--strict-mcp-config \
|
|
||||||
--permission-mode dontAsk \
|
|
||||||
--output-format text \
|
|
||||||
--model claude-haiku-4-5 \
|
|
||||||
--name qs-ask \
|
|
||||||
--system-prompt 'Answer the question in one or two short sentences, under 240 characters. Lead with the specific fact asked for, including units. The query is often bare keywords rather than a question ("shadow hunter knife classic wow"); treat that as a request for the key facts about that thing, not as unanswerable. No preamble, no caveats, no markdown, no follow-up offers. Reply with exactly UNKNOWN only when you genuinely do not know, or the answer depends on live data you do not have.' \
|
|
||||||
2>/dev/null) \
|
2>/dev/null) \
|
||||||
|| { log "claude exited non-zero (auth expired? run 'claude' once to log in)"; exit 0; }
|
|| { log "codex exited non-zero (check 'codex login status' as fred)"; exit 0; }
|
||||||
|
|
||||||
ans=$(printf '%s' "$ans" | ${pkgs.coreutils}/bin/tr -d '\r' | ${pkgs.gnused}/bin/sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')
|
ans=$(printf '%s' "$ans" | ${pkgs.coreutils}/bin/tr -d '\r' | ${pkgs.gnused}/bin/sed -e 's/^[[:space:]]*//' -e 's/[[:space:]]*$//')
|
||||||
|
|
||||||
|
|
|
||||||
|
|
@ -344,19 +344,16 @@ in
|
||||||
'https://en.wikipedia.org/w/api.php?action=query&format=json&generator=search&gsrlimit=3&prop=extracts&exintro&explaintext&redirects=1'
|
'https://en.wikipedia.org/w/api.php?action=query&format=json&generator=search&gsrlimit=3&prop=extracts&exintro&explaintext&redirects=1'
|
||||||
'';
|
'';
|
||||||
|
|
||||||
# Launcher LLM answers, via the mediaserver. The API key lives on the
|
# Launcher LLM answers, via the mediaserver. Codex uses fred's ChatGPT
|
||||||
# server ONLY — the desktops just pipe a question over the same SSH
|
# login there; desktops only pipe questions over the existing SSH path.
|
||||||
# path the server monitor already uses, so no laptop ever holds a
|
# The remote end is `qs-ask` in hosts/FredOS-Mediaserver.nix.
|
||||||
# credential and there's one place to rotate it. The remote end is
|
|
||||||
# `qs-ask` in hosts/FredOS-Mediaserver.nix.
|
|
||||||
#
|
#
|
||||||
# The query goes over STDIN, not as an argv element: `ssh host cmd arg`
|
# The query goes over STDIN, not as an argv element: `ssh host cmd arg`
|
||||||
# re-parses arg through the remote shell, so a question containing
|
# re-parses arg through the remote shell, so a question containing
|
||||||
# quotes would break. stdin sidesteps remote quoting entirely.
|
# quotes would break. stdin sidesteps remote quoting entirely.
|
||||||
#
|
#
|
||||||
# Prints nothing when the key is missing, the server is unreachable, the
|
# Prints nothing when login is missing, the server is unreachable, the
|
||||||
# call fails, or the model says it doesn't know — each of those falls
|
# call fails, or the model says it doesn't know; Wikipedia then remains.
|
||||||
# through to the Wikipedia answer.
|
|
||||||
askFetchScript = pkgs.writeShellScript "ask-fetch" ''
|
askFetchScript = pkgs.writeShellScript "ask-fetch" ''
|
||||||
q="$1"
|
q="$1"
|
||||||
[ -z "$q" ] && exit 0
|
[ -z "$q" ] && exit 0
|
||||||
|
|
@ -367,8 +364,7 @@ in
|
||||||
'';
|
'';
|
||||||
|
|
||||||
# Launcher weather. Neither of the other two answer paths can serve
|
# Launcher weather. Neither of the other two answer paths can serve
|
||||||
# this: qs-ask runs with no tools at all, so Claude has no live data and
|
# this: qs-ask has no live data, and Wikipedia answers "weather in kristiansand"
|
||||||
# correctly refuses, and Wikipedia answers "weather in kristiansand"
|
|
||||||
# with the city's founding date. open-meteo needs no key — geocode the
|
# with the city's founding date. open-meteo needs no key — geocode the
|
||||||
# place name, then read current conditions at those coordinates.
|
# place name, then read current conditions at those coordinates.
|
||||||
#
|
#
|
||||||
|
|
@ -2025,8 +2021,8 @@ in
|
||||||
// the slow paths count: Wikipedia lands inside the debounce
|
// the slow paths count: Wikipedia lands inside the debounce
|
||||||
// and would just flicker the spinner.
|
// and would just flicker the spinner.
|
||||||
readonly property bool answerPending: askProc.running || wxProc.running
|
readonly property bool answerPending: askProc.running || wxProc.running
|
||||||
readonly property string pendingSource: wxProc.running ? "open-meteo" : "Claude"
|
readonly property string pendingSource: wxProc.running ? "open-meteo" : "ChatGPT"
|
||||||
// Claude answered the current query, so the slower-arriving
|
// ChatGPT answered the current query, so the slower-arriving
|
||||||
// Wikipedia result must not overwrite it.
|
// Wikipedia result must not overwrite it.
|
||||||
property bool askWon: false
|
property bool askWon: false
|
||||||
property var answerCache: ({})
|
property var answerCache: ({})
|
||||||
|
|
@ -2116,10 +2112,8 @@ in
|
||||||
}
|
}
|
||||||
}
|
}
|
||||||
|
|
||||||
// Claude, tried first. The script is the feature gate: with
|
// ChatGPT, tried alongside Wikipedia. Without a Codex login
|
||||||
// no API key it exits immediately and prints nothing, which
|
// the script prints nothing and Wikipedia remains visible.
|
||||||
// is the same signal as a failed call — so the Wikipedia
|
|
||||||
// path stays the default with zero configuration.
|
|
||||||
Process {
|
Process {
|
||||||
id: askProc
|
id: askProc
|
||||||
stdout: StdioCollector {
|
stdout: StdioCollector {
|
||||||
|
|
@ -2172,7 +2166,7 @@ in
|
||||||
return;
|
return;
|
||||||
}
|
}
|
||||||
answerQuery = q;
|
answerQuery = q;
|
||||||
// Weather takes the open-meteo path alone: Claude has no
|
// Weather takes the open-meteo path alone: ChatGPT has no
|
||||||
// network to look it up with, Wikipedia answers with the
|
// network to look it up with, Wikipedia answers with the
|
||||||
// city's founding date, and the cache has no notion of
|
// city's founding date, and the cache has no notion of
|
||||||
// age so it must not serve a temperature twice.
|
// age so it must not serve a temperature twice.
|
||||||
|
|
@ -2191,10 +2185,7 @@ in
|
||||||
return;
|
return;
|
||||||
}
|
}
|
||||||
// Both at once. Wikipedia fills the card in well under a
|
// Both at once. Wikipedia fills the card in well under a
|
||||||
// second; Claude takes ~6s through Claude Code and
|
// second; ChatGPT supersedes it when Codex finishes.
|
||||||
// supersedes it on arrival. Racing them beats chaining:
|
|
||||||
// Claude-then-fallback would leave the card blank for
|
|
||||||
// six seconds on every question.
|
|
||||||
askWon = false;
|
askWon = false;
|
||||||
killAsk();
|
killAsk();
|
||||||
askProc.command = [Commands.askFetch, q];
|
askProc.command = [Commands.askFetch, q];
|
||||||
|
|
@ -2218,7 +2209,7 @@ in
|
||||||
fact: text.slice(0, 400),
|
fact: text.slice(0, 400),
|
||||||
lead: "",
|
lead: "",
|
||||||
url: "", // nothing to open
|
url: "", // nothing to open
|
||||||
source: "Claude"
|
source: "ChatGPT"
|
||||||
};
|
};
|
||||||
cacheAnswer(res);
|
cacheAnswer(res);
|
||||||
// Typing may have moved on while the call was out.
|
// Typing may have moved on while the call was out.
|
||||||
|
|
@ -2365,7 +2356,7 @@ in
|
||||||
: "https://en.wikipedia.org/wiki/" + encodeURIComponent(title.replace(/ /g, "_")),
|
: "https://en.wikipedia.org/wiki/" + encodeURIComponent(title.replace(/ /g, "_")),
|
||||||
source: "Wikipedia"
|
source: "Wikipedia"
|
||||||
};
|
};
|
||||||
// Claude already answered this query — don't overwrite
|
// ChatGPT already answered this query — don't overwrite
|
||||||
// the better answer in the cache or on screen.
|
// the better answer in the cache or on screen.
|
||||||
if (askWon) return;
|
if (askWon) return;
|
||||||
cacheAnswer(res);
|
cacheAnswer(res);
|
||||||
|
|
@ -2801,7 +2792,7 @@ in
|
||||||
anchors.rightMargin: 14
|
anchors.rightMargin: 14
|
||||||
spacing: 4
|
spacing: 4
|
||||||
|
|
||||||
// Article title — Wikipedia only; a Claude
|
// Article title — Wikipedia only; a ChatGPT
|
||||||
// answer has no page behind it.
|
// answer has no page behind it.
|
||||||
Row {
|
Row {
|
||||||
spacing: 8
|
spacing: 8
|
||||||
|
|
@ -2842,12 +2833,12 @@ in
|
||||||
|
|
||||||
// Always say who answered — the two sources
|
// Always say who answered — the two sources
|
||||||
// carry very different trust. The spinner
|
// carry very different trust. The spinner
|
||||||
// runs for as long as Claude's ~6s call is
|
// runs for as long as the Codex call is
|
||||||
// out, which is the only cue that a
|
// out, which is the only cue that a
|
||||||
// Wikipedia answer already on screen is
|
// Wikipedia answer already on screen is
|
||||||
// still provisional — and that a card that
|
// still provisional — and that a card that
|
||||||
// never changes was Claude declining, not
|
// never changes was ChatGPT declining or
|
||||||
// Claude being slow.
|
// the call failing.
|
||||||
Row {
|
Row {
|
||||||
spacing: 5
|
spacing: 5
|
||||||
SIcon {
|
SIcon {
|
||||||
|
|
|
||||||
Loading…
Add table
Add a link
Reference in a new issue