mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-07-31 04:12:10 +03:00
* chore(release): open v3.8.34 development cycle * chore(quality): release-green pre-flight validator + nightly signal (C+D) (#4622) C — scripts/quality/validate-release-green.mjs (npm run check:release-green): reproduces the release-equivalent validation (typecheck, eslint, db-rules, public-creds, full unit, vitest, ratchets, optional --with-build package-artifact) against the current working tree and classifies each red as HARD (real defect, exit 1) vs DRIFT (ratchet — reported, never affects exit / never blocks). Pure helpers exported + orchestration behind a direct-run guard; unit-tested. D — .github/workflows/nightly-release-green.yml: runs C on the active release branch nightly (and on workflow_dispatch) and opens/updates a single tracking issue on HARD failures. Never a required check, never touches a contributor PR. Closes the gap where the full gate (ci.yml) only ran on the release PR, so reds accrued silently on release/** and surfaced in 40-min layers at release time. Non-blocking by construction; drift is the maintainer's to rebaseline at release. Co-authored-by: Diego Rodrigues de Sa e Souza <diego.souza@cdwasolutions.com.br> * fix(providers): show revealed connection API keys (#4583) Integrated into release/v3.8.34 * fix(resilience): respect upstream retry hint toggle (#4585) Integrated into release/v3.8.34 * feat(settings): expose stream recovery feature flags (#4586) Integrated into release/v3.8.34 * fix(logs): make active request stale sweep configurable (#4599) Integrated into release/v3.8.34 * fix(plugin): auto-prefix providerId with 'opencode-' for OC 1.17.8+ native gate (#4527) Integrated into release/v3.8.34 (supersedes #4445) * fix(models): treat unknown output caps as unset (#4584) Integrated into release/v3.8.34 * fix(executors): strip temperature for GitHub Copilot gpt-5.4 family (#4564) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(oauth): update Qwen OAuth URLs from chat.qwen.ai to qwen.ai (#4561) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(api/settings): prevent cached /api/settings responses (port from 9router#951) (#4566) Integrated into release/v3.8.34 (rebuilt onto tip) * feat(audio): MiniMax T2A v2 TTS dispatch in audioSpeech (port #1043) (#4553) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(dashboard): surface manual config CTA when Open Claw CLI auto-detect fails (#4562) Integrated into release/v3.8.34 (rebuilt onto tip) * feat(providers): optional model ID for custom API-key validation (#4555) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(cli): align data dir and env loading with runtime (#4607) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(quota): expose Bailian quota windows (#4610) Integrated into release/v3.8.34 (rebuilt onto tip) * fix: retain provider cooldowns for configured max window (#4588) Integrated into release/v3.8.34 (rebuilt — bundled commits stripped) * fix: reject invalid provider cooldown bounds (#4589) Integrated into release/v3.8.34 (rebuilt — bundled commits stripped) * fix: preserve production combo metrics on shadow eviction (#4590) Integrated into release/v3.8.34 (rebuilt — bundled commits stripped) * fix(stream): estimate input tokens when upstream reports prompt_tokens=0 (#4615) Integrated into release/v3.8.34 (rebuilt onto tip) * fix(catalog): shorten no-thinking gateway prefix to no-think/ (#4525) Integrated into release/v3.8.34 (rebuilt — kept only the prefix rename, dropped stale-base reverts) * fix(relay): apply IP rate limit to bifrost sidecar (#4593) Integrated into release/v3.8.34 (rebuilt onto tip; merge before #4612) * fix(bifrost): finalize SSE relay usage after stream (#4612) Integrated into release/v3.8.34 (rebuilt + reconciled with #4593) * feat(compression): per-request `x-omniroute-compression` header (Phase 3) (#4645) * docs(compression): Phase 3 per-request header design spec Approved brainstorming output for the x-omniroute-compression header: header-first precedence, name-first combo matching (Decision A), explicit value bypasses auto-trigger (Decision B), DerivedPlan.source, and the X-OmniRoute-Compression response header. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(compression): Phase 3 per-request header implementation plan 4-task TDD plan (resolver header-first + source, parser, chatCore wiring + response header, docs/file-size) with full code and exact commands. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(compression): header-first resolver + plan source (Phase 3 core) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(compression): resolveCompressionHeader parser (Phase 3) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * feat(compression): wire x-omniroute-compression header + response header (Phase 3) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * refactor(compression): extract plan-resolution leaf (planResolution.ts) under size cap (Phase 3) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * docs(compression): document x-omniroute-compression header (Phase 3) Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> * fix(compression): harden named-combo map + trim engine: header id (Phase 3 review) Addresses gemini-code-assist review on #4645: - Extract buildNamedComboLookup (pure) so a blank/whitespace/null combo name contributes only its id key (no '' key, no throw that disables all combos). - Trim the engine:<id> header value so 'engine: rtk' resolves. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> --------- Co-authored-by: Diego Rodrigues de Sa e Souza <diego.souza@cdwasolutions.com.br> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Diego Rodrigues de Sa e Souza <souzamiriamrodrigues790@gmail.com> * fix: exclude exhausted connections from auto scoring (#4592) Integrated into release/v3.8.34 (rebuilt + opt-in gate fix) * fix(dashboard): memoize compatible provider groups (#4613) Integrated into release/v3.8.34 (rebuilt + test added) * fix(dashboard): isolate quota widget refresh clock (#4611) Integrated into release/v3.8.34 (rebuilt + jsdom test) * fix(dashboard): gate topology side effects behind widget visibility (#4606) Integrated into release/v3.8.34 (rebuilt + jsdom test) * fix(dashboard): keep play_arrow spinning on provider Test All buttons (#4563) Integrated into release/v3.8.34 (rebuilt onto tip; UI-cosmetic per owner) * fix(db): schedule retention cleanup + fix cleanup table/column names (extracted from #4428) (#4691) Integrated into release/v3.8.34 (cleanup core extracted from #4428, credit @oyi77) * fix(telemetry): back off live-WS event forwarding when the sidecar is unreachable (#4604) (#4687) Co-authored-by: Diego Rodrigues de Sa e Souza <souzamiriamrodrigues790@gmail.com> * fix(api): serve GET /v1/models/{model} as JSON, not the HTML dashboard (#4674) (#4677) Co-authored-by: Diego Rodrigues de Sa e Souza <souzamiriamrodrigues790@gmail.com> * feat(opencode): add go deepseek reasoning variants (#4647) Integrated into release/v3.8.34 * fix(executors): robust deepseek-web tool-call parsing and agentic context retention (#4644) Integrated into release/v3.8.34 * fix(cli): authenticate `omniroute logs` and honor active context (#4638) Integrated into release/v3.8.34 (authored by Rahul Sharma, AI co-author trailer stripped per project policy) * fix(proxy): apply pipelining:0 + connections cap to the direct dispatcher (#4580) (#4684) Co-authored-by: Diego Rodrigues de Sa e Souza <souzamiriamrodrigues790@gmail.com> * fix(executors): Firecrawl web_fetch 500 with include_metadata=true (#4692) Integrated into release/v3.8.34 * fix(routing): include all noAuth models in auto-combos + add reka-flash + best-free template (#4621) Integrated into release/v3.8.34 (dead getFirstRegistryModelId dropped, rebuilt onto tip) * fix(dashboard): gate home topology live-WS networking (#4596) (#4618) Integrated into release/v3.8.34 (adapted onto #4606's extracted topology section: default-hidden flip + enabled gate on useLiveDashboard) * fix(cli): align `omniroute` env loading with the runtime data dir (#4597) (#4619) Integrated into release/v3.8.34 (data-dir.mjs refactor reconciled with #4607; loadEnvFile aligned to getDefaultDataDir) * chore(quality): reconcile file-size baseline for #4644 (deepseek-web.ts 1117->1125) (#4695) file-size reconcile for #4644 * Support quota scraping for OpenCode Go and Ollama Cloud (#4642) Integrated into release/v3.8.34 (Ollama Cloud + OpenCode Go dashboard quota scraping; rebuilt onto tip, gates green: typecheck/public-creds/file-size/lint/docs-sync + 31 tests) * feat(executors): land M365 Copilot pure framing + connection helpers (#4042) (#4696) Land M365 pure modules ahead of draft #4400 * deps: bump production + development groups; migrate js-yaml to v5 ESM (#4697) Incorporates Dependabot #4667 + #4668 + js-yaml v5 ESM migration into release/v3.8.34 * fix: noAuth provider validation + kimi executor routing (#4699) Integrated into release/v3.8.34 (noAuth in NOAUTH_PROVIDERS dynamic check + remove misrouted kimi web alias; 9 tests) * refactor(imageGeneration): extract 8 provider families to co-located files (#4609) Integrated into release/v3.8.34 (extraction completed: added missing imports/exports per module, main imports handlers locally; 145 image-gen tests pass, typecheck/cycles/file-size green) * chore(release): v3.8.34 — finalize changelog, rebaseline drift, fix release-green reds - Finalize CHANGELOG [3.8.34] (43 bullets, full contributor attribution) + seed i18n mirrors - Rebaseline inherited cycle drift surfaced by release-green pre-flight: eslint warnings 3900->3907, cognitive-complexity 797->801 (release-finalize touches no prod code; all drift is from this cycle's contributor merges) - fix(providers): keep reka-flash-3 as the Reka provider default. #4621 inserted reka-flash at the head of the model list, silently changing the default from reka-flash-3 (the free-tier model) to reka-flash; reorder so reka-flash-3 stays default, reka-flash retained. - test: align provider-models-config / provider-models-route / web-cookie-providers-new with #4621 (reka-flash now in the Reka catalog) and #4699 (the `kimi` API-key provider correctly falls through to DefaultExecutor instead of KimiWebExecutor) - chore(quality): allowlist the COMPRESSION_GUIDE doc name in check-fabricated-docs (false-positive env-var match; docs/compression/COMPRESSION_GUIDE.md exists) * fix(release-green): resolve release-PR full-CI reds for v3.8.34 Surfaced only on the release PR (these gates don't run on PR->release fast-gates): - fix(quota): complete HTML-comment sanitization in opencodeOllamaUsage SSR reset-time parsing — strip any <!--...--> generically instead of the two literal React hydration markers, so no partial "<!--" can survive (CodeQL js/incomplete-multi-character- sanitization, HIGH, introduced by #4642). Regression test added. - test(codex): correct the Codex-fingerprint body key order assertion to match the canonical bodyFieldOrder (prompt_cache_key precedes include); #4584 flipped the two and integration tests don't run on fast-gates so it never executed until the release PR. - chore(quality): rebaseline inherited cycle drift surfaced by full CI — zizmorFindings 152->155 (+3 unpinned-uses in nightly-release-green.yml from #4622, same @vN convention as ci.yml) and openapiCoverage.pct 38.4->37.8 (-0.6, contributor routes added faster than openapi docs). Release-finalize touches no prod routes. * fix(release-green): complete CodeQL sanitization + rebaseline complexity drift - fix(quota): handle unterminated HTML comments in opencodeOllamaUsage SSR reset-time parsing — the `(?:-->|$)` arm consumes a trailing "<!--" with no closing "-->", so no partial "<!--" can survive (CodeQL js/incomplete-multi-character-sanitization persisted with the plain <!--...--> form because an unclosed comment could still leave "<!--"). - chore(quality): rebaseline cyclomatic complexity 1915->1916 (+1) — inherited v3.8.34 cycle drift (contributor feature branches); check:complexity does not run on PR->release fast-gates so it surfaced only on the release PR. Release-finalize adds 0 complexity (measured 1916 with/without the regex tweak). dead-code/cognitive/type-coverage/ compression-budget/codeql ratchets all pass. --------- Co-authored-by: Diego Rodrigues de Sa e Souza <diego.souza@cdwasolutions.com.br> Co-authored-by: Randi <55005611+rdself@users.noreply.github.com> Co-authored-by: Hernan Javier Ardila Sanchez <hjasgr@gmail.com> Co-authored-by: KooshaPari <42529354+KooshaPari@users.noreply.github.com> Co-authored-by: Abhishek Divekar <adivekar@utexas.edu> Co-authored-by: Rahul sharma <sharmaR0810@gmail.com> Co-authored-by: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Co-authored-by: Diego Rodrigues de Sa e Souza <souzamiriamrodrigues790@gmail.com> Co-authored-by: Ronald Estacion <DevEstacion@users.noreply.github.com> Co-authored-by: Igor <60442260+BugsBag@users.noreply.github.com> Co-authored-by: Oonishi <275808243+ponkcore@users.noreply.github.com> Co-authored-by: Paijo <14921983+oyi77@users.noreply.github.com> Co-authored-by: Jan Leon <Jan.gaschler@gmail.com>
532 lines
15 KiB
TypeScript
532 lines
15 KiB
TypeScript
// Tool-call translation for web-cookie providers (deepseek-web, chatgpt-web, ...).
|
|
//
|
|
// The web UIs accept only a single plain prompt string and have no native function
|
|
// calling — they reply with tool invocations as raw text. To let agentic clients use
|
|
// these providers we (a) serialize the OpenAI `tools` array into a system-prompt
|
|
// contract on the request side, and (b) parse the upstream `<tool>{...}</tool>` text
|
|
// back into OpenAI `tool_calls` on the response side. (#2820)
|
|
|
|
export interface OpenAIToolCall {
|
|
id: string;
|
|
type: "function";
|
|
function: { name: string; arguments: string };
|
|
}
|
|
|
|
interface OpenAIToolDef {
|
|
type?: string;
|
|
function?: {
|
|
name?: string;
|
|
description?: string;
|
|
parameters?: unknown;
|
|
};
|
|
}
|
|
|
|
const TOOL_BLOCK_RE = /<tool>\s*([\s\S]*?)\s*<\/tool>/g;
|
|
// Some web-cookie models (e.g. ds-web) wrap calls as `<tool_call name="...">{json}</tool_call>`
|
|
// instead of the canonical `<tool>{json}</tool>`. Capture the JSON body — the real tool name
|
|
// lives there, never in the tag's `name="..."` attribute (#3260).
|
|
const TOOL_CALL_TAG_RE = /<tool_call(?:\s+[^>]*)?\s*>\s*([\s\S]*?)\s*<\/tool_call>/g;
|
|
|
|
interface ToolParseCandidate {
|
|
raw: string;
|
|
start: number;
|
|
end: number;
|
|
requireRequestedTool: boolean;
|
|
}
|
|
|
|
export interface RequestedToolName {
|
|
original: string;
|
|
normalized: string;
|
|
}
|
|
|
|
function toRecord(value: unknown): Record<string, unknown> | null {
|
|
return value && typeof value === "object" && !Array.isArray(value)
|
|
? (value as Record<string, unknown>)
|
|
: null;
|
|
}
|
|
|
|
export function getRequestedToolNames(tools: unknown): RequestedToolName[] {
|
|
if (!Array.isArray(tools)) return [];
|
|
const names: RequestedToolName[] = [];
|
|
const seen = new Set<string>();
|
|
for (const tool of tools) {
|
|
const record = toRecord(tool);
|
|
const fn = toRecord(record?.function);
|
|
const name = typeof fn?.name === "string" ? fn.name.trim() : "";
|
|
if (!name || seen.has(name)) continue;
|
|
seen.add(name);
|
|
names.push({ original: name, normalized: normalizeToolName(name) });
|
|
}
|
|
return names;
|
|
}
|
|
|
|
function normalizeToolName(name: string): string {
|
|
return name.toLowerCase().replace(/[^a-z0-9]/g, "");
|
|
}
|
|
|
|
function levenshteinDistance(a: string, b: string): number {
|
|
if (a === b) return 0;
|
|
if (!a) return b.length;
|
|
if (!b) return a.length;
|
|
let previous = Array.from({ length: b.length + 1 }, (_, i) => i);
|
|
let current = Array<number>(b.length + 1);
|
|
for (let i = 1; i <= a.length; i += 1) {
|
|
current[0] = i;
|
|
for (let j = 1; j <= b.length; j += 1) {
|
|
const cost = a[i - 1] === b[j - 1] ? 0 : 1;
|
|
current[j] = Math.min(current[j - 1] + 1, previous[j] + 1, previous[j - 1] + cost);
|
|
}
|
|
const temp = previous;
|
|
previous = current;
|
|
current = temp;
|
|
}
|
|
return previous[b.length];
|
|
}
|
|
|
|
function scoreToolName(emitted: string, requested: RequestedToolName): number {
|
|
if (emitted === requested.original) return 1;
|
|
const normalized = normalizeToolName(emitted);
|
|
if (!normalized || !requested.normalized) return 0;
|
|
if (normalized === requested.normalized) return 0.98;
|
|
|
|
const shorter = Math.min(normalized.length, requested.normalized.length);
|
|
const longer = Math.max(normalized.length, requested.normalized.length);
|
|
if (shorter >= 4) {
|
|
if (normalized.includes(requested.normalized) || requested.normalized.includes(normalized)) {
|
|
return 0.86 - (longer - shorter) / Math.max(longer, 1) / 4;
|
|
}
|
|
}
|
|
|
|
const distance = levenshteinDistance(normalized, requested.normalized);
|
|
const similarity = 1 - distance / Math.max(longer, 1);
|
|
return similarity >= 0.72 ? similarity : 0;
|
|
}
|
|
|
|
export function resolveRequestedToolName(
|
|
emitted: string,
|
|
requestedTools: RequestedToolName[]
|
|
): string | null {
|
|
if (requestedTools.length === 0) return emitted;
|
|
|
|
let best: { name: string; score: number } | null = null;
|
|
let secondBest = 0;
|
|
for (const requested of requestedTools) {
|
|
const score = scoreToolName(emitted, requested);
|
|
if (!best || score > best.score) {
|
|
secondBest = best?.score ?? 0;
|
|
best = { name: requested.original, score };
|
|
} else if (score > secondBest) {
|
|
secondBest = score;
|
|
}
|
|
}
|
|
|
|
if (!best || best.score < 0.72) return null;
|
|
// Avoid correcting to an arbitrary tool when the fuzzy match is ambiguous.
|
|
if (best.score < 0.98 && best.score - secondBest < 0.08) return null;
|
|
return best.name;
|
|
}
|
|
|
|
function stripCodeFence(value: string): string {
|
|
return value
|
|
.trim()
|
|
.replace(/^```(?:json|javascript|js|python)?\s*/i, "")
|
|
.replace(/\s*```$/i, "")
|
|
.trim();
|
|
}
|
|
|
|
function convertSingleQuotedStrings(value: string): string {
|
|
let result = "";
|
|
let inSingle = false;
|
|
let inDouble = false;
|
|
let escaped = false;
|
|
|
|
for (const ch of value) {
|
|
if (escaped) {
|
|
result += ch === '"' && inSingle ? '\\"' : ch;
|
|
escaped = false;
|
|
continue;
|
|
}
|
|
|
|
if (ch === "\\") {
|
|
result += ch;
|
|
escaped = true;
|
|
continue;
|
|
}
|
|
|
|
if (ch === '"') {
|
|
if (inSingle) {
|
|
result += '\\"';
|
|
} else {
|
|
inDouble = !inDouble;
|
|
result += ch;
|
|
}
|
|
continue;
|
|
}
|
|
|
|
if (ch === "'" && !inDouble) {
|
|
inSingle = !inSingle;
|
|
result += '"';
|
|
continue;
|
|
}
|
|
|
|
result += ch;
|
|
}
|
|
|
|
return result;
|
|
}
|
|
|
|
function replacePythonLiterals(value: string): string {
|
|
let result = "";
|
|
let inString = false;
|
|
let escaped = false;
|
|
let token = "";
|
|
|
|
const flushToken = () => {
|
|
if (token === "True") result += "true";
|
|
else if (token === "False") result += "false";
|
|
else if (token === "None") result += "null";
|
|
else result += token;
|
|
token = "";
|
|
};
|
|
|
|
for (const ch of value) {
|
|
if (escaped) {
|
|
if (token) flushToken();
|
|
result += ch;
|
|
escaped = false;
|
|
continue;
|
|
}
|
|
|
|
if (ch === "\\") {
|
|
if (token) flushToken();
|
|
result += ch;
|
|
escaped = inString;
|
|
continue;
|
|
}
|
|
|
|
if (ch === '"') {
|
|
if (token) flushToken();
|
|
inString = !inString;
|
|
result += ch;
|
|
continue;
|
|
}
|
|
|
|
if (!inString && /[A-Za-z]/.test(ch)) {
|
|
token += ch;
|
|
continue;
|
|
}
|
|
|
|
if (token) flushToken();
|
|
result += ch;
|
|
}
|
|
|
|
if (token) flushToken();
|
|
return result;
|
|
}
|
|
|
|
function normalizeLooseJson(value: string): string {
|
|
return replacePythonLiterals(convertSingleQuotedStrings(value))
|
|
.replace(/([{,]\s*)([A-Za-z_][A-Za-z0-9_-]*)(\s*:)/g, '$1"$2"$3')
|
|
.replace(/,\s*([}\]])/g, "$1");
|
|
}
|
|
|
|
export function parseLooseJsonObject(raw: string): Record<string, unknown> | null {
|
|
const trimmed = stripCodeFence(raw);
|
|
for (const candidate of [trimmed, normalizeLooseJson(trimmed)]) {
|
|
try {
|
|
return toRecord(JSON.parse(candidate));
|
|
} catch {
|
|
// Try the next, more permissive form.
|
|
}
|
|
}
|
|
return null;
|
|
}
|
|
|
|
function findBareJsonCandidates(text: string): ToolParseCandidate[] {
|
|
const candidates: ToolParseCandidate[] = [];
|
|
let start = -1;
|
|
let depth = 0;
|
|
let quote: '"' | "'" | "" = "";
|
|
let escaped = false;
|
|
|
|
for (let i = 0; i < text.length; i += 1) {
|
|
const ch = text[i];
|
|
|
|
if (depth === 0 && ch !== "{") {
|
|
continue;
|
|
}
|
|
|
|
if (escaped) {
|
|
escaped = false;
|
|
continue;
|
|
}
|
|
|
|
if (quote) {
|
|
if (ch === "\\") {
|
|
escaped = true;
|
|
} else if (ch === quote) {
|
|
quote = "";
|
|
}
|
|
continue;
|
|
}
|
|
|
|
if (ch === '"' || ch === "'") {
|
|
quote = ch;
|
|
continue;
|
|
}
|
|
|
|
if (ch === "{") {
|
|
if (depth === 0) start = i;
|
|
depth += 1;
|
|
continue;
|
|
}
|
|
|
|
if (ch === "}" && depth > 0) {
|
|
depth -= 1;
|
|
if (depth === 0 && start >= 0) {
|
|
const raw = text.slice(start, i + 1);
|
|
if (
|
|
/[{,]\s*["']?(name|command)["']?\s*:/i.test(raw) &&
|
|
/[{,]\s*["']?arguments["']?\s*:/i.test(raw)
|
|
) {
|
|
candidates.push({ raw, start, end: i + 1, requireRequestedTool: true });
|
|
}
|
|
start = -1;
|
|
}
|
|
}
|
|
}
|
|
|
|
return candidates;
|
|
}
|
|
|
|
function rangesOverlap(
|
|
a: { start: number; end: number },
|
|
b: { start: number; end: number }
|
|
): boolean {
|
|
return a.start < b.end && b.start < a.end;
|
|
}
|
|
|
|
export function stripRanges(text: string, ranges: Array<{ start: number; end: number }>): string {
|
|
let content = text;
|
|
const sorted = [...ranges].sort((a, b) => b.start - a.start);
|
|
for (const range of sorted) {
|
|
const lineStart = content.lastIndexOf("\n", range.start - 1) + 1;
|
|
const nextLineBreak = content.indexOf("\n", range.end);
|
|
const lineEnd = nextLineBreak === -1 ? content.length : nextLineBreak;
|
|
const beforeOnLine = content.slice(lineStart, range.start);
|
|
const afterOnLine = content.slice(range.end, lineEnd);
|
|
const removeWholeLine = beforeOnLine.trim() === "" && afterOnLine.trim() === "";
|
|
const start = removeWholeLine ? lineStart : range.start;
|
|
const end =
|
|
removeWholeLine && nextLineBreak !== -1
|
|
? nextLineBreak + 1
|
|
: removeWholeLine
|
|
? lineEnd
|
|
: range.end;
|
|
content = `${content.slice(0, start)}${content.slice(end)}`;
|
|
}
|
|
return content.replace(/\n{3,}/g, "\n\n").trim();
|
|
}
|
|
|
|
export function toArgumentsString(value: unknown): string {
|
|
if (value === undefined) return "{}";
|
|
if (typeof value === "string") {
|
|
const parsed = parseLooseJsonObject(value);
|
|
return parsed ? JSON.stringify(parsed) : value;
|
|
}
|
|
try {
|
|
return JSON.stringify(value);
|
|
} catch {
|
|
return "{}";
|
|
}
|
|
}
|
|
|
|
/**
|
|
* Serialize an OpenAI `tools` array into a system-prompt block that instructs the
|
|
* web UI model how to invoke a tool (emit a `<tool>{...}</tool>` block). Returns an
|
|
* empty string when there are no usable tools.
|
|
*/
|
|
export function serializeToolsToPrompt(tools: unknown): string {
|
|
if (!Array.isArray(tools) || tools.length === 0) return "";
|
|
|
|
const lines: string[] = [];
|
|
for (const t of tools as OpenAIToolDef[]) {
|
|
const fn = t?.function;
|
|
if (!fn?.name) continue;
|
|
const desc = typeof fn.description === "string" && fn.description ? fn.description : "";
|
|
let params = "";
|
|
try {
|
|
params = fn.parameters ? JSON.stringify(fn.parameters) : "";
|
|
} catch {
|
|
params = "";
|
|
}
|
|
lines.push(
|
|
`- ${fn.name}${desc ? `: ${desc}` : ""}${params ? `\n parameters: ${params}` : ""}`
|
|
);
|
|
}
|
|
|
|
if (lines.length === 0) return "";
|
|
|
|
return [
|
|
"You can call tools. To call a tool, reply with a single line containing a <tool> block",
|
|
'with JSON: <tool>{"name": "<tool_name>", "arguments": { ... }}</tool>',
|
|
"Only emit the <tool> block when you actually want to call a tool; otherwise answer normally.",
|
|
"",
|
|
"Available tools:",
|
|
...lines,
|
|
].join("\n");
|
|
}
|
|
|
|
/**
|
|
* Parse `<tool>{...}</tool>` blocks out of upstream text into OpenAI `tool_calls`.
|
|
* When a requested `tools[]` set is provided, also accepts bare JSON tool-call
|
|
* objects emitted by web models that ignored the `<tool>` wrapper contract.
|
|
* Returns the content with the blocks stripped, plus the tool calls (or null when
|
|
* there are none). `arguments` is always a JSON *string*, matching the OpenAI API.
|
|
*
|
|
* `idSeed` makes generated ids deterministic for callers that need stability; when
|
|
* omitted, ids are still unique within a single call (index-based).
|
|
*/
|
|
export function parseToolCallsFromText(
|
|
text: string,
|
|
idSeed = "call",
|
|
requestedTools?: unknown
|
|
): { content: string; toolCalls: OpenAIToolCall[] | null } {
|
|
const requestedToolNames = getRequestedToolNames(requestedTools);
|
|
const canParseBareJson = requestedToolNames.length > 0;
|
|
if (
|
|
typeof text !== "string" ||
|
|
(!text.includes("<tool>") && !text.includes("<tool_call") && !canParseBareJson)
|
|
) {
|
|
return { content: text ?? "", toolCalls: null };
|
|
}
|
|
|
|
const candidates: ToolParseCandidate[] = [];
|
|
const toolBlockRanges: Array<{ start: number; end: number }> = [];
|
|
|
|
let blockMatch: RegExpExecArray | null;
|
|
TOOL_BLOCK_RE.lastIndex = 0;
|
|
while ((blockMatch = TOOL_BLOCK_RE.exec(text)) !== null) {
|
|
const range = { start: blockMatch.index, end: TOOL_BLOCK_RE.lastIndex };
|
|
toolBlockRanges.push(range);
|
|
candidates.push({
|
|
raw: blockMatch[1].trim(),
|
|
start: range.start,
|
|
end: range.end,
|
|
requireRequestedTool: false,
|
|
});
|
|
}
|
|
|
|
TOOL_CALL_TAG_RE.lastIndex = 0;
|
|
while ((blockMatch = TOOL_CALL_TAG_RE.exec(text)) !== null) {
|
|
const range = { start: blockMatch.index, end: TOOL_CALL_TAG_RE.lastIndex };
|
|
toolBlockRanges.push(range);
|
|
candidates.push({
|
|
raw: blockMatch[1].trim(),
|
|
start: range.start,
|
|
end: range.end,
|
|
requireRequestedTool: false,
|
|
});
|
|
}
|
|
|
|
if (canParseBareJson) {
|
|
for (const candidate of findBareJsonCandidates(text)) {
|
|
if (!toolBlockRanges.some((range) => rangesOverlap(range, candidate))) {
|
|
candidates.push(candidate);
|
|
}
|
|
}
|
|
}
|
|
|
|
candidates.sort((a, b) => a.start - b.start);
|
|
|
|
const toolCalls: OpenAIToolCall[] = [];
|
|
const acceptedRanges: Array<{ start: number; end: number }> = [];
|
|
for (const candidate of candidates) {
|
|
const parsed = parseLooseJsonObject(candidate.raw);
|
|
const emittedName =
|
|
parsed && typeof parsed.name === "string"
|
|
? parsed.name
|
|
: parsed && typeof parsed.command === "string"
|
|
? parsed.command
|
|
: null;
|
|
if (!emittedName) continue;
|
|
const name =
|
|
resolveRequestedToolName(emittedName, requestedToolNames) ||
|
|
(candidate.requireRequestedTool ? null : emittedName);
|
|
if (!name || (candidate.requireRequestedTool && requestedToolNames.length === 0)) continue;
|
|
const args = toArgumentsString(parsed?.arguments);
|
|
toolCalls.push({
|
|
id: `${idSeed}_${toolCalls.length}`,
|
|
type: "function",
|
|
function: { name, arguments: args },
|
|
});
|
|
acceptedRanges.push({ start: candidate.start, end: candidate.end });
|
|
}
|
|
|
|
if (toolCalls.length === 0) {
|
|
return { content: text, toolCalls: null };
|
|
}
|
|
|
|
const content = stripRanges(text, acceptedRanges);
|
|
return { content, toolCalls };
|
|
}
|
|
|
|
// ── Shared helpers for web-cookie executors ────────────────────────────────
|
|
|
|
interface ToolPrepResult {
|
|
hasTools: boolean;
|
|
requestedTools: unknown;
|
|
effectiveMessages: Array<{ role: string; content: unknown }>;
|
|
}
|
|
|
|
/**
|
|
* Extract tools from an OpenAI request body and prepend a tool-system-prompt
|
|
* to the messages array when tools are present. Every web-cookie executor
|
|
* that wants tool-call support calls this once before building its upstream
|
|
* request body.
|
|
*/
|
|
export function prepareToolMessages(
|
|
bodyObj: Record<string, unknown>,
|
|
messages: Array<{ role: string; content: unknown }>
|
|
): ToolPrepResult {
|
|
const requestedTools = bodyObj.tools;
|
|
const hasTools = Array.isArray(requestedTools) && requestedTools.length > 0;
|
|
if (!hasTools) return { hasTools: false, requestedTools, effectiveMessages: messages };
|
|
|
|
const toolPrompt = serializeToolsToPrompt(requestedTools);
|
|
return {
|
|
hasTools: true,
|
|
requestedTools,
|
|
effectiveMessages: [{ role: "system", content: toolPrompt }, ...messages],
|
|
};
|
|
}
|
|
|
|
interface ToolCompletionResult {
|
|
content: string;
|
|
toolCalls: OpenAIToolCall[] | null;
|
|
finishReason: string;
|
|
}
|
|
|
|
/**
|
|
* Parse tool calls from a model's text response. Returns the cleaned content
|
|
* (with `<tool>` blocks stripped), the parsed tool calls (or null), and the
|
|
* appropriate finish_reason. Every web-cookie executor calls this on the
|
|
* collected response text when `hasTools` is true.
|
|
*/
|
|
export function buildToolAwareResult(
|
|
rawContent: string,
|
|
requestedTools: unknown,
|
|
idSeed = "call"
|
|
): ToolCompletionResult {
|
|
const { content, toolCalls } = parseToolCallsFromText(
|
|
rawContent,
|
|
`${idSeed}-${Date.now()}`,
|
|
requestedTools
|
|
);
|
|
return {
|
|
content,
|
|
toolCalls,
|
|
finishReason: toolCalls ? "tool_calls" : "stop",
|
|
};
|
|
}
|