Sprint 6d — Migrate Gemini Live to @google/genai SDK

feat(geminiLive): rewrite with GoogleGenAI SDK (vertexai: true, apiKey)
  replaces raw WebSocket to generativelanguage.googleapis.com
feat(geminiLive): restore full setup config (systemInstruction,
  inputAudioTranscription, outputAudioTranscription, VAD)
fix(geminiLive): buildSetupFrame → SDK config object (no manual JSON)
fix(useT2LiveSession): cancelTokenRef for idempotent startDialogue,
  closeAllRef for stable unmount cleanup
chore: add @google/genai@^1.50.1 dependency
test: 11 geminiLive tests rewritten with SDK mock
  292/292 backend tests green
This commit is contained in:
Hermann_Kitio 2026-04-27 02:25:58 +03:00
parent d89b0b1e89
commit 0662e766d4
6 changed files with 970 additions and 331 deletions

415
package-lock.json generated
View file

@ -8,6 +8,7 @@
"name": "expria-backend",
"version": "1.0.0",
"dependencies": {
"@google/genai": "^1.50.1",
"@hono/node-server": "^1.13.7",
"@hono/node-ws": "^1.3.0",
"@supabase/supabase-js": "^2.49.4",
@ -541,6 +542,29 @@
"node": ">=18"
}
},
"node_modules/@google/genai": {
"version": "1.50.1",
"resolved": "https://registry.npmjs.org/@google/genai/-/genai-1.50.1.tgz",
"integrity": "sha512-YbkX7H9+1Pt8wOt7DDREy8XSoiL6fRDzZQRyaVBarFf8MR3zHGqVdvM4cLbDXqPhxqvegZShgfxb8kw9C7YhAQ==",
"license": "Apache-2.0",
"dependencies": {
"google-auth-library": "^10.3.0",
"p-retry": "^4.6.2",
"protobufjs": "^7.5.4",
"ws": "^8.18.0"
},
"engines": {
"node": ">=20.0.0"
},
"peerDependencies": {
"@modelcontextprotocol/sdk": "^1.25.2"
},
"peerDependenciesMeta": {
"@modelcontextprotocol/sdk": {
"optional": true
}
}
},
"node_modules/@hono/node-server": {
"version": "1.19.14",
"resolved": "https://registry.npmjs.org/@hono/node-server/-/node-server-1.19.14.tgz",
@ -647,6 +671,70 @@
"node": ">=14"
}
},
"node_modules/@protobufjs/aspromise": {
"version": "1.1.2",
"resolved": "https://registry.npmjs.org/@protobufjs/aspromise/-/aspromise-1.1.2.tgz",
"integrity": "sha512-j+gKExEuLmKwvz3OgROXtrJ2UG2x8Ch2YZUxahh+s1F2HZ+wAceUNLkvy6zKCPVRkU++ZWQrdxsUeQXmcg4uoQ==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/base64": {
"version": "1.1.2",
"resolved": "https://registry.npmjs.org/@protobufjs/base64/-/base64-1.1.2.tgz",
"integrity": "sha512-AZkcAA5vnN/v4PDqKyMR5lx7hZttPDgClv83E//FMNhR2TMcLUhfRUBHCmSl0oi9zMgDDqRUJkSxO3wm85+XLg==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/codegen": {
"version": "2.0.4",
"resolved": "https://registry.npmjs.org/@protobufjs/codegen/-/codegen-2.0.4.tgz",
"integrity": "sha512-YyFaikqM5sH0ziFZCN3xDC7zeGaB/d0IUb9CATugHWbd1FRFwWwt4ld4OYMPWu5a3Xe01mGAULCdqhMlPl29Jg==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/eventemitter": {
"version": "1.1.0",
"resolved": "https://registry.npmjs.org/@protobufjs/eventemitter/-/eventemitter-1.1.0.tgz",
"integrity": "sha512-j9ednRT81vYJ9OfVuXG6ERSTdEL1xVsNgqpkxMsbIabzSo3goCjDIveeGv5d03om39ML71RdmrGNjG5SReBP/Q==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/fetch": {
"version": "1.1.0",
"resolved": "https://registry.npmjs.org/@protobufjs/fetch/-/fetch-1.1.0.tgz",
"integrity": "sha512-lljVXpqXebpsijW71PZaCYeIcE5on1w5DlQy5WH6GLbFryLUrBD4932W/E2BSpfRJWseIL4v/KPgBFxDOIdKpQ==",
"license": "BSD-3-Clause",
"dependencies": {
"@protobufjs/aspromise": "^1.1.1",
"@protobufjs/inquire": "^1.1.0"
}
},
"node_modules/@protobufjs/float": {
"version": "1.0.2",
"resolved": "https://registry.npmjs.org/@protobufjs/float/-/float-1.0.2.tgz",
"integrity": "sha512-Ddb+kVXlXst9d+R9PfTIxh1EdNkgoRe5tOX6t01f1lYWOvJnSPDBlG241QLzcyPdoNTsblLUdujGSE4RzrTZGQ==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/inquire": {
"version": "1.1.0",
"resolved": "https://registry.npmjs.org/@protobufjs/inquire/-/inquire-1.1.0.tgz",
"integrity": "sha512-kdSefcPdruJiFMVSbn801t4vFK7KB/5gd2fYvrxhuJYg8ILrmn9SKSX2tZdV6V+ksulWqS7aXjBcRXl3wHoD9Q==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/path": {
"version": "1.1.2",
"resolved": "https://registry.npmjs.org/@protobufjs/path/-/path-1.1.2.tgz",
"integrity": "sha512-6JOcJ5Tm08dOHAbdR3GrvP+yUUfkjG5ePsHYczMFLq3ZmMkAD98cDgcT2iA1lJ9NVwFd4tH/iSSoe44YWkltEA==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/pool": {
"version": "1.1.0",
"resolved": "https://registry.npmjs.org/@protobufjs/pool/-/pool-1.1.0.tgz",
"integrity": "sha512-0kELaGSIDBKvcgS4zkjz1PeddatrjYcmMWOlAuAPwAeccUrPHdUqo/J6LiymHHEiJT5NrF1UVwxY14f+fy4WQw==",
"license": "BSD-3-Clause"
},
"node_modules/@protobufjs/utf8": {
"version": "1.1.0",
"resolved": "https://registry.npmjs.org/@protobufjs/utf8/-/utf8-1.1.0.tgz",
"integrity": "sha512-Vvn3zZrhQZkkBE8LSuW3em98c0FwgO4nxzv6OdSxPKJIEKY2bGbHn+mhGIPerzI4twdxaP8/0+06HBpwf345Lw==",
"license": "BSD-3-Clause"
},
"node_modules/@rollup/rollup-android-arm-eabi": {
"version": "4.60.1",
"resolved": "https://registry.npmjs.org/@rollup/rollup-android-arm-eabi/-/rollup-android-arm-eabi-4.60.1.tgz",
@ -1117,6 +1205,12 @@
"undici-types": "~6.21.0"
}
},
"node_modules/@types/retry": {
"version": "0.12.0",
"resolved": "https://registry.npmjs.org/@types/retry/-/retry-0.12.0.tgz",
"integrity": "sha512-wWKOClTTiizcZhXnPY4wikVAwmdYHp8q6DmC+EJUzAMsycb7HB32Kh9RN4+0gExjmPmZSAQjgURXIGATPegAvA==",
"license": "MIT"
},
"node_modules/@types/ws": {
"version": "8.18.1",
"resolved": "https://registry.npmjs.org/@types/ws/-/ws-8.18.1.tgz",
@ -1275,6 +1369,15 @@
"url": "https://opencollective.com/vitest"
}
},
"node_modules/agent-base": {
"version": "7.1.4",
"resolved": "https://registry.npmjs.org/agent-base/-/agent-base-7.1.4.tgz",
"integrity": "sha512-MnA+YT8fwfJPgBx3m60MNqakm30XOkyIoH1y6huTQvC0PwZG7ki8NacLBcrPbNoo8vEZy7Jpuk7+jMO+CUovTQ==",
"license": "MIT",
"engines": {
"node": ">= 14"
}
},
"node_modules/ansi-regex": {
"version": "6.2.2",
"resolved": "https://registry.npmjs.org/ansi-regex/-/ansi-regex-6.2.2.tgz",
@ -1333,6 +1436,35 @@
"node": "18 || 20 || >=22"
}
},
"node_modules/base64-js": {
"version": "1.5.1",
"resolved": "https://registry.npmjs.org/base64-js/-/base64-js-1.5.1.tgz",
"integrity": "sha512-AKpaYlHn8t4SVbOHCy+b5+KKgvR4vrsD8vbvrbiQJps7fKDTkjkDry6ji0rUJjC0kzbNePLwzxq8iypo41qeWA==",
"funding": [
{
"type": "github",
"url": "https://github.com/sponsors/feross"
},
{
"type": "patreon",
"url": "https://www.patreon.com/feross"
},
{
"type": "consulting",
"url": "https://feross.org/support"
}
],
"license": "MIT"
},
"node_modules/bignumber.js": {
"version": "9.3.1",
"resolved": "https://registry.npmjs.org/bignumber.js/-/bignumber.js-9.3.1.tgz",
"integrity": "sha512-Ko0uX15oIUS7wJ3Rb30Fs6SkVbLmPBAKdlm7q9+ak9bbIeFf0MwuBsQV6z7+X768/cHsfg+WlysDWJcmthjsjQ==",
"license": "MIT",
"engines": {
"node": "*"
}
},
"node_modules/brace-expansion": {
"version": "5.0.5",
"resolved": "https://registry.npmjs.org/brace-expansion/-/brace-expansion-5.0.5.tgz",
@ -1346,6 +1478,12 @@
"node": "18 || 20 || >=22"
}
},
"node_modules/buffer-equal-constant-time": {
"version": "1.0.1",
"resolved": "https://registry.npmjs.org/buffer-equal-constant-time/-/buffer-equal-constant-time-1.0.1.tgz",
"integrity": "sha512-zRpUiDwd/xk6ADqPMATG8vc9VPrkck7T07OIx0gnjmJAnHnTVXNQG3vfvWNuiZIkwu9KrKdA1iJKfsfTVxE6NA==",
"license": "BSD-3-Clause"
},
"node_modules/cac": {
"version": "6.7.14",
"resolved": "https://registry.npmjs.org/cac/-/cac-6.7.14.tgz",
@ -1447,11 +1585,19 @@
"node": ">= 8"
}
},
"node_modules/data-uri-to-buffer": {
"version": "4.0.1",
"resolved": "https://registry.npmjs.org/data-uri-to-buffer/-/data-uri-to-buffer-4.0.1.tgz",
"integrity": "sha512-0R9ikRb668HB7QDxT1vkpuUBtqc53YyAwMwGeUFKRojY/NWKvdZ+9UYtRfGmhqNbRkTSVpMbmyhXipFFv2cb/A==",
"license": "MIT",
"engines": {
"node": ">= 12"
}
},
"node_modules/debug": {
"version": "4.4.3",
"resolved": "https://registry.npmjs.org/debug/-/debug-4.4.3.tgz",
"integrity": "sha512-RGwwWnwQvkVfavKVt22FGLw+xYSdzARwm0ru6DhTVA3umU5hZc28V3kO4stgYryrTlLpuvgI9GiijltAjNbcqA==",
"dev": true,
"license": "MIT",
"dependencies": {
"ms": "^2.1.3"
@ -1508,6 +1654,15 @@
"dev": true,
"license": "MIT"
},
"node_modules/ecdsa-sig-formatter": {
"version": "1.0.11",
"resolved": "https://registry.npmjs.org/ecdsa-sig-formatter/-/ecdsa-sig-formatter-1.0.11.tgz",
"integrity": "sha512-nagl3RYrbNv6kQkeJIpt6NJZy8twLB/2vtz6yN9Z4vRKHN4/QZJIEbqohALSgwKdnksuY3k5Addp5lg8sVoVcQ==",
"license": "Apache-2.0",
"dependencies": {
"safe-buffer": "^5.0.1"
}
},
"node_modules/emoji-regex": {
"version": "9.2.2",
"resolved": "https://registry.npmjs.org/emoji-regex/-/emoji-regex-9.2.2.tgz",
@ -1614,6 +1769,12 @@
"node": ">=12.0.0"
}
},
"node_modules/extend": {
"version": "3.0.2",
"resolved": "https://registry.npmjs.org/extend/-/extend-3.0.2.tgz",
"integrity": "sha512-fjquC59cD7CyW6urNXK0FBufkZcoiGG80wTuPujX590cB5Ttln20E2UB4S/WARVqhXffZl2LNgS+gQdPIIim/g==",
"license": "MIT"
},
"node_modules/fdir": {
"version": "6.5.0",
"resolved": "https://registry.npmjs.org/fdir/-/fdir-6.5.0.tgz",
@ -1632,6 +1793,29 @@
}
}
},
"node_modules/fetch-blob": {
"version": "3.2.0",
"resolved": "https://registry.npmjs.org/fetch-blob/-/fetch-blob-3.2.0.tgz",
"integrity": "sha512-7yAQpD2UMJzLi1Dqv7qFYnPbaPx7ZfFK6PiIxQ4PfkGPyNyl2Ugx+a/umUonmKqjhM4DnfbMvdX6otXq83soQQ==",
"funding": [
{
"type": "github",
"url": "https://github.com/sponsors/jimmywarting"
},
{
"type": "paypal",
"url": "https://paypal.me/jimmywarting"
}
],
"license": "MIT",
"dependencies": {
"node-domexception": "^1.0.0",
"web-streams-polyfill": "^3.0.3"
},
"engines": {
"node": "^12.20 || >= 14.13"
}
},
"node_modules/foreground-child": {
"version": "3.3.1",
"resolved": "https://registry.npmjs.org/foreground-child/-/foreground-child-3.3.1.tgz",
@ -1649,6 +1833,18 @@
"url": "https://github.com/sponsors/isaacs"
}
},
"node_modules/formdata-polyfill": {
"version": "4.0.10",
"resolved": "https://registry.npmjs.org/formdata-polyfill/-/formdata-polyfill-4.0.10.tgz",
"integrity": "sha512-buewHzMvYL29jdeQTVILecSaZKnt/RJWjoZCF5OW60Z67/GmSLBkOFM7qh1PI3zFNtJbaZL5eQu1vLfazOwj4g==",
"license": "MIT",
"dependencies": {
"fetch-blob": "^3.1.2"
},
"engines": {
"node": ">=12.20.0"
}
},
"node_modules/fsevents": {
"version": "2.3.3",
"resolved": "https://registry.npmjs.org/fsevents/-/fsevents-2.3.3.tgz",
@ -1673,6 +1869,34 @@
"url": "https://github.com/sponsors/ljharb"
}
},
"node_modules/gaxios": {
"version": "7.1.4",
"resolved": "https://registry.npmjs.org/gaxios/-/gaxios-7.1.4.tgz",
"integrity": "sha512-bTIgTsM2bWn3XklZISBTQX7ZSddGW+IO3bMdGaemHZ3tbqExMENHLx6kKZ/KlejgrMtj8q7wBItt51yegqalrA==",
"license": "Apache-2.0",
"dependencies": {
"extend": "^3.0.2",
"https-proxy-agent": "^7.0.1",
"node-fetch": "^3.3.2"
},
"engines": {
"node": ">=18"
}
},
"node_modules/gcp-metadata": {
"version": "8.1.2",
"resolved": "https://registry.npmjs.org/gcp-metadata/-/gcp-metadata-8.1.2.tgz",
"integrity": "sha512-zV/5HKTfCeKWnxG0Dmrw51hEWFGfcF2xiXqcA3+J90WDuP0SvoiSO5ORvcBsifmx/FoIjgQN3oNOGaQ5PhLFkg==",
"license": "Apache-2.0",
"dependencies": {
"gaxios": "^7.0.0",
"google-logging-utils": "^1.0.0",
"json-bigint": "^1.0.0"
},
"engines": {
"node": ">=18"
}
},
"node_modules/get-intrinsic": {
"version": "1.3.0",
"resolved": "https://registry.npmjs.org/get-intrinsic/-/get-intrinsic-1.3.0.tgz",
@ -1778,6 +2002,32 @@
"url": "https://github.com/sponsors/isaacs"
}
},
"node_modules/google-auth-library": {
"version": "10.6.2",
"resolved": "https://registry.npmjs.org/google-auth-library/-/google-auth-library-10.6.2.tgz",
"integrity": "sha512-e27Z6EThmVNNvtYASwQxose/G57rkRuaRbQyxM2bvYLLX/GqWZ5chWq2EBoUchJbCc57eC9ArzO5wMsEmWftCw==",
"license": "Apache-2.0",
"dependencies": {
"base64-js": "^1.3.0",
"ecdsa-sig-formatter": "^1.0.11",
"gaxios": "^7.1.4",
"gcp-metadata": "8.1.2",
"google-logging-utils": "1.1.3",
"jws": "^4.0.0"
},
"engines": {
"node": ">=18"
}
},
"node_modules/google-logging-utils": {
"version": "1.1.3",
"resolved": "https://registry.npmjs.org/google-logging-utils/-/google-logging-utils-1.1.3.tgz",
"integrity": "sha512-eAmLkjDjAFCVXg7A1unxHsLf961m6y17QFqXqAXGj/gVkKFrEICfStRfwUlGNfeCEjNRa32JEWOUTlYXPyyKvA==",
"license": "Apache-2.0",
"engines": {
"node": ">=14"
}
},
"node_modules/gopd": {
"version": "1.2.0",
"resolved": "https://registry.npmjs.org/gopd/-/gopd-1.2.0.tgz",
@ -1840,6 +2090,19 @@
"dev": true,
"license": "MIT"
},
"node_modules/https-proxy-agent": {
"version": "7.0.6",
"resolved": "https://registry.npmjs.org/https-proxy-agent/-/https-proxy-agent-7.0.6.tgz",
"integrity": "sha512-vK9P5/iUfdl95AI+JVyUuIcVtd4ofvtrOr3HNtM2yxC9bnMbEdp3x01OhQNnjb8IJYi38VlTE3mBXwcfvywuSw==",
"license": "MIT",
"dependencies": {
"agent-base": "^7.1.2",
"debug": "4"
},
"engines": {
"node": ">= 14"
}
},
"node_modules/iceberg-js": {
"version": "0.8.1",
"resolved": "https://registry.npmjs.org/iceberg-js/-/iceberg-js-0.8.1.tgz",
@ -1943,6 +2206,42 @@
"dev": true,
"license": "MIT"
},
"node_modules/json-bigint": {
"version": "1.0.0",
"resolved": "https://registry.npmjs.org/json-bigint/-/json-bigint-1.0.0.tgz",
"integrity": "sha512-SiPv/8VpZuWbvLSMtTDU8hEfrZWg/mH/nV/b4o0CYbSxu1UIQPLdwKOCIyLQX+VIPO5vrLX3i8qtqFyhdPSUSQ==",
"license": "MIT",
"dependencies": {
"bignumber.js": "^9.0.0"
}
},
"node_modules/jwa": {
"version": "2.0.1",
"resolved": "https://registry.npmjs.org/jwa/-/jwa-2.0.1.tgz",
"integrity": "sha512-hRF04fqJIP8Abbkq5NKGN0Bbr3JxlQ+qhZufXVr0DvujKy93ZCbXZMHDL4EOtodSbCWxOqR8MS1tXA5hwqCXDg==",
"license": "MIT",
"dependencies": {
"buffer-equal-constant-time": "^1.0.1",
"ecdsa-sig-formatter": "1.0.11",
"safe-buffer": "^5.0.1"
}
},
"node_modules/jws": {
"version": "4.0.1",
"resolved": "https://registry.npmjs.org/jws/-/jws-4.0.1.tgz",
"integrity": "sha512-EKI/M/yqPncGUUh44xz0PxSidXFr/+r0pA70+gIYhjv+et7yxM+s29Y+VGDkovRofQem0fs7Uvf4+YmAdyRduA==",
"license": "MIT",
"dependencies": {
"jwa": "^2.0.1",
"safe-buffer": "^5.0.1"
}
},
"node_modules/long": {
"version": "5.3.2",
"resolved": "https://registry.npmjs.org/long/-/long-5.3.2.tgz",
"integrity": "sha512-mNAgZ1GmyNhD7AuqnTG3/VQ26o760+ZYBPKjPvugO8+nLbYfX6TVpJPseBvopbdY+qpZ/lKUnmEc1LeZYS3QAA==",
"license": "Apache-2.0"
},
"node_modules/loupe": {
"version": "3.2.1",
"resolved": "https://registry.npmjs.org/loupe/-/loupe-3.2.1.tgz",
@ -2034,7 +2333,6 @@
"version": "2.1.3",
"resolved": "https://registry.npmjs.org/ms/-/ms-2.1.3.tgz",
"integrity": "sha512-6FlzubTLZG3J2a/NVCAleEhjzq5oxgHyaCU9yYXvcLsvoVaHJq/s5xXI6/XXP6tz7R9xAOtHnSO/tXtF3WRTlA==",
"dev": true,
"license": "MIT"
},
"node_modules/nanoid": {
@ -2056,6 +2354,44 @@
"node": "^10 || ^12 || ^13.7 || ^14 || >=15.0.1"
}
},
"node_modules/node-domexception": {
"version": "1.0.0",
"resolved": "https://registry.npmjs.org/node-domexception/-/node-domexception-1.0.0.tgz",
"integrity": "sha512-/jKZoMpw0F8GRwl4/eLROPA3cfcXtLApP0QzLmUT/HuPCZWyB7IY9ZrMeKw2O/nFIqPQB3PVM9aYm0F312AXDQ==",
"deprecated": "Use your platform's native DOMException instead",
"funding": [
{
"type": "github",
"url": "https://github.com/sponsors/jimmywarting"
},
{
"type": "github",
"url": "https://paypal.me/jimmywarting"
}
],
"license": "MIT",
"engines": {
"node": ">=10.5.0"
}
},
"node_modules/node-fetch": {
"version": "3.3.2",
"resolved": "https://registry.npmjs.org/node-fetch/-/node-fetch-3.3.2.tgz",
"integrity": "sha512-dRB78srN/l6gqWulah9SrxeYnxeddIG30+GOqK/9OlLVyLg3HPnr6SqOWTWOXKRwC2eGYCkZ59NNuSgvSrpgOA==",
"license": "MIT",
"dependencies": {
"data-uri-to-buffer": "^4.0.0",
"fetch-blob": "^3.1.4",
"formdata-polyfill": "^4.0.10"
},
"engines": {
"node": "^12.20.0 || ^14.13.1 || >=16.0.0"
},
"funding": {
"type": "opencollective",
"url": "https://opencollective.com/node-fetch"
}
},
"node_modules/object-inspect": {
"version": "1.13.4",
"resolved": "https://registry.npmjs.org/object-inspect/-/object-inspect-1.13.4.tgz",
@ -2068,6 +2404,19 @@
"url": "https://github.com/sponsors/ljharb"
}
},
"node_modules/p-retry": {
"version": "4.6.2",
"resolved": "https://registry.npmjs.org/p-retry/-/p-retry-4.6.2.tgz",
"integrity": "sha512-312Id396EbJdvRONlngUx0NydfrIQ5lsYu0znKVUzVvArzEIt08V1qhtyESbGVd1FGX7UKtiFp5uwKZdM8wIuQ==",
"license": "MIT",
"dependencies": {
"@types/retry": "0.12.0",
"retry": "^0.13.1"
},
"engines": {
"node": ">=8"
}
},
"node_modules/package-json-from-dist": {
"version": "1.0.1",
"resolved": "https://registry.npmjs.org/package-json-from-dist/-/package-json-from-dist-1.0.1.tgz",
@ -2168,6 +2517,30 @@
"node": "^10 || ^12 || >=14"
}
},
"node_modules/protobufjs": {
"version": "7.5.5",
"resolved": "https://registry.npmjs.org/protobufjs/-/protobufjs-7.5.5.tgz",
"integrity": "sha512-3wY1AxV+VBNW8Yypfd1yQY9pXnqTAN+KwQxL8iYm3/BjKYMNg4i0owhEe26PWDOMaIrzeeF98Lqd5NGz4omiIg==",
"hasInstallScript": true,
"license": "BSD-3-Clause",
"dependencies": {
"@protobufjs/aspromise": "^1.1.2",
"@protobufjs/base64": "^1.1.2",
"@protobufjs/codegen": "^2.0.4",
"@protobufjs/eventemitter": "^1.1.0",
"@protobufjs/fetch": "^1.1.0",
"@protobufjs/float": "^1.0.2",
"@protobufjs/inquire": "^1.1.0",
"@protobufjs/path": "^1.1.2",
"@protobufjs/pool": "^1.1.0",
"@protobufjs/utf8": "^1.1.0",
"@types/node": ">=13.7.0",
"long": "^5.0.0"
},
"engines": {
"node": ">=12.0.0"
}
},
"node_modules/qs": {
"version": "6.15.1",
"resolved": "https://registry.npmjs.org/qs/-/qs-6.15.1.tgz",
@ -2193,6 +2566,15 @@
"url": "https://github.com/privatenumber/resolve-pkg-maps?sponsor=1"
}
},
"node_modules/retry": {
"version": "0.13.1",
"resolved": "https://registry.npmjs.org/retry/-/retry-0.13.1.tgz",
"integrity": "sha512-XQBQ3I8W1Cge0Seh+6gjj03LbmRFWuoszgK9ooCpwYIrhhoO80pfq4cUkU5DkknwfOfFteRwlZ56PYOGYyFWdg==",
"license": "MIT",
"engines": {
"node": ">= 4"
}
},
"node_modules/rollup": {
"version": "4.60.1",
"resolved": "https://registry.npmjs.org/rollup/-/rollup-4.60.1.tgz",
@ -2238,6 +2620,26 @@
"fsevents": "~2.3.2"
}
},
"node_modules/safe-buffer": {
"version": "5.2.1",
"resolved": "https://registry.npmjs.org/safe-buffer/-/safe-buffer-5.2.1.tgz",
"integrity": "sha512-rp3So07KcdmmKbGvgaNxQSJr7bGVSVk5S9Eq1F+ppbRo70+YeaDxkw5Dd8NPN+GD6bjnYm2VuPuCXmpuYvmCXQ==",
"funding": [
{
"type": "github",
"url": "https://github.com/sponsors/feross"
},
{
"type": "patreon",
"url": "https://www.patreon.com/feross"
},
{
"type": "consulting",
"url": "https://feross.org/support"
}
],
"license": "MIT"
},
"node_modules/semver": {
"version": "7.7.4",
"resolved": "https://registry.npmjs.org/semver/-/semver-7.7.4.tgz",
@ -2833,6 +3235,15 @@
}
}
},
"node_modules/web-streams-polyfill": {
"version": "3.3.3",
"resolved": "https://registry.npmjs.org/web-streams-polyfill/-/web-streams-polyfill-3.3.3.tgz",
"integrity": "sha512-d2JWLCivmZYTSIoge9MsgFCZrt571BikcWGYkjC1khllbTeDlGqZ2D8vD8E/lJa8WGWbb7Plm8/XJYV7IJHZZw==",
"license": "MIT",
"engines": {
"node": ">= 8"
}
},
"node_modules/which": {
"version": "2.0.2",
"resolved": "https://registry.npmjs.org/which/-/which-2.0.2.tgz",

View file

@ -11,6 +11,7 @@
"test:coverage": "vitest run --coverage"
},
"dependencies": {
"@google/genai": "^1.50.1",
"@hono/node-server": "^1.13.7",
"@hono/node-ws": "^1.3.0",
"@supabase/supabase-js": "^2.49.4",

View file

@ -1,5 +1,50 @@
import { describe, it, expect, beforeEach, afterEach, vi } from "vitest";
import { EventEmitter } from "node:events";
// ─── Mock du SDK @google/genai ───────────────────────────────────────────────
//
// On capture les callbacks passés à `ai.live.connect` pour pouvoir simuler les
// événements (onopen, onmessage, onerror, onclose) depuis les tests. La
// fabrique `clientFactory` injectée dans openGeminiLiveSession permet de
// remplacer `new GoogleGenAI(...)` par un stub.
interface CapturedConnect {
model: string;
config: Record<string, unknown>;
callbacks: {
onopen?: () => void;
onmessage?: (msg: unknown) => void;
onerror?: (err: unknown) => void;
onclose?: (evt: unknown) => void;
};
session: {
sendRealtimeInput: ReturnType<typeof vi.fn>;
close: ReturnType<typeof vi.fn>;
};
}
let capturedConnect: CapturedConnect | null = null;
function makeFakeClient() {
return {
live: {
connect: vi.fn(async (params: CapturedConnect) => {
const session = {
sendRealtimeInput: vi.fn(),
close: vi.fn(),
};
capturedConnect = {
model: params.model,
config: params.config,
callbacks: params.callbacks,
session,
};
return session;
}),
},
};
}
import {
openGeminiLiveSession,
buildT2SystemPrompt,
@ -30,6 +75,33 @@ const SUJET_OPTS = {
"Vous cherchez un appartement de 2 pièces dans le centre-ville, votre budget est limité et vous souhaitez emménager le mois prochain.",
};
/** Helper : ouvre une session avec un client mocké et retourne la capture. */
async function openWithMock(
client: FakeWs,
extra: Partial<{
onSessionEnd: (transcript: string) => void | Promise<void>;
timeoutMs: number;
warningMs: number;
}> = {},
) {
capturedConnect = null;
// eslint-disable-next-line @typescript-eslint/no-explicit-any
openGeminiLiveSession(client, {
...SUJET_OPTS,
apiKey: "test-key",
clientFactory: () => makeFakeClient() as any,
...extra,
});
// Le `await live.connect()` est dans un `.then()` du code prod ; on laisse
// les microtasks se vider avant de retourner la capture.
await Promise.resolve();
await Promise.resolve();
if (!capturedConnect) {
throw new Error("Le mock du SDK n'a pas capturé de connect()");
}
return capturedConnect;
}
describe("buildT2SystemPrompt", () => {
it("substitue role et contexte dans le template", () => {
const prompt = buildT2SystemPrompt(SUJET_OPTS);
@ -42,170 +114,120 @@ describe("buildT2SystemPrompt", () => {
});
});
describe("openGeminiLiveSession", () => {
let originalKey: string | undefined;
describe("openGeminiLiveSession (SDK)", () => {
beforeEach(() => {
originalKey = process.env.GEMINI_API_KEY;
process.env.GEMINI_API_KEY = "test-key";
vi.useFakeTimers();
});
afterEach(() => {
if (originalKey === undefined) {
delete process.env.GEMINI_API_KEY;
} else {
process.env.GEMINI_API_KEY = originalKey;
}
vi.useRealTimers();
vi.restoreAllMocks();
capturedConnect = null;
});
it("envoie le setup frame avec prompt dynamique + VAD + transcriptions", () => {
it("appelle live.connect avec le modèle + config Live (audio + system + transcripts + VAD)", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const capture = await openWithMock(client);
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
});
gemini.emit("open");
expect(gemini.sent).toHaveLength(1);
const setup = JSON.parse(gemini.sent[0] as string);
expect(setup.setup.model).toMatch(/gemini/);
expect(setup.setup.systemInstruction.parts[0].text).toContain(
expect(capture.model).toMatch(/gemini/);
const config = capture.config;
expect(config.responseModalities).toContain("AUDIO");
expect(config.systemInstruction).toContain(
"un bailleur qui propose un appartement",
);
expect(setup.setup.generationConfig.responseModalities).toContain("AUDIO");
expect(setup.setup.inputAudioTranscription).toEqual({});
expect(setup.setup.outputAudioTranscription).toEqual({});
expect(
setup.setup.realtimeInputConfig.automaticActivityDetection,
).toMatchObject({
disabled: false,
startOfSpeechSensitivity: "START_SENSITIVITY_LOW",
endOfSpeechSensitivity: "END_SENSITIVITY_LOW",
silenceDurationMs: 2000,
expect(config.inputAudioTranscription).toEqual({});
expect(config.outputAudioTranscription).toEqual({});
// eslint-disable-next-line @typescript-eslint/no-explicit-any
const vad: any = (config.realtimeInputConfig as any)
?.automaticActivityDetection;
expect(vad?.disabled).toBe(false);
expect(vad?.silenceDurationMs).toBe(2000);
});
it("forwarde un chunk audio client {type:'audio'} via session.sendRealtimeInput (PCM 16k base64)", async () => {
const client = new FakeWs();
const capture = await openWithMock(client);
capture.callbacks.onopen?.();
const base64 = "AQIDBA=="; // base64 de [1,2,3,4]
client.emit("message", JSON.stringify({ type: "audio", data: base64 }));
expect(capture.session.sendRealtimeInput).toHaveBeenCalledTimes(1);
expect(capture.session.sendRealtimeInput).toHaveBeenCalledWith({
audio: { data: base64, mimeType: "audio/pcm;rate=16000" },
});
});
it("forwarde un chunk audio client (Buffer) vers Gemini", () => {
it("forwarde un message Gemini (audio inlineData) au client en JSON", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
});
gemini.emit("open");
const capture = await openWithMock(client);
capture.callbacks.onopen?.();
const audioChunk = Buffer.from([0x01, 0x02, 0x03, 0x04]);
client.emit("message", audioChunk);
const geminiMsg = {
serverContent: {
modelTurn: {
parts: [
{
inlineData: { data: "EAYE", mimeType: "audio/pcm;rate=24000" },
},
],
},
},
};
capture.callbacks.onmessage?.(geminiMsg);
// [0] = setup, [1] = chunk audio
expect(gemini.sent).toHaveLength(2);
expect(gemini.sent[1]).toBe(audioChunk);
});
it("forwarde un chunk audio Gemini (Buffer non-JSON) vers le client sans accumuler de transcript", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const onSessionEnd = vi.fn();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
onSessionEnd,
});
gemini.emit("open");
const examinerAudio = Buffer.from([0x10, 0x20, 0x30]);
gemini.emit("message", examinerAudio);
expect(client.sent).toHaveLength(1);
expect(client.sent[0]).toBe(examinerAudio);
// Fin de session via signal client → transcript vide
client.emit("message", JSON.stringify({ type: "end" }));
await vi.runAllTimersAsync();
expect(onSessionEnd).toHaveBeenCalledWith("");
expect(JSON.parse(client.sent[0] as string)).toEqual(geminiMsg);
});
it("accumule inputTranscription et outputTranscription depuis Gemini", async () => {
it("accumule input/outputTranscription et reconstruit le transcript chronologique", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const onSessionEnd = vi.fn();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
onSessionEnd,
});
gemini.emit("open");
const capture = await openWithMock(client, { onSessionEnd });
capture.callbacks.onopen?.();
gemini.emit(
"message",
JSON.stringify({
serverContent: {
inputTranscription: { text: "Bonjour, je voudrais louer." },
},
}),
);
gemini.emit(
"message",
JSON.stringify({
serverContent: {
outputTranscription: { text: "Bonjour, cest pour quel quartier ?" },
},
}),
);
gemini.emit(
"message",
JSON.stringify({
serverContent: {
inputTranscription: { text: "Le centre-ville." },
},
}),
);
capture.callbacks.onmessage?.({
serverContent: {
inputTranscription: { text: "Bonjour, je voudrais louer." },
},
});
capture.callbacks.onmessage?.({
serverContent: {
outputTranscription: { text: "Bonjour, cest pour quel quartier ?" },
},
});
capture.callbacks.onmessage?.({
serverContent: { inputTranscription: { text: "Le centre-ville." } },
});
client.emit("message", JSON.stringify({ type: "end" }));
await vi.runAllTimersAsync();
expect(onSessionEnd).toHaveBeenCalledTimes(1);
const transcript = onSessionEnd.mock.calls[0][0] as string;
expect(transcript).toBe(
expect(onSessionEnd.mock.calls[0][0]).toBe(
"Candidat : Bonjour, je voudrais louer.\nExaminateur : Bonjour, cest pour quel quartier ?\nCandidat : Le centre-ville.",
);
});
it("ferme Gemini après onSessionEnd, sans fermer le client (réservé à lappelant)", async () => {
it("ferme la session SDK après onSessionEnd, sans fermer le client", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const onSessionEnd = vi.fn();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
onSessionEnd,
});
gemini.emit("open");
const capture = await openWithMock(client, { onSessionEnd });
capture.callbacks.onopen?.();
client.emit("message", JSON.stringify({ type: "end" }));
await vi.runAllTimersAsync();
expect(gemini.closed).toBe(true);
expect(gemini.closeCode).toBe(1000);
expect(capture.session.close).toHaveBeenCalledTimes(1);
expect(client.closed).toBe(false);
});
it("warning à 180 s puis timeout à 210 s déclenche endSession", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const onSessionEnd = vi.fn();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
onSessionEnd,
});
gemini.emit("open");
const capture = await openWithMock(client, { onSessionEnd });
capture.callbacks.onopen?.();
// Avancer à 180 s → warning au client
await vi.advanceTimersByTimeAsync(180_000);
const warningFrame = client.sent.find(
(f) => typeof f === "string" && f.includes('"warning"'),
@ -217,22 +239,16 @@ describe("openGeminiLiveSession", () => {
});
expect(onSessionEnd).not.toHaveBeenCalled();
// Avancer à 210 s total → timeout déclenche endSession
await vi.advanceTimersByTimeAsync(30_000);
expect(onSessionEnd).toHaveBeenCalledTimes(1);
expect(gemini.closed).toBe(true);
expect(capture.session.close).toHaveBeenCalled();
});
it("signal end client déclenche endSession une seule fois (idempotent)", async () => {
it("signal end client est idempotent (un seul onSessionEnd)", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
const onSessionEnd = vi.fn();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
onSessionEnd,
});
gemini.emit("open");
const capture = await openWithMock(client, { onSessionEnd });
capture.callbacks.onopen?.();
client.emit("message", JSON.stringify({ type: "end" }));
client.emit("message", JSON.stringify({ type: "end" }));
@ -241,47 +257,47 @@ describe("openGeminiLiveSession", () => {
expect(onSessionEnd).toHaveBeenCalledTimes(1);
});
it("fermeture Gemini avant fin → close client 4006 GEMINI_DISCONNECTED", () => {
it("onclose SDK avant fin → close client 4006 GEMINI_DISCONNECTED", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
});
gemini.emit("open");
const capture = await openWithMock(client);
capture.callbacks.onopen?.();
gemini.emit("close");
capture.callbacks.onclose?.({ code: 1000 });
expect(client.closed).toBe(true);
expect(client.closeCode).toBe(4006);
expect(client.closeReason).toBe("GEMINI_DISCONNECTED");
});
it("erreur Gemini → close client 4006 GEMINI_DISCONNECTED", () => {
it("onerror SDK → close client 4006", async () => {
const client = new FakeWs();
const gemini = new FakeWs();
openGeminiLiveSession(client, {
...SUJET_OPTS,
geminiFactory: () => gemini,
});
gemini.emit("open");
const capture = await openWithMock(client);
capture.callbacks.onopen?.();
gemini.emit("error", new Error("boom"));
capture.callbacks.onerror?.(new Error("boom"));
expect(client.closed).toBe(true);
expect(client.closeCode).toBe(4006);
});
it("absence de GEMINI_API_KEY → close client 4005 GEMINI_CONFIG sans appel à la factory", () => {
it("absence de GEMINI_API_KEY → close client 4005 GEMINI_CONFIG sans appel à live.connect", () => {
const originalKey = process.env.GEMINI_API_KEY;
delete process.env.GEMINI_API_KEY;
capturedConnect = null;
const client = new FakeWs();
const factory = vi.fn(() => new FakeWs());
const factory = vi.fn(() => makeFakeClient());
openGeminiLiveSession(client, { ...SUJET_OPTS, geminiFactory: factory });
openGeminiLiveSession(client, {
...SUJET_OPTS,
// eslint-disable-next-line @typescript-eslint/no-explicit-any
clientFactory: factory as any,
});
expect(factory).not.toHaveBeenCalled();
expect(client.closed).toBe(true);
expect(client.closeCode).toBe(4005);
expect(client.closeReason).toBe("GEMINI_CONFIG");
if (originalKey !== undefined) process.env.GEMINI_API_KEY = originalKey;
});
});

View file

@ -1,9 +1,38 @@
import { WebSocket as NodeWebSocket } from "ws";
/**
* geminiLive.ts Sprint 6d.
*
* Migration du WebSocket brut (`wss://generativelanguage.googleapis.com/...`)
* vers le SDK officiel `@google/genai` v1.50.x. Motif : Google a migré les
* clés API vers le mode "Vertex AI Express", incompatible avec l'endpoint WS
* historique (réponse 403 systématique). Le SDK gère l'auth automatiquement
* et accepte les clés Express bound à un service account.
*
* Interface publique (consommée par `routes/t2live.ts`) :
* - openGeminiLiveSession(clientWs, opts) : ouvre une session Live et
* proxifie les messages dans les deux sens entre le client (navigateur)
* et Gemini, accumule les transcripts, gère timeouts + close codes.
* - WebSocketLike : interface minimale pour le client WS (Hono adapter).
* - buildT2SystemPrompt({role, contexte}) : prompt dynamique T2 Live.
* - GEMINI_LIVE_MODEL, T2_SESSION_TIMEOUT_MS, T2_SESSION_WARNING_MS.
*
* Cf. docs/IMPLEMENTATION_T2_LIVE.md §3, docs/Prompt_t2live.md §3.
*/
export const GEMINI_LIVE_URL =
"wss://generativelanguage.googleapis.com/ws/google.ai.generativelanguage.v1beta.GenerativeService.BidiGenerateContent";
import {
GoogleGenAI,
Modality,
StartSensitivity,
EndSensitivity,
type Session,
} from "@google/genai";
export const GEMINI_LIVE_MODEL = "models/gemini-2.5-flash-native-audio-latest";
/**
* Modèle Live cible. `gemini-3.1-flash-live-preview` est le choix par défaut
* (Sprint 6d), à valider sur Express Mode via `test-gemini-live.js`. Fallback
* documenté : `gemini-2.0-flash-live-001` (modèle Live garanti sur Express
* d'après la doc Vertex Express).
*/
export const GEMINI_LIVE_MODEL = "gemini-3.1-flash-live-preview";
/** Timeout total session WS T2 Live : 3 min 30 (durée TCF) + marge évaluation. */
export const T2_SESSION_TIMEOUT_MS = 210_000;
@ -36,7 +65,6 @@ Règles à respecter impérativement :
/**
* Subset minimal d'une WebSocket compatible avec :
* - le wrapper exposé par @hono/node-ws (côté client navigateur)
* - la WebSocket de `ws` (côté Gemini)
* - les fakes basés sur EventEmitter dans les tests
*/
export interface WebSocketLike {
@ -59,34 +87,32 @@ export interface OpenGeminiLiveSessionOptions {
timeoutMs?: number;
/** Override warning (par défaut T2_SESSION_WARNING_MS). */
warningMs?: number;
/** Injection pour les tests — fabrique de WebSocket vers Gemini. */
geminiFactory?: (url: string) => WebSocketLike;
/** Surcharge la clé API (par défaut : process.env.GEMINI_API_KEY). */
apiKey?: string;
/**
* Injection pour les tests fabrique de client SDK. Permet de remplacer
* `new GoogleGenAI(...)` par un mock dans les tests sans toucher au code prod.
*/
clientFactory?: (apiKey: string) => GoogleGenAI;
}
function buildSetupFrame(systemPrompt: string): string {
return JSON.stringify({
setup: {
model: GEMINI_LIVE_MODEL,
systemInstruction: {
parts: [{ text: systemPrompt }],
},
generationConfig: {
responseModalities: ["AUDIO"],
},
inputAudioTranscription: {},
outputAudioTranscription: {},
realtimeInputConfig: {
automaticActivityDetection: {
disabled: false,
startOfSpeechSensitivity: "START_SENSITIVITY_LOW",
endOfSpeechSensitivity: "END_SENSITIVITY_LOW",
silenceDurationMs: 2000,
},
},
},
});
/**
* Forme minimale d'un message Live retourné par le SDK. On n'exporte pas
* `LiveServerMessage` du SDK pour ne pas coupler les tests à son shape exact.
*/
interface LiveServerMessage {
serverContent?: {
modelTurn?: {
parts?: Array<{
inlineData?: { data?: string; mimeType?: string };
}>;
};
inputTranscription?: { text?: string };
outputTranscription?: { text?: string };
interrupted?: boolean;
turnComplete?: boolean;
};
setupComplete?: unknown;
}
interface TranscriptEntry {
@ -104,54 +130,6 @@ function reconstructTranscript(entries: TranscriptEntry[]): string {
.join("\n");
}
/**
* Tente de parser un message Gemini en JSON pour en extraire les transcripts.
* Retourne null si non-JSON (chunks audio binaires).
*/
function tryParseGeminiMessage(data: unknown): {
inputText?: string;
outputText?: string;
} | null {
let text: string;
if (typeof data === "string") {
text = data;
} else if (data instanceof Buffer) {
// Heuristique : tenter de parser comme JSON UTF-8 ; si ça échoue, c'est binaire.
try {
text = data.toString("utf8");
if (!text.startsWith("{")) return null;
} catch {
return null;
}
} else if (typeof data === "object" && data !== null && "toString" in data) {
try {
text = (data as { toString: () => string }).toString();
if (!text.startsWith("{")) return null;
} catch {
return null;
}
} else {
return null;
}
try {
const parsed = JSON.parse(text) as {
serverContent?: {
inputTranscription?: { text?: string };
outputTranscription?: { text?: string };
};
};
const sc = parsed.serverContent;
if (!sc) return {};
return {
inputText: sc.inputTranscription?.text,
outputText: sc.outputTranscription?.text,
};
} catch {
return null;
}
}
/**
* Détecte un signal de fin de session envoyé par le client : `{type:'end'}`.
*/
@ -178,19 +156,53 @@ function isEndSignal(data: unknown): boolean {
}
/**
* Ouvre une session Gemini Live et proxifie les messages
* Parse un message client `{type:'audio', data: base64}` et renvoie le base64
* si le format est valide, sinon null.
*/
function parseAudioChunk(data: unknown): string | null {
let text: string;
if (typeof data === "string") {
text = data;
} else if (data instanceof Buffer) {
try {
text = data.toString("utf8");
} catch {
return null;
}
} else {
return null;
}
if (!text.startsWith("{")) return null;
try {
const parsed = JSON.parse(text) as { type?: string; data?: unknown };
if (parsed.type === "audio" && typeof parsed.data === "string") {
return parsed.data;
}
return null;
} catch {
return null;
}
}
/**
* Ouvre une session Gemini Live via le SDK et proxifie les messages
* dans les deux sens entre le client (navigateur) et Gemini.
*
* - À l'open Gemini : envoie le setup frame avec prompt dynamique + VAD
* + inputAudioTranscription + outputAudioTranscription.
* - Forward transparent des frames audio dans les deux directions.
* - Accumule les transcripts (input = candidat, output = examinateur IA).
* - Détecte signal client `{type:'end'}` déclenche fin de session.
* - Timeout 210 s : warning client à 180 s, fin auto à 210 s.
* - En fin de session : appelle `onSessionEnd(transcript)` puis ferme Gemini.
* Le client WS n'est PAS fermé ici — c'est l'appelant qui décide (envoi du
* rapport puis close 1000).
* - Erreur Gemini close client 4006 GEMINI_DISCONNECTED.
* - Init : `new GoogleGenAI({ vertexai: true, apiKey })` mode Vertex Express
* (compatible avec les clés API auto-bound à un service account).
* - Setup config : modèle + responseModalities AUDIO + systemInstruction
* + inputAudioTranscription + outputAudioTranscription + VAD.
* - Forward client Gemini : parse `{type:'audio', data: base64}`
* `session.sendRealtimeInput({audio: {data, mimeType: 'audio/pcm;rate=16000'}})`.
* - Forward Gemini client : `clientWs.send(JSON.stringify(msg))` (le frontend
* parse `serverContent.modelTurn.parts[].inlineData.data`).
* - Accumule input/outputTranscription pour la correction finale.
* - Détecte `{type:'end'}` du client fin de session.
* - Timer 210 s : warning à 180 s, fin auto à 210 s.
* - En fin : `onSessionEnd(transcript)` puis ferme la session SDK. Le client WS
* n'est PAS fermé ici — c'est l'appelant qui décide (envoi du rapport puis
* close 1000).
* - Erreur SDK / close Gemini close client 4006 GEMINI_DISCONNECTED.
* - GEMINI_API_KEY absente close client 4005 GEMINI_CONFIG.
*/
export function openGeminiLiveSession(
@ -211,17 +223,14 @@ export function openGeminiLiveSession(
contexte: opts.contexte,
});
const url = `${GEMINI_LIVE_URL}?key=${apiKey}`;
const factory =
opts.geminiFactory ??
((u: string) => new NodeWebSocket(u) as unknown as WebSocketLike);
const geminiWs = factory(url);
const ai =
opts.clientFactory?.(apiKey) ?? new GoogleGenAI({ vertexai: true, apiKey });
const transcriptEntries: TranscriptEntry[] = [];
let sessionEnded = false;
let warningTimer: ReturnType<typeof setTimeout> | null = null;
let timeoutTimer: ReturnType<typeof setTimeout> | null = null;
let session: Session | null = null;
const clearTimers = () => {
if (warningTimer !== null) {
@ -238,10 +247,12 @@ export function openGeminiLiveSession(
if (sessionEnded) return;
sessionEnded = true;
clearTimers();
try {
geminiWs.close(1000);
} catch {
/* ignore */
if (session) {
try {
session.close();
} catch {
/* ignore */
}
}
if (opts.onSessionEnd) {
try {
@ -255,105 +266,153 @@ export function openGeminiLiveSession(
}
};
geminiWs.on("open", () => {
console.log("[T2] Gemini WS opened");
try {
geminiWs.send(buildSetupFrame(systemPrompt));
console.log("[T2] Setup frame sent");
// Démarrer les timers une fois la session Gemini effectivement ouverte.
warningTimer = setTimeout(() => {
if (sessionEnded) return;
try {
clientWs.send(
JSON.stringify({
type: "warning",
message: "30 secondes restantes",
}),
);
} catch {
/* ignore */
}
}, warningMs);
timeoutTimer = setTimeout(() => {
void endSession();
}, timeoutMs);
} catch {
try {
clientWs.close(4005, "GEMINI_CONFIG");
} catch {
/* ignore */
}
const handleSdkMessage = (msg: LiveServerMessage) => {
// Accumuler transcripts pour la correction finale.
const sc = msg.serverContent;
if (sc?.inputTranscription?.text && sc.inputTranscription.text.length > 0) {
transcriptEntries.push({
speaker: "candidat",
text: sc.inputTranscription.text,
});
}
});
geminiWs.on("message", (data) => {
// Tentative d'extraction des transcripts — si JSON, on accumule ;
// dans tous les cas (JSON ou audio binaire), on forward au client.
const parsed = tryParseGeminiMessage(data);
if (parsed) {
if (parsed.inputText && parsed.inputText.length > 0) {
transcriptEntries.push({
speaker: "candidat",
text: parsed.inputText,
});
}
if (parsed.outputText && parsed.outputText.length > 0) {
transcriptEntries.push({
speaker: "examinateur",
text: parsed.outputText,
});
}
if (
sc?.outputTranscription?.text &&
sc.outputTranscription.text.length > 0
) {
transcriptEntries.push({
speaker: "examinateur",
text: sc.outputTranscription.text,
});
}
// Forward verbatim au client. Le frontend parse serverContent.modelTurn.
try {
clientWs.send(data);
clientWs.send(JSON.stringify(msg));
} catch {
void endSession();
}
});
};
// ── Ouverture de la session SDK ──────────────────────────────────────
ai.live
.connect({
model: GEMINI_LIVE_MODEL,
config: {
responseModalities: [Modality.AUDIO],
systemInstruction: systemPrompt,
inputAudioTranscription: {},
outputAudioTranscription: {},
realtimeInputConfig: {
automaticActivityDetection: {
disabled: false,
startOfSpeechSensitivity: StartSensitivity.START_SENSITIVITY_LOW,
endOfSpeechSensitivity: EndSensitivity.END_SENSITIVITY_LOW,
silenceDurationMs: 2000,
},
},
},
callbacks: {
onopen: () => {
console.log("[T2] Session Gemini ouverte (SDK)");
// Démarrer les timers une fois la session effectivement ouverte.
warningTimer = setTimeout(() => {
if (sessionEnded) return;
try {
clientWs.send(
JSON.stringify({
type: "warning",
message: "30 secondes restantes",
}),
);
} catch {
/* ignore */
}
}, warningMs);
timeoutTimer = setTimeout(() => {
void endSession();
}, timeoutMs);
},
onmessage: (msg: LiveServerMessage) => {
handleSdkMessage(msg);
},
onerror: (err: unknown) => {
console.log(
"[T2] Erreur SDK :",
err instanceof Error ? err.message : String(err),
);
if (!sessionEnded) {
clearTimers();
sessionEnded = true;
try {
clientWs.close(4006, "GEMINI_DISCONNECTED");
} catch {
/* ignore */
}
}
},
onclose: () => {
console.log("[T2] Session Gemini fermée (SDK)");
if (!sessionEnded) {
clearTimers();
try {
clientWs.close(4006, "GEMINI_DISCONNECTED");
} catch {
/* ignore */
}
}
},
},
})
.then((s: Session) => {
session = s;
})
.catch((err: unknown) => {
console.log(
"[T2] live.connect a échoué :",
err instanceof Error ? err.message : String(err),
);
sessionEnded = true;
clearTimers();
try {
clientWs.close(4006, "GEMINI_DISCONNECTED");
} catch {
/* ignore */
}
});
// ── Forward client → Gemini ──────────────────────────────────────────
clientWs.on("message", (data) => {
if (isEndSignal(data)) {
void endSession();
return;
}
try {
geminiWs.send(data);
} catch {
void endSession();
}
});
geminiWs.on("close", () => {
console.log("[T2] Gemini closed");
if (!sessionEnded) {
clearTimers();
const audioBase64 = parseAudioChunk(data);
if (audioBase64 !== null && session !== null && !sessionEnded) {
try {
clientWs.close(4006, "GEMINI_DISCONNECTED");
} catch {
/* ignore */
session.sendRealtimeInput({
audio: {
data: audioBase64,
mimeType: "audio/pcm;rate=16000",
},
});
} catch (err) {
console.log(
"[T2] sendRealtimeInput a échoué :",
err instanceof Error ? err.message : String(err),
);
void endSession();
}
}
// Tout autre message client est ignoré (ex: ping keep-alive frontend).
});
clientWs.on("close", () => {
clearTimers();
sessionEnded = true;
try {
geminiWs.close(1000);
} catch {
/* ignore */
}
});
geminiWs.on("error", (err) => {
console.log("[T2] Gemini error:", (err as Error)?.message);
if (!sessionEnded) {
clearTimers();
sessionEnded = true;
if (session) {
try {
clientWs.close(4006, "GEMINI_DISCONNECTED");
session.close();
} catch {
/* ignore */
}
@ -363,10 +422,12 @@ export function openGeminiLiveSession(
clientWs.on("error", () => {
clearTimers();
sessionEnded = true;
try {
geminiWs.close(1011);
} catch {
/* ignore */
if (session) {
try {
session.close();
} catch {
/* ignore */
}
}
});
}

View file

@ -230,8 +230,8 @@ export async function runT2LiveCorrection(args: {
}
export interface CreateT2LiveRoutesOptions {
/** Injection pour les tests : fabrique de WebSocket vers Gemini. */
geminiFactory?: OpenGeminiLiveSessionOptions["geminiFactory"];
/** Injection pour les tests : fabrique de client SDK Gemini (Sprint 6d). */
clientFactory?: OpenGeminiLiveSessionOptions["clientFactory"];
/** Injection pour les tests : override timeout/warning. */
timeoutMs?: number;
warningMs?: number;
@ -313,7 +313,7 @@ export default function createT2LiveRoutes(
openGeminiLiveSession(adapter, {
role: sujetNonNull.role!,
contexte: sujetNonNull.contexte!,
geminiFactory: opts.geminiFactory,
clientFactory: opts.clientFactory,
timeoutMs: opts.timeoutMs,
warningMs: opts.warningMs,
onSessionEnd: async (transcript) => {

150
test-gemini-live.js Normal file
View file

@ -0,0 +1,150 @@
// test-gemini-live.js — Sprint 6d : debug du setup frame Gemini Live via SDK.
//
// Usage :
// node --env-file=.env test-gemini-live.js minimal
// node --env-file=.env test-gemini-live.js +system
// node --env-file=.env test-gemini-live.js +transcription
// node --env-file=.env test-gemini-live.js +vad
//
// Chaque mode part du `minimal` qui doit fonctionner avec une clé Express
// Mode et ajoute UN champ. Si le mode reçoit `setupComplete` → le champ est
// accepté. Si l'ouverture échoue → c'est ce champ qui pose problème.
//
// Migration Sprint 6d : passage du WebSocket brut au SDK officiel
// `@google/genai` qui gère l'auth Express Mode automatiquement.
import {
GoogleGenAI,
Modality,
StartSensitivity,
EndSensitivity,
} from "@google/genai";
const MODES = ["minimal", "+system", "+transcription", "+vad"];
const mode = process.argv[2] ?? "minimal";
if (!MODES.includes(mode)) {
console.error(
`❌ Mode inconnu : "${mode}". Modes valides : ${MODES.join(", ")}`,
);
process.exit(1);
}
const KEY = process.env.GEMINI_API_KEY;
if (!KEY) {
console.error("❌ GEMINI_API_KEY manquante dans l'env");
process.exit(1);
}
// Modèle par défaut Sprint 6d. Fallback documenté : `gemini-2.0-flash-live-001`.
const MODEL = "gemini-3.1-flash-live-preview";
const SAMPLE_PROMPT =
"Tu joues le rôle d'un bailleur. Tu réponds uniquement en français. " +
"Tu attends que ton interlocuteur s'adresse à toi avant de parler.";
function buildConfig(mode) {
// Base minimal — équivalent au mode `minimal` qui doit fonctionner.
const config = {
responseModalities: [Modality.AUDIO],
};
if (mode === "+system") {
config.systemInstruction = SAMPLE_PROMPT;
}
if (mode === "+transcription") {
config.inputAudioTranscription = {};
config.outputAudioTranscription = {};
}
if (mode === "+vad") {
config.realtimeInputConfig = {
automaticActivityDetection: {
disabled: false,
startOfSpeechSensitivity: StartSensitivity.START_SENSITIVITY_LOW,
endOfSpeechSensitivity: EndSensitivity.END_SENSITIVITY_LOW,
silenceDurationMs: 2000,
},
};
}
return config;
}
const ai = new GoogleGenAI({ vertexai: true, apiKey: KEY });
console.log(`→ Mode : ${mode}`);
console.log(`→ Modèle : ${MODEL}`);
console.log("→ Connexion à Gemini Live (via SDK)…");
let setupCompleteReceived = false;
let resolved = false;
const config = buildConfig(mode);
console.log("→ Config envoyée :");
console.log(JSON.stringify(config, null, 2));
const timeoutId = setTimeout(() => {
if (!resolved) {
console.log("⏱ Timeout 15 s — pas de setupComplete reçu.");
process.exit(setupCompleteReceived ? 0 : 1);
}
}, 15000);
try {
const session = await ai.live.connect({
model: MODEL,
config,
callbacks: {
onopen: () => {
console.log("✅ Connexion ouverte");
},
onmessage: (msg) => {
// Compat : selon la version du SDK, setupComplete arrive soit comme
// propriété directe, soit dans serverContent. On loggue tout.
console.log("📨 Message reçu :", JSON.stringify(msg).slice(0, 600));
if (msg.setupComplete || msg?.serverContent?.setupComplete) {
setupCompleteReceived = true;
resolved = true;
console.log(
`\n🎉 [${mode}] ACCEPTÉ — setupComplete reçu (modèle ${MODEL}).`,
);
clearTimeout(timeoutId);
try {
session.close();
} catch {
/* ignore */
}
process.exit(0);
}
},
onerror: (err) => {
console.log("❌ Erreur :", err?.message ?? err);
},
onclose: (evt) => {
console.log(
`🔒 Fermeture${evt?.code ? ` — code ${evt.code}` : ""}${evt?.reason ? ` reason: ${evt.reason}` : ""}`,
);
if (!setupCompleteReceived) {
console.log(`\n⚠ [${mode}] REJETÉ — fermeture avant setupComplete.`);
console.log(
"→ Le ou les champs ajoutés par ce mode ne sont pas acceptés.",
);
}
resolved = true;
clearTimeout(timeoutId);
process.exit(setupCompleteReceived ? 0 : 1);
},
},
});
// Conserver la session vivante jusqu'au timeout/setupComplete.
void session;
} catch (err) {
resolved = true;
clearTimeout(timeoutId);
console.log(
"❌ live.connect a échoué :",
err instanceof Error ? err.message : String(err),
);
process.exit(1);
}