lingxinai-psychology 1.0.5

Python SDK for LingxinAI Psychology Open APIs

LingxinAI Psychology Python SDK

Python SDK for LingxinAI Psychology Open APIs.

Install

pip install lingxinai-psychology

Quick Start

from lingxinai.psychology import PsychologyClient
from lingxinai.psychology.types import OpenStudentQueryRequest

client = PsychologyClient.builder() \
    .admin("your-api-key", "your-api-secret") \
    .build()

students = client.open_api().student().list(OpenStudentQueryRequest(
    orgCode="school001",
    pageNum=0,
    pageSize=20,
))

print(students.content)

Streaming Chat

Use completionsStream when the reply must be displayed as it is generated.

from lingxinai.psychology.types import ChatQueryRequest

for chunk in student_client.open_api().chat().completionsStream(ChatQueryRequest(
    botId=237,
    content="我最近压力很大，想聊一聊",
)):
    print(chunk, end="", flush=True)
print()

Client Identity

Create one client per identity. Do not mix admin, user, student, or parent calls on the same client.

admin_client = PsychologyClient.builder() \
    .admin("your-api-key", "your-api-secret") \
    .build()

user_client = PsychologyClient.builder() \
    .user("your-api-key", "your-api-secret", "external-user-001") \
    .build()

student_client = PsychologyClient.builder() \
    .student("your-api-key", "your-api-secret", 10001) \
    .build()

parent_client = PsychologyClient.builder() \
    .access_token(parent_login_response.token) \
    .build()

API Groups

• auth() - login helpers
• org() - organization APIs
• dept() - department/class APIs
• student() - student APIs
• orgUser() - organization staff APIs
• device() - device APIs
• bot() - bot APIs
• chat() - chat APIs
• course() - micro course APIs
• algo() - HTTP ASR/TTS and realtime ASR/TTS WebSocket APIs
• crisis() - crisis warning APIs
• meditation() - meditation APIs
• overview() - dashboard/statistics APIs
• parent() - parent-side APIs
• quickInterview() - quick interview APIs
• activity() - activity APIs
• scale() - scale/interview runtime APIs

Realtime ASR

from lingxinai.psychology.types import RealtimeAsrOptions

ws = student_client.open_api().algo().realtimeAsr(RealtimeAsrOptions(
    voiceFormat=1,
    inputSampleRate=16000,
))

ws.send('{"type":"start"}')
ws.send(audio_chunk, opcode=0x2)
ws.send('{"type":"end"}')
print(ws.recv())
ws.close()

Bidirectional TTS

Use bidirectionalTts when text arrives in chunks and audio must be streamed back at the same time.

import json

ws = student_client.open_api().algo().bidirectionalTts()
ws.send(json.dumps({"type": "start", "voiceCode": "zh_female_linjianvhai_moon_bigtts", "format": "mp3"}))
ws.send(json.dumps({"type": "text", "text": "我们先做一次简单的"}))
ws.send(json.dumps({"type": "text", "text": "放松练习。"}))
ws.send(json.dumps({"type": "end"}))
ws.close()

Realtime ASR/TTS Smoke Test

The repository includes a script for checking both realtime WebSocket APIs. It reads credentials from .env.local, writes TTS audio to tmp/realtime-tts-output.mp3, and feeds realtime-voice-smoke.wav to realtime ASR.

python test/realtime_asr_tts_stream.py

Optional arguments:

python test/realtime_asr_tts_stream.py \
  --asr-audio /path/to/16k-mono-pcm.wav \
  --voice-code zh_female_linjianvhai_moon_bigtts \
  --text "你好，这是实时语音合成测试。"

Verification

python -m unittest discover -s test
python -m build --sdist --wheel
python -m twine check dist/*