initial public

Author: lspahija

.gitignore

@@ -0,0 +1,133 @@
+# Byte-compiled / optimized / DLL files
+__pycache__/
+*.py[cod]
+*$py.class
+
+# C extensions
+*.so
+
+# Distribution / packaging
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+pip-wheel-metadata/
+share/python-wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+MANIFEST
+
+# PyInstaller
+#  Usually these files are written by a python script from a template
+#  before PyInstaller builds the exe, so as to inject date/other infos into it.
+*.manifest
+*.spec
+
+# Installer logs
+pip-log.txt
+pip-delete-this-directory.txt
+
+# Unit test / coverage reports
+htmlcov/
+.tox/
+.nox/
+.coverage
+.coverage.*
+.cache
+nosetests.xml
+coverage.xml
+*.cover
+*.py,cover
+.hypothesis/
+.pytest_cache/
+
+# Translations
+*.mo
+*.pot
+
+# Django stuff:
+*.log
+local_settings.py
+db.sqlite3
+db.sqlite3-journal
+
+# Flask stuff:
+instance/
+.webassets-cache
+
+# Scrapy stuff:
+.scrapy
+
+# Sphinx documentation
+docs/_build/
+
+# PyBuilder
+target/
+
+# Jupyter Notebook
+.ipynb_checkpoints
+
+# IPython
+profile_default/
+ipython_config.py
+
+# pyenv
+.python-version
+
+# pipenv
+#   According to pypa/pipenv#598, it is recommended to include Pipfile.lock in version control.
+#   However, in case of collaboration, if having platform-specific dependencies or dependencies
+#   having no cross-platform support, pipenv may install dependencies that don't work, or not
+#   install all needed dependencies.
+#Pipfile.lock
+
+# PEP 582; used by e.g. github.com/David-OConnor/pyflow
+__pypackages__/
+
+# Celery stuff
+celerybeat-schedule
+celerybeat.pid
+
+# SageMath parsed files
+*.sage.py
+
+# Environments
+.env
+.venv
+env/
+venv/
+ENV/
+env.bak/
+venv.bak/
+
+# Spyder project settings
+.spyderproject
+.spyproject
+
+# Rope project settings
+.ropeproject
+
+# mkdocs documentation
+/site
+
+# mypy
+.mypy_cache/
+.dmypy.json
+dmypy.json
+
+# Pyre type checker
+.pyre/
+
+.idea
+
+.DS_Store

Dockerfile

@@ -0,0 +1,13 @@
+FROM tiangolo/uvicorn-gunicorn-fastapi:python3.10-slim
+
+WORKDIR /
+
+ENV MAX_WORKERS=5
+
+COPY ./requirements.txt /app/requirements.txt
+
+RUN pip install --no-cache-dir --upgrade -r /app/requirements.txt
+
+RUN apt-get update && apt-get install -y ffmpeg
+
+COPY ./app /app

README.md

@@ -0,0 +1,18 @@
+# The Ultimate AI Assistant
+
+Now you can use ChatGPT from anywhere! Talk to ChatGPT when driving in the car. Bounce ideas off the ultimate
+personal assistant.
+
+To get it running:
+```
+(while inside ultimate-ai-assistant project root directory)
+docker build -t ultimate_ai_assistant .  
+docker run -d -e OPENAI_API_KEY='<YOUR_API_KEY>' -p 8000:80 ultimate_ai_assistant
+```
+
+and then navigate to `localhost:8000`
+
+
+Check out the demo at https://ultimate-ai-assistant.up.railway.app/
+
+Star the repo if you like it!

app/main.py

@@ -0,0 +1,112 @@
+import json
+import time
+from fastapi import FastAPI, UploadFile, BackgroundTasks, Header
+from fastapi.responses import FileResponse
+import os
+import openai
+import shutil
+import uuid
+from gtts import gTTS
+import ffmpeg
+import base64
+from fastapi.staticfiles import StaticFiles
+
+AI_COMPLETION_MODEL = "gpt-3.5-turbo"
+app = FastAPI()
+
+
+@app.post("/inference")
+async def infer(audio: UploadFile, background_tasks: BackgroundTasks,
+                conversation: str = Header(default=None)) -> FileResponse:
+    print("received request")
+    start_time = time.time()
+
+    user_prompt = await transcribe(audio)
+    ai_response = await get_completion(user_prompt, conversation)
+
+    output_audio_filepath = to_audio(ai_response)
+    background_tasks.add_task(delete_file, output_audio_filepath)
+
+    print('total processing time:', time.time() - start_time, 'seconds')
+
+    return FileResponse(path=output_audio_filepath, media_type="audio/mpeg",
+                        headers={"text": construct_response_header(user_prompt, ai_response)})
+
+
+app.mount("/", StaticFiles(directory="app/static", html=True), name="static")
+
+
+async def transcribe(audio):
+    start_time = time.time()
+    initial_filepath = f"/tmp/{uuid.uuid4()}{audio.filename}"
+
+    with open(initial_filepath, "wb+") as file_object:
+        shutil.copyfileobj(audio.file, file_object)
+
+    converted_filepath = f"/tmp/ffmpeg-{uuid.uuid4()}{audio.filename}"
+
+    print("running through ffmpeg")
+    (
+        ffmpeg
+        .input(initial_filepath)
+        .output(converted_filepath, loglevel="error")
+        .run()
+    )
+    print("ffmpeg done")
+
+    delete_file(initial_filepath)
+
+    read_file = open(converted_filepath, "rb")
+
+    print("calling whisper")
+    transcription = (await openai.Audio.atranscribe("whisper-1", read_file))["text"]
+    print("STT response received from whisper in", time.time() - start_time, 'seconds')
+    print('user prompt:', transcription)
+
+    delete_file(converted_filepath)
+
+    return transcription
+
+
+async def get_completion(user_prompt, conversation_thus_far):
+    start_time = time.time()
+    messages = [
+        {"role": "system", "content": "You are a helpful assistant."},
+        {"role": "user",
+         "content": "You are a helpful assistant with a voice interface. Keep your responses limited to a single sentence of reasonable length. Make sure your response is in English, regardless of the language that the user's input is in."},
+    ]
+
+    messages.extend(json.loads(base64.b64decode(conversation_thus_far)))
+
+    messages.append({"role": "user", "content": user_prompt})
+
+    print("calling", AI_COMPLETION_MODEL)
+    res = await openai.ChatCompletion.acreate(model=AI_COMPLETION_MODEL, messages=messages, timeout=15)
+    print("response received from", AI_COMPLETION_MODEL, "in", time.time() - start_time, "seconds")
+
+    completion = res['choices'][0]['message']['content']
+    print(AI_COMPLETION_MODEL, "response:", completion)
+
+    return completion
+
+
+def to_audio(text):
+    start_time = time.time()
+
+    tts = gTTS(text)
+    filepath = f"/tmp/{uuid.uuid4()}.mp3"
+    tts.save(filepath)
+
+    print('TTS time:', time.time() - start_time, 'seconds')
+    return filepath
+
+
+def delete_file(filepath: str):
+    os.remove(filepath)
+
+
+def construct_response_header(user_prompt, ai_response):
+    return base64.b64encode(
+        json.dumps(
+            [{"role": "user", "content": user_prompt}, {"role": "assistant", "content": ai_response}]).encode(
+            'utf-8')).decode("utf-8")

app/static/favicon-16.png

@@ -0,0 +1,25 @@
+<!DOCTYPE html>
+<html lang="en">
+<head>
+    <link href="style.css" rel="stylesheet" type="text/css">
+    <script src="https://kit.fontawesome.com/80276f8bce.js" crossorigin="anonymous"></script>
+    <script type="text/javascript" src="script.js"></script>
+    <!-- Cloudflare Web Analytics -->
+    <script defer src='https://static.cloudflareinsights.com/beacon.min.js'
+            data-cf-beacon='{"token": "2929d361bab44c8392e78bf5421d742f"}'></script>
+    <!-- End Cloudflare Web Analytics -->
+    <link rel="icon" type="image/png" sizes="16x16" href="./favicon-16.png">
+    <link rel="icon" type="image/png" sizes="32x32" href="./favicon-32.png">
+    <link rel="icon" type="image/png" sizes="96x96" href="./favicon-96.png">
+</head>
+<meta charset="UTF-8">
+<title>Record audio</title>
+</head>
+<body>
+<div id="container">
+    <button id="record-button"><i class="fas fa-microphone"></i></button>
+    <div id="error-message">Error occurred</div>
+    <div id="spinner"></div>
+</div>
+</body>
+</html>