Update space

app.py

@@ -1,63 +1,256 @@
 import gradio as gr
-from huggingface_hub import InferenceClient
+import os
+import string
+from pymongo import MongoClient
+from openai import AsyncOpenAI, OpenAI
+import copy
+from constants import *
+import asyncio
+import string as st
+from opik.integrations.openai import track_openai
+from opik import track
+from bson.objectid import ObjectId
+import opik
 
-"""
-For more information on `huggingface_hub` Inference API support, please check the docs: https://huggingface.co/docs/huggingface_hub/v0.22.2/en/guides/inference
-"""
-client = InferenceClient("HuggingFaceH4/zephyr-7b-beta")
 
+oClient = opik.Opik()
+mdb = MongoClient(
+    os.getenv("MONGO_URI")
+)  # , "mongodb://localhost:27017/"))  # Default to localhost if not set
+aclient = track_openai(AsyncOpenAI())
+client = track_openai(OpenAI())
+db = mdb["Mindware"]
 
-def respond(
-    message,
-    history: list[tuple[str, str]],
-    system_message,
-    max_tokens,
-    temperature,
-    top_p,
-):
-    messages = [{"role": "system", "content": system_message}]
 
-    for val in history:
-        if val[0]:
-            messages.append({"role": "user", "content": val[0]})
-        if val[1]:
-            messages.append({"role": "assistant", "content": val[1]})
+def purge(d):
+    """
+    Recursively collect all leaf nodes.
+    """
+    result = {}
+    for k, v in d.items():
+        if k == "chat_history":
+            pass
+        if isinstance(v, dict):
+            result.update(purge(v))
+        elif isinstance(v, list):
+            for idx, d in enumerate(v):
 
-    messages.append({"role": "user", "content": message})
+                if isinstance(d, dict):
+                    try:
+                        for k1 in d.keys():
+                            result[k1] = []
+                        for k2, v2 in d.items():
+                            result[k2].append(v2)
+                    except Exception as e:
+                        print("Error! Error!", e)
+                        if k not in result.keys():
+                            result[k] = []
+                        result[k].append(d)
+                else:
+                    result[k] = v
 
-    response = ""
+            else:
+                result[k] = v
+        else:
+            result[k] = v
+    return result
+
+
+def deploy(d):
+    """
+    Recursively deploy all leaf nodes.
+    """
+    result = {}
+    result.update(purge(d))
+
+    return result
 
-    for message in client.chat_completion(
-        messages,
-        max_tokens=max_tokens,
+
+async def chat(prompt, model="gpt-4"):
+    text = await aclient.chat.completions.create(
+        model=model, messages=[{"role": "user", "content": prompt}]
+    )
+    return text.choices[0].message.content
+
+
+async def chat_generator(prompt: str, user: dict = None):
+    response = await aclient.chat.completions.create(
+        model="gpt-4",
+        messages=[
+            {
+                "role": "user",
+                "content": prompt.format(**deploy(user)) if user else prompt,
+            }
+        ],
         stream=True,
-        temperature=temperature,
-        top_p=top_p,
-    ):
-        token = message.choices[0].delta.content
+    )
+    # reply_chunks = []
+    async for chunk in response:
+        if chunk and chunk.choices[0].delta.content:
+
+            # reply_chunks.append(chunk.choices[0].delta.content)
+            yield chunk.choices[0].delta.content
+
+
+def get_or_init_user(reply, userId):
+
+    users = list(db["users"].find({"userId": userId}))
+    if not users:
+        user = dict(**copy.deepcopy(USER_TEMPLATE))
+        user.update({"userId": userId})
+        user.update({"user_query": reply})
+        print("user created:", user)
+
+        # user.update({"chat_history": history})
+
+    else:
+        user = users[0]
+        user.pop("_id")
+        user = dict(users[0])
+        user.update({"user_query": reply})
+
+    return user
+
+
+async def search(query, n=5):
+    embed = await aclient.embeddings.create(input=query, model="text-embedding-3-small")
+
+    query_embedding = embed.data[0].embedding
+    pipeline = [
+        {
+            "$vectorSearch": {
+                "queryVector": query_embedding,
+                "path": "embedding",
+                "index": "arrestor",
+                "score": {"$meta": "vectorSearchScore"},
+                "filter": {"class": "THERAPIST"},
+                "numCandidates": 850,
+                "limit": n,
+            }
+        }
+    ]
+
+    projection = [{"$project": {"embedding": 0}}]
+
+    pipeline += projection
+
+    docs = db["runway"].aggregate(pipeline)
+    return list(docs)
+
+
+async def agentic_search(query, n=5):
+    results = await search(query, n * 5)
+    tasks = [chat(ARAG_PROMPT.format(query=query, doc=doc)) for doc in results]
+    is_context = asyncio.gather(*tasks)
+
+    docs = []
+    for doc, reply in zip(results, is_context):
+        if reply == "True":
+            docs.append(doc)
+            if len(docs) >= n:
+                break
+
+    return list(docs)
+
+
+async def update_user(response, user):
+    user["last_question"] = response
+    user["chat_history"].append({"role": "user", "content": user["user_query"]})
+    user["chat_history"].append({"role": "assistant", "content": response})
+    db["users"].delete_many({"userId": user["userId"]})
+    db["users"].insert_one(user)
+
+    print("Updated", user["userId"], user["user_query"], "reply:", response)
+
+
+async def add_background_tasks(task):
+    """a dummy wrapper to be replaced with FastAPI background tasks"""
+    await task
+
+
+punc_removal = str.maketrans("", "", string.punctuation.replace("_", ""))
+
+
+async def escalate(user):
+    print(f"user {user.get('name')} is not working, escalating to clinician")
+
+
+async def update_docs(user):
+    print(f"updating docs for {user['name']}: {user['cache']}")
+
+
+@track
+async def handle_chat(reply, userId):
+    """
+    Handle the chat response and update the user
+    """
+    user = get_or_init_user(reply, userId)
+    prompt = BASE_PROMPT
+    tasks = [
+        chat(prompt=p.format(**deploy(user)))
+        for p in [INTENT_PROMPT, RISK_PROMPT, CACHE_PROMPT, INTENSITY_PROMPT]
+    ]
+    responses = await asyncio.gather(*tasks)
+
+    if responses[2]:
+        user["cache"] = responses[2]
+        await add_background_tasks(update_docs(user))
+
+    intent = responses[0].upper().translate(punc_removal).replace(" ", "_")
+
+    if intent == "ACTIVE_SPEAKING":
+        prompt += SPEAKING_PROMPT
+    elif intent == "VALIDATION_SEEK":
+        prompt += VALIDATION_PROMPT
+    elif intent == "OVERWHELMED":
+        prompt += OVERWHELMED_PROMPT
+        await asyncio.sleep(5)
+    elif intent == "REMOTE_REFERRAL":
+        results = await search(user["cache"], n=5)
+        prompt += REMOTE_PROMPT.format(results=results, **deploy(user))
+    elif intent == "NEUTRAL_STOP":
+        prompt += STOP_PROMPT
+    elif intent == "END_OF_NARRATIVE":
+        prompt += END_PROMPT
+    else:
+        print("Unknown response of intent detection:", responses[0])
+
+    if responses[1].upper().translate(punc_removal).replace(" ", "_") == "HIGH_RISK":
+        prompt += HIGH_RISK_PROMPT
+        await add_background_tasks(escalate(user))
+
+    response = ""
+    async for word in chat_generator(prompt, user):
+        if word:
+            response += word
+            yield word
+
+    await add_background_tasks(update_user(response, user))
+
+    return
+
 
-        response += token
-        yield response
+async def respond(message, history, id):
+    """
+    Respond to the chat message and return the response.
+    """
+    reply = ""
+    async for r in handle_chat(message, id):
+        if r:
+            reply += r
+            yield reply
 
 
 """
 For information on how to customize the ChatInterface, peruse the gradio docs: https://www.gradio.app/docs/chatinterface
 """
-demo = gr.ChatInterface(
-    respond,
-    additional_inputs=[
-        gr.Textbox(value="You are a friendly Chatbot.", label="System message"),
-        gr.Slider(minimum=1, maximum=2048, value=512, step=1, label="Max new tokens"),
-        gr.Slider(minimum=0.1, maximum=4.0, value=0.7, step=0.1, label="Temperature"),
-        gr.Slider(
-            minimum=0.1,
-            maximum=1.0,
-            value=0.95,
-            step=0.05,
-            label="Top-p (nucleus sampling)",
-        ),
-    ],
-)
+with gr.Blocks() as demo:
+    id = gr.Textbox(str(ObjectId()), label="userID")
+    gr.ChatInterface(
+        fn=respond,
+        type="messages",
+        additional_inputs=[id],
+    )
 
 
 if __name__ == "__main__":

requirements.txt

@@ -1 +1,230 @@
-huggingface_hub==0.25.2
+huggingface_hub==0.25.2
+aiofiles==24.1.0
+aiohappyeyeballs==2.6.1
+aiohttp==3.12.14
+aiosignal==1.4.0
+annotated-types==0.7.0
+anyio==4.9.0
+argon2-cffi==25.1.0
+argon2-cffi-bindings==25.1.0
+arrow==1.3.0
+asgiref==3.8.1
+asttokens==3.0.0
+async-lru==2.0.5
+attrs==25.3.0
+Authlib==1.3.1
+babel==2.17.0
+beautifulsoup4==4.13.4
+bleach==6.2.0
+boto3-stubs==1.39.9
+botocore-stubs==1.38.46
+Brotli==1.1.0
+certifi==2025.7.14
+cffi==1.17.1
+charset-normalizer==3.4.2
+click==8.2.1
+cobble==0.1.4
+colorama==0.4.6
+coloredlogs==15.0.1
+comm==0.2.2
+contourpy==1.3.2
+cryptography==45.0.5
+cycler==0.12.1
+debugpy==1.8.15
+decorator==5.2.1
+defusedxml==0.7.1
+dill==0.3.8
+distro==1.9.0
+dnspython==2.7.0
+executing==2.2.0
+fastapi==0.116.1
+fastjsonschema==2.21.1
+ffmpy==0.6.1
+filelock==3.18.0
+flake8==3.9.2
+flatbuffers==25.2.10
+fonttools==4.59.0
+fqdn==1.5.1
+frozenlist==1.7.0
+fsspec==2025.7.0
+gradio==5.41.0
+gradio_client==1.11.0
+groovy==0.1.2
+h11==0.16.0
+httpcore==1.0.9
+httpx==0.27.0
+huggingface-hub==0.34.3
+humanfriendly==10.0
+idna==3.10
+imageio==2.37.0
+imageio-ffmpeg==0.6.0
+importlib_metadata==8.7.0
+impyute==0.0.8
+iniconfig==2.1.0
+ipykernel==6.29.5
+ipython==9.4.0
+ipython_pygments_lexers==1.1.1
+ipywidgets==8.1.7
+isoduration==20.11.0
+jedi==0.19.2
+Jinja2==3.1.6
+jiter==0.10.0
+json5==0.12.0
+jsonpointer==3.0.0
+jsonschema==4.25.0
+jsonschema-specifications==2025.4.1
+jupyter==1.1.1
+jupyter-console==6.6.3
+jupyter-events==0.12.0
+jupyter-lsp==2.2.6
+jupyter_client==8.6.3
+jupyter_core==5.8.1
+jupyter_server==2.16.0
+jupyter_server_terminals==0.5.3
+jupyterlab==4.4.5
+jupyterlab_pygments==0.3.0
+jupyterlab_server==2.27.3
+jupyterlab_widgets==3.0.15
+kiwisolver==1.4.8
+lark==1.2.2
+litellm==1.74.4
+lxml==6.0.0
+magika==0.6.2
+mammoth==1.9.1
+markdown-it-py==3.0.0
+markdownify==1.1.0
+markitdown==0.1.2
+MarkupSafe==3.0.2
+matplotlib==3.10.3
+matplotlib-inline==0.1.7
+mccabe==0.6.1
+mdurl==0.1.2
+mistune==3.1.3
+moviepy==2.2.1
+mpmath==1.3.0
+multidict==6.6.3
+multiprocess==0.70.16
+mypy-boto3-bedrock-runtime==1.39.7
+nbclient==0.10.2
+nbconvert==7.16.6
+nbformat==5.10.4
+nbqa==1.8.5
+nest-asyncio==1.6.0
+notebook==7.4.5
+notebook_shim==0.2.4
+numpy==2.3.1
+onnxruntime==1.22.1
+openai==1.97.0
+opik==1.8.6
+orjson==3.11.1
+overrides==7.7.0
+packaging==25.0
+pandas==2.3.1
+pandas-stubs==2.2.2.240603
+pandocfilters==1.5.1
+parso==0.8.4
+pdfminer.six==20250506
+pillow==11.3.0
+platformdirs==4.3.8
+pluggy==1.6.0
+proglog==0.1.12
+prometheus_client==0.22.1
+prompt_toolkit==3.0.51
+propcache==0.3.2
+protobuf==6.31.1
+psutil==7.0.0
+pure_eval==0.2.3
+pycodestyle==2.12.0
+pycodestyle_magic==0.5
+pycparser==2.22
+pydantic==2.11.7
+pydantic-settings==2.10.1
+pydantic_core==2.33.2
+pydub==0.25.1
+pyflakes==2.3.1
+Pygments==2.19.2
+pymongo==4.13.2
+pyparsing==3.2.3
+pyreadline3==3.5.4
+pytest==8.4.1
+python-dateutil==2.9.0.post0
+python-dotenv==1.1.1
+python-json-logger==3.3.0
+python-multipart==0.0.20
+python-pptx==1.0.2
+pytz==2025.2
+pywin32==311
+pywinpty==2.0.15
+PyYAML==6.0.2
+pyzmq==27.0.0
+RapidFuzz==3.13.0
+referencing==0.36.2
+regex==2024.11.6
+requests==2.32.4
+rfc3339-validator==0.1.4
+rfc3986-validator==0.1.1
+rfc3987-syntax==1.1.0
+rich==14.0.0
+rpds-py==0.26.0
+ruff==0.12.7
+safehttpx==0.1.6
+schwab-py==1.3.0
+schwabdev==2.1.1
+semantic-version==2.10.0
+Send2Trash==1.8.3
+sentry-sdk==2.33.0
+shellingham==1.5.4
+six==1.17.0
+sniffio==1.3.1
+soupsieve==2.7
+sqlparse==0.5.0
+stack-data==0.6.3
+starlette==0.47.1
+sympy==1.14.0
+tenacity==9.1.2
+terminado==0.18.1
+tiktoken==0.9.0
+tinycss2==1.4.0
+tokenize-rt==5.2.0
+tokenizers==0.21.2
+tomli==2.0.1
+tomlkit==0.13.3
+tornado==6.5.1
+tqdm==4.67.1
+traitlets==5.14.3
+typer==0.16.0
+types-awscrt==0.27.4
+types-bleach==6.1.0.20240331
+types-colorama==0.4.15.20240311
+types-croniter==2.0.0.20240423
+types-decorator==5.1.8.20240310
+types-docutils==0.21.0.20240423
+types-html5lib==1.1.11.20240228
+types-jsonschema==4.22.0.20240610
+types-Markdown==3.6.0.20240316
+types-Pillow==10.2.0.20240520
+types-psutil==5.9.5.20240516
+types-Pygments==2.18.0.20240506
+types-python-dateutil==2.9.0.20240316
+types-pytz==2024.1.0.20240417
+types-PyYAML==6.0.12.20240311
+types-requests==2.32.0.20240602
+types-s3transfer==0.13.0
+types-setuptools==70.0.0.20240524
+types-tqdm==4.66.0.20240417
+typing-inspection==0.4.1
+typing_extensions==4.14.1
+tzdata==2025.2
+uri-template==1.3.0
+urllib3==2.5.0
+uuid6==2025.0.1
+uvicorn==0.35.0
+wcwidth==0.2.13
+webcolors==24.11.1
+webencodings==0.5.1
+websocket-client==1.8.0
+websockets==12.0
+widgetsnbextension==4.0.14
+xlsxwriter==3.2.5
+yarl==1.20.1
+zipp==3.23.0