title	Chatbot AI v2
emoji	🚀
colorFrom	red
colorTo	red
sdk	docker
app_port	8501
tags	streamlit
pinned	false
short_description	https://huggingface.co/docs/hub/en/spaces-sdks-streamlit

Welcome to Streamlit!

Edit /src/streamlit_app.py to customize this app to your heart's desire. ❤️

If you have any questions, checkout our documentation and community forums.

Returning

The venv for this project lives outside the repo, at ~/.venvs/chatbot_ai_v2 — this keeps the local folder free of large site-packages/metadata to clean up. It's pinned to Python 3.8 because torch dropped Intel-Mac wheels after 2.2.2, and 3.8 is the last interpreter pip can still resolve that version for automatically.

cd chatbot_ai_v2

# one-time only, if ~/.venvs/chatbot_ai_v2 doesn't exist yet
python3.8 -m venv ~/.venvs/chatbot_ai_v2

# every time you return: activate the venv
source ~/.venvs/chatbot_ai_v2/bin/activate

# one-time per venv: install deps (pip resolves torch==2.2.2 + transformers==4.46.3
# automatically on Python 3.8 / Intel Mac — no manual pinning needed)
pip install -r requirements.txt

# run the app
streamlit run src/streamlit_app.py

Tip: add this to ~/.zshrc to activate with one word instead of the full path:

activate() { source ~/.venvs/"$1"/bin/activate; }

then just run activate chatbot_ai_v2.

Deploying to the Space is unaffected by any of this — Hugging Face builds the Dockerfile, which installs from requirements.txt directly on its own (Linux) infrastructure.

Enabling real AI replies

By default this Space has no HF_TOKEN configured, so it runs on free local sentiment analysis plus canned empathetic replies — $0 cost to the maintainer. To get real LLM-generated replies:

• Paste your own token: enter a Hugging Face token in the sidebar. It's used for your session only and is never stored.
• Or duplicate the Space: click "Duplicate this Space," make your copy private, and add HF_TOKEN under Settings → Variables and secrets so it's always on.

Get a free token at https://huggingface.co/settings/tokens.

For local testing, .env and .streamlit/secrets.toml are gitignored, so either is a safe place to stash your own HF_TOKEN without risking a commit.

Note: the app reads HF_TOKEN via os.environ, and a plain .env file is not auto-loaded (no python-dotenv dependency). To use one locally, source it into your shell before running the app:

# chatbot_ai_v2/.env
HF_TOKEN=hf_your_real_token_here

set -a; source .env; set +a
streamlit run src/streamlit_app.py