I’m leveraging the DeepLearning.ai platform to complete Agentic AI with Andrew Ng. I’m setting-up my own environment on a MacBook Pro using VS Code, Conda, and the provided requirements.txt file (below), which includes the package mistralai. There is an error when I try to install this particular package (see below). PyPI currently shows this package as quarantined, which means pip cannot install it right now. PyPI says quarantined projects are “not installable by clients.”
# In the active conda environment with Python 3.12.13 installed:
$ python -m pip install -r requirements.txt
Collecting aisuite==0.1.11 (from -r requirements.txt (line 2)) Using cached aisuite-0.1.11-py3-none-any.whl.metadata (9.4 kB) Collecting anthropic (from -r requirements.txt (line 3)) Using cached anthropic-0.101.0-py3-none-any.whl.metadata (3.2 kB) Collecting docstring-parser (from -r requirements.txt (line 4)) Using cached docstring_parser-0.18.0-py3-none-any.whl.metadata (3.5 kB) Collecting markdown (from -r requirements.txt (line 5)) Using cached markdown-3.10.2-py3-none-any.whl.metadata (5.1 kB)
ERROR: Could not find a version that satisfies the requirement mistralai (from versions: none)
ERROR: No matching distribution found for mistralai
- Is the
mistralaipackage essential? - Is there a work-around for the installation error?
For reference, here are the contents of the requirements.txt file. (Note that the “W” in “Wikipedia” should probably be in lowercase.
# === Agent + LLM Tools ===
aisuite==0.1.11
anthropic
docstring-parser
markdown
mistralai
openai
qrcode
tavily-python>=0.7.12
textstat
vertexai
# === Web Framework + API ===
fastapi
pydantic
pydantic[email]
python-dotenv
python-multipart
requests
sqlalchemy
uvicorn
# === Notebook Experience ===
ipywidgets
jupyter_server
nbclassic
notebook
# === Data Analysis / Display (Optional Enhancements) ===
duckdb
matplotlib
pandas
seaborn
tabulate
tinydb
# === Machine Learning / NLP (Optional Enhancements) ===
jinja2
psycopg2-binary
scikit-learn
Wikipedia
Thank you!