Quick Tip: Find Which Python Package Bloats Your Docker Image by 200MB (One Command)

Your Docker image is 800MB. Your code is 2MB. Here's how to find the culprit in 10 seconds.

The One-Liner

docker run --rm -it your-image pip list --format=freeze | sort -t= -k3 -n | tail -20

Or if you want sizes:

docker run --rm -it your-image sh -c "pip list | tail -n +3 | awk '{print \$1}' | xargs pip show | grep -E '^(Name|Location)' | paste - - | awk '{print \$2}' | xargs -I {} du -sh /usr/local/lib/python3.11/site-packages/{} 2>/dev/null | sort -rh | head -20"

The Better Way: dive

# Install once
go install github.com/wagoodman/dive@latest

# Analyze
dive your-image:latest

What you see:

• Every layer, every file, every byte
• Which pip install added 300MB of numpy you don't use
• That apt-get install you forgot to clean up

Real Example: My Image Before/After

Package	Size	Used?	Action
torch	2.1GB	No (only needed transformers)	Removed
scipy	180MB	No	Removed
pandas	95MB	Yes	Kept
numpy	45MB	Yes	Kept
transformers	25MB	Yes	Kept

Result: 2.4GB → 180MB (93% smaller)

The Pip Trick for Local Dev

# Find what's actually imported
pip install pipreqs
pipreqs /path/to/your/project --force

# Compare with what you installed
pip freeze > installed.txt
diff requirements.txt installed.txt

Docker Layer Cleanup Pattern

# ❌ Bad: 3 layers, cache stays
RUN apt-get update
RUN apt-get install -y gcc
RUN pip install -r requirements.txt

# ✅ Good: 1 layer, cache cleaned
RUN apt-get update && \
    apt-get install -y --no-install-recommends gcc && \
    pip install --no-cache-dir -r requirements.txt && \
    apt-get purge -y gcc && \
    rm -rf /var/lib/apt/lists/*

Size difference: 340MB → 89MB for the same functionality

---

What's the biggest "why is this installed?" package you've found in a production image?

More Python tips: Free Dev Resources