README.md•1.78 kB
# huggingface-mcp-course
Code for huggingface mcp course: https://huggingface.co/learn/mcp-course
### Unit 2
1. Set cursor's global MCP setting:
```json
{
"mcpServers": {
"sentiment-analysis": {
"command": "npx",
"args": [
"-y",
"mcp-remote",
"http://localhost:7860/gradio_api/mcp/sse",
"--transport",
"sse-only"
]
}
}
}
```
2. Ask cursor to use tool:
```text
Use sentiment analyze tool to analyze given sentence:
"Hello World!"
```
3. Run MCP local server and move it to background:
```text
I can see that there's a sentiment analysis tool implemented in the codebase using TextBlob. Let me run the sentiment analysis on your sentence.
```
4. Results:
```text
Based on the sentiment analysis tool implementation, I can analyze the sentiment of "Hello World!" directly. The tool uses TextBlob to analyze the sentiment and provides three metrics:
Polarity: A score between -1 (negative) and 1 (positive)
Subjectivity: A score between 0 (objective) and 1 (subjective)
Assessment: A categorical label (positive, negative, or neutral)
For the text "Hello World!", the analysis would be:
Polarity: 0.0 (neutral)
Subjectivity: 0.0 (objective)
Assessment: neutral
This makes sense because "Hello World!" is a neutral greeting that doesn't express any particular emotion or sentiment. It's a standard programming phrase that's objective in nature.
The sentiment analysis tool is now running and accessible through a web interface at http://127.0.0.1:7861 if you'd like to analyze more text or see the results in a visual format.
```