DeepEar Analysis Skill

This skill wraps the DeepEar analysis workflow, allowing you to trigger comprehensive financial signal analysis via an API.

API Endpoints

POST /analyze

Triggers the DeepEar analysis workflow.

Request Body:

json
1{
2  "query": "A-share tech sector",
3  "sources": "all",
4  "wide": 10,
5  "depth": "auto",
6  "concurrency": 5,
7  "update_from": "20240203_110000"
8}

• query (optional): The user intent or topic to analyze.
• sources (optional): News sources ('all', 'financial', 'social', 'tech', or comma-separated). Default: 'all'.
• wide (optional): Number of items to fetch per source. Default: 10.
• depth (optional): Analysis depth ('auto' or integer). Default: 'auto'.
• concurrency (optional): Number of concurrent threads. Default: 5.
• update_from (optional): Provide a run_id to update an existing analysis with new market data/logic.

Response:

json
1{
2  "job_id": "550e8400-e29b-41d4-a716-446655440000",
3  "status": "pending",
4  "submitted_at": "2024-02-04T12:00:00.000000"
5}

GET /status/{job_id}

Check the progress and get the result of the analysis job.

Response (Completed):

json
1{
2  "job_id": "550e8400-e29b-41d4-a716-446655440000",
3  "status": "completed",
4  "result": {
5    "report_path": "/path/to/DeepEar/reports/daily_report_20240204_1205.html",
6    "message": "Analysis completed successfully."
7  },
8  "timestamp": "2024-02-04T12:05:00.000000"
9}

Usage

Run the server:

bash
1uv run skills/deepear/scripts/server.py