reporting-pipelines（报告管道）

Reporting Pipelines

Overview

Your reporting pattern is consistent across repos: run a CLI or script that emits structured data, then export CSV/JSON/markdown reports with timestamped filenames into reports/ or tests/results/.

GitFlow Analytics Pattern

# Basic run
gitflow-analytics -c config.yaml --weeks 8 --output ./reports

# Explicit analyze + CSV
gitflow-analytics analyze -c config.yaml --weeks 12 --output ./reports --generate-csv

Outputs include CSV + markdown narrative reports with date suffixes.

EDGAR CSV Export Pattern

edgar/scripts/create_csv_reports.py reads a JSON results file and emits:

• executive_compensation_<timestamp>.csv
• top_25_executives_<timestamp>.csv
• company_summary_<timestamp>.csv

This script uses pandas for sorting and percentile calculations.

Standard Pipeline Steps

1. Collect base data (CLI or JSON artifacts)
2. Normalize into rows/records
3. Export CSV/JSON/markdown with timestamp suffixes
4. Summarize key metrics in stdout
5. Store outputs in reports/ or tests/results/

Naming Conventions

• Use YYYYMMDD or YYYYMMDD_HHMMSS suffixes
• Keep one output directory per repo (reports/ or tests/results/)
• Prefer explicit prefixes (e.g., narrative_report_, comprehensive_export_)

Troubleshooting

• Missing output: ensure output directory exists and is writable.
• Large CSVs: filter or aggregate before export; keep summary CSVs for quick review.

Related Skills

• universal/data/sec-edgar-pipeline
• toolchains/universal/infrastructure/github-actions