Finance Report Analyzer
Generate financial analysis reports from uploaded Excel/PDF files with LLM-powered analysis, competitor analysis, web search for news, and inline SVG sparkline trend charts.
Architecture
The tool works in a hybrid pipeline:
- Python script → Extract data, build tables/charts, render HTML/PDF (deterministic, zero-cost)
- LLM (OpenClaw agent) → Write in-depth analysis for each section + competitor analysis + news summary
- Web search → Fetch recent company news and hot topics
Report Sections
| # | Section | Source |
|---|---|---|
| 一 | 盈利能力分析 | Script(tables) + LLM(analysis) |
| 二 | 资产负债分析 | Script(tables) + LLM(analysis) |
| 三 | 现金流分析 | Script(tables) + LLM(analysis) |
| 四 | 每股指标与效率 | Script(tables) + LLM(analysis) |
| 五 | 行业分析 | LLM |
| 六 | 关键竞争对手分析 | LLM(competitor) |
| 七 | 风险与机遇分析 | LLM |
| 八 | 近期热点新闻 | Web search + LLM |
Workflow (OpenClaw Agent)
Step 1: Acquire Data File
Try in order:
- Feishu chat file attachment — Download via API:
# Get token TOKEN=$(curl -s -X POST 'https://open.feishu.cn/open-apis/auth/v3/tenant_access_token/internal' \ -H 'Content-Type: application/json' \ -d '{"app_id":"APP_ID","app_secret":"APP_SECRET"}' | python3 -c "import json,sys; print(json.load(sys.stdin)['tenant_access_token'])") # Get file_key from message curl -s "https://open.feishu.cn/open-apis/im/v1/messages/{message_id}" -H "Authorization: Bearer $TOKEN" # Download curl -s "https://open.feishu.cn/open-apis/im/v1/messages/{message_id}/resources/{file_key}?type=file" \ -H "Authorization: Bearer $TOKEN" -o /tmp/data.xlsx - Feishu Doc/Bitable link — Use feishu_doc/feishu_bitable tools
- Local file — Use directly
Step 2: Extract Financial Data (JSON)
python3 scripts/generate_report.py /tmp/data.xlsx --company "公司名" --ticker "000001.SZ" --output-dir /tmp/reports --jsonThis outputs structured JSON with all financial metrics organized by category.
Step 3: Web Search for Company News & Competitors
web_search("{company} {year} 最新新闻 业绩 研发 重大事件")
web_search("{company} 竞争对手 市场份额 行业对比")Summarize 5-8 key news items as HTML list. Save to /tmp/reports/news.html.
Step 4: LLM Analysis
Based on the JSON data from Step 2 and search results from Step 3, write professional analysis for 7 sections. Save as JSON to /tmp/reports/analysis.json:
{
"profitability": "HTML text - revenue trends, margins, profitability inflection points...",
"balance_sheet": "HTML text - asset structure, leverage, liquidity...",
"cash_flow": "HTML text - operating CF trends, capex, FCF, cash reserves...",
"per_share": "HTML text - EPS/BPS trends, efficiency ratios, workforce...",
"industry": "HTML text wrapped in <div class=\"analysis-box\">...",
"competitor": "HTML text - competitor table + strategic analysis (see format below)",
"risk": "...",
"growth": "增长动力分析 HTML...",
"rd_analysis": "研发效率分析 HTML...",
"dupont": "杜邦分析(ROE拆解) HTML..."
}Analysis guidelines:
- Use
<strong>for key conclusions as the opening line - Use
<br><br>for paragraph breaks - Reference specific numbers from the data
- Compare year-over-year trends
- Highlight inflection points and turning points
- For risk section, use the two-column float layout with
.risk-list
Competitor section format:
<div class="analysis-box"><strong>Summary of competitive landscape</strong></div>
<table>
<tr><th style="text-align:left">竞争对手</th><th style="text-align:left">核心竞品</th><th style="text-align:left">竞争领域</th><th style="text-align:left">威胁程度</th></tr>
<tr><td style="text-align:left">Company A</td><td style="text-align:left">Product</td><td style="text-align:left">Domain</td><td style="text-align:left; color:#dc2626">直接竞争</td></tr>
</table>
<div class="analysis-box">Detailed competitive analysis...</div>Step 5: Generate Final Report
python3 scripts/generate_report.py /tmp/data.xlsx \
--company "公司名" --ticker "000001.SZ" \
--output-dir /tmp/reports \
--analysis-json /tmp/reports/analysis.json \
--news-html /tmp/reports/news.html \
-o html,pdfStep 6: Deliver Files via Feishu API
# Upload and send file
FK=$(curl -s -X POST 'https://open.feishu.cn/open-apis/im/v1/files' \
-H "Authorization: Bearer $TOKEN" \
-F 'file_type=stream' -F "file_name=report.html" -F "file=@/tmp/reports/report.html" \
| python3 -c "import json,sys; print(json.load(sys.stdin)['data']['file_key'])")
curl -s -X POST 'https://open.feishu.cn/open-apis/im/v1/messages?receive_id_type=chat_id' \
-H "Authorization: Bearer $TOKEN" -H 'Content-Type: application/json' \
-d "{\"receive_id\":\"CHAT_ID\",\"msg_type\":\"file\",\"content\":\"{\\\"file_key\\\":\\\"$FK\\\"}\"}"CLI Reference
# JSON data export (for LLM pipeline)
python3 scripts/generate_report.py input.xlsx --json --company NAME --ticker TICKER --output-dir DIR
# Full report with LLM analysis + competitor + news
python3 scripts/generate_report.py input.xlsx \
--analysis-json analysis.json \
--news-html news.html \
--company NAME --ticker TICKER \
-o html,pdf --output-dir DIR
# Standalone (rule-based analysis, no LLM needed)
python3 scripts/generate_report.py input.xlsx --company NAME --ticker TICKER -o html,pdfOutput Formats
| Flag | Output | Requires |
|---|---|---|
-o html | HTML only | (built-in) |
-o pdf | HTML + PDF | wkhtmltopdf |
-o doc | HTML + DOCX | pandoc |
-o md | HTML + Markdown | pandoc or markdownify |
Report Features
- LLM-powered analysis: Deep, contextual financial commentary (when using --analysis-json)
- Competitor analysis: Key competitors table + strategic competitive assessment
- Company news section: Recent hot topics and events (when using --news-html)
- Sparkline trend charts: SVG mini-charts in each data row
- Forecast markers: Predicted values with ⟡ symbol and yellow background
- PDF compatible: No emoji/flex/gradient/CSS variables (works with wkhtmltopdf)
- Fallback mode: Rule-based analysis when no LLM analysis is provided