Travel Expense Reimbursement Organizer

Problem

Travel reimbursement files often mix PDF invoices, hotel statements, ride receipts, and file names with dates, cities, seat classes, and amounts. This skill targets that business operations workflow: it turns a folder of travel expense documents into a cleaned reimbursement list.

How it works

• File scanning and normalization: reads the target folder and strips noisy numbering, underscores, and brackets from file names.
• Expense extraction: identifies dates, routes, amounts, and categories such as ride fares, flights, hotel stays, high-speed rail, and rental cars.
• PDF parsing: uses pymupdf to read ride receipts, hotel folios, and train tickets, extracting origin/destination, hotel name, check-in/check-out dates, stations, and seat classes.
• Flight discounting: requires an online lookup of the Y-cabin full fare for Air China, China Southern, or China Eastern on Ctrip, then subtracts airport construction and fuel surcharges before calculating the discount.
• Allowances and deduplication: applies 80 CNY/day for Guangzhou, Shenzhen, or Shanghai, and 60 CNY/day elsewhere; detects duplicate ride or rental entries and keeps the more complete record.
• Output: produces a compact single-line report by default, or a Markdown table with subtotals for manual review.

Boundaries

The skill depends on readable PDFs or sufficiently informative file names. Flight full fares must come from the required Ctrip lookup rather than estimates. If a hotel name lacks a city, a PDF is hard to parse, or rental item amounts do not reconcile exactly, manual correction may be needed.