04e1592fe3
- fill 命令生成 Excel 后正确移动到归档目录而非保留在报销输出 - 新增 build_formula_cached_values 支持填充公式单元格的缓存值 - 完成后清理报销输出目录避免残留临时文件 - .gitignore 移除 *.json 使 analyzed_result.json 可被追踪
582 lines
20 KiB
Python
582 lines
20 KiB
Python
#!/usr/bin/env python3
|
||
"""
|
||
差旅报销自动化脚本
|
||
|
||
Python 只做两类确定性工作：
|
||
1. extract：整理输入附件，提取 PDF 原文到本次归档目录的 TXT
|
||
2. fill：读取大模型生成的 JSON，填入 Excel 模板，并完成本次归档目录命名
|
||
|
||
中间的发票理解、金额分类、日期/事由/人员判断全部由大模型完成。
|
||
模板使用 .xlsx，完全依赖 Python 包实现跨平台，不依赖 LibreOffice。
|
||
"""
|
||
|
||
import argparse
|
||
import json
|
||
import re
|
||
import shutil
|
||
import zipfile
|
||
from datetime import date
|
||
from pathlib import Path
|
||
from tempfile import NamedTemporaryFile
|
||
from xml.etree import ElementTree as ET
|
||
|
||
import pdfplumber
|
||
|
||
|
||
OUTPUT_DIR_NAME = "报销输出"
|
||
RAW_TEXT_FILENAME = "pdf_raw_text.txt"
|
||
ANALYZED_RESULT_FILENAME = "analyzed_result.json"
|
||
TEMPLATE_FILE = Path("差旅费明细表.xlsx")
|
||
ATTACHMENT_SUFFIXES = {".pdf", ".jpg", ".jpeg", ".png"}
|
||
|
||
|
||
SCHEMA_EXAMPLE = {
|
||
"meta": {
|
||
"出差人": "从文本中提取或询问",
|
||
"出差事由": "从文本中提取或询问",
|
||
"出发日期": "从文本中提取或询问",
|
||
"结束日期": "从文本中提取或询问",
|
||
"出发地": "从文本中提取或询问",
|
||
"目的地": "从文本中提取或询问",
|
||
"出行方式": "从文本中提取或询问",
|
||
"出差天数": "从文本中提取或询问",
|
||
"事由简写": "从文本中提取或询问，适合文件名使用的简短事由描述"
|
||
},
|
||
"expenses": {
|
||
"住宿": {
|
||
"金额": 510.0,
|
||
"单据张数": 1,
|
||
"说明": "广州市番禺明乐宾馆住宿费"
|
||
},
|
||
"交通": {
|
||
"金额": 2013.77,
|
||
"单据张数": 5,
|
||
"说明": "机票1690.00 + 滴滴323.77"
|
||
},
|
||
"饮食": {
|
||
"金额": 0.0,
|
||
"单据张数": 0,
|
||
"说明": "未发现餐饮发票"
|
||
},
|
||
"其他": {
|
||
"金额": 0.0,
|
||
"单据张数": 0,
|
||
"说明": "无其他费用"
|
||
}
|
||
},
|
||
"summary": {
|
||
"附件总张数": 6,
|
||
"报销总额": 2523.77,
|
||
"预借差旅费": 0.0,
|
||
"补领金额": 2523.77,
|
||
"退还金额": 0.0
|
||
},
|
||
"evidence": [
|
||
{
|
||
"文件": "住宿.pdf",
|
||
"类别": "住宿",
|
||
"金额": 510.0,
|
||
"依据": "价税合计（小写）¥510.00"
|
||
},
|
||
{
|
||
"文件": "机票预订出行订单尾号87377629-电子普通发票.pdf",
|
||
"类别": "交通",
|
||
"金额": 1690.0,
|
||
"依据": "价税合计（小写）¥1690.00"
|
||
}
|
||
],
|
||
"questions": []
|
||
}
|
||
|
||
|
||
def command_extract(_args):
|
||
archive_dir = prepare_archive_dir()
|
||
output_dir = archive_dir / OUTPUT_DIR_NAME
|
||
output_dir.mkdir(exist_ok=True)
|
||
raw_text_file = output_dir / RAW_TEXT_FILENAME
|
||
|
||
pdf_files = sorted(archive_dir.glob("*.pdf"))
|
||
if not pdf_files:
|
||
raise SystemExit(f"{archive_dir} 中未找到 PDF 文件。请把 PDF 放到根目录后重新运行 extract。")
|
||
|
||
with open(raw_text_file, "w", encoding="utf-8") as output:
|
||
output.write("=" * 60 + "\n")
|
||
output.write("PDF 原始内容汇总\n")
|
||
output.write("=" * 60 + "\n\n")
|
||
|
||
for pdf_path in pdf_files:
|
||
output.write(f"\n{'=' * 60}\n")
|
||
output.write(f"文件: {pdf_path}\n")
|
||
output.write("=" * 60 + "\n")
|
||
|
||
try:
|
||
with pdfplumber.open(pdf_path) as pdf:
|
||
for page_index, page in enumerate(pdf.pages, start=1):
|
||
text = page.extract_text() or ""
|
||
output.write(f"\n--- 第 {page_index} 页 ---\n")
|
||
output.write(text)
|
||
output.write("\n")
|
||
except Exception as exc:
|
||
output.write(f"\n[读取失败: {exc}]\n")
|
||
|
||
print(f"本次附件目录：{archive_dir}")
|
||
print(f"已提取 {len(pdf_files)} 个 PDF 文件到 {raw_text_file}")
|
||
print(f"请将分析结果保存为：{output_dir / ANALYZED_RESULT_FILENAME}")
|
||
|
||
|
||
def prepare_archive_dir():
|
||
root_attachments = find_root_attachments()
|
||
today_dir = Path(date.today().strftime("%Y%m%d"))
|
||
|
||
if root_attachments:
|
||
archive_dir = choose_archive_dir(today_dir)
|
||
archive_dir.mkdir(exist_ok=True)
|
||
for attachment in root_attachments:
|
||
target = unique_file_path(archive_dir / attachment.name)
|
||
shutil.move(str(attachment), target)
|
||
print(f"已移动 {len(root_attachments)} 个输入文件到 {archive_dir}")
|
||
return archive_dir
|
||
|
||
existing_today_dir = find_existing_today_dir(today_dir)
|
||
if existing_today_dir:
|
||
return existing_today_dir
|
||
|
||
raise SystemExit("当前根目录未找到 PDF/JPG/JPEG/PNG 文件。请把本次发票、行程单或截图放到根目录后再运行 extract。")
|
||
|
||
|
||
def find_root_attachments():
|
||
return sorted(
|
||
path for path in Path(".").iterdir()
|
||
if path.is_file() and path.suffix.lower() in ATTACHMENT_SUFFIXES
|
||
)
|
||
|
||
|
||
def choose_archive_dir(today_dir):
|
||
if not today_dir.exists():
|
||
return today_dir
|
||
|
||
if has_attachments(today_dir):
|
||
answer = input(f"{today_dir} 已存在且包含附件，是否合并本次根目录附件？[y/N] ").strip().lower()
|
||
if answer in {"y", "yes"}:
|
||
return today_dir
|
||
return next_numbered_dir(today_dir)
|
||
|
||
return today_dir
|
||
|
||
|
||
def find_existing_today_dir(today_dir):
|
||
candidates = [today_dir] + sorted(Path(".").glob(f"{today_dir.name}-*"))
|
||
candidates = [path for path in candidates if path.is_dir() and list(path.glob("*.pdf"))]
|
||
if not candidates:
|
||
return None
|
||
return candidates[-1]
|
||
|
||
|
||
def has_attachments(directory):
|
||
return any(
|
||
path.is_file() and path.suffix.lower() in ATTACHMENT_SUFFIXES
|
||
for path in directory.iterdir()
|
||
)
|
||
|
||
|
||
def next_numbered_dir(base_dir):
|
||
index = 1
|
||
while True:
|
||
candidate = Path(f"{base_dir.name}-{index}")
|
||
if not candidate.exists():
|
||
return candidate
|
||
index += 1
|
||
|
||
|
||
def unique_file_path(path):
|
||
if not path.exists():
|
||
return path
|
||
|
||
index = 1
|
||
while True:
|
||
candidate = path.with_name(f"{path.stem}-{index}{path.suffix}")
|
||
if not candidate.exists():
|
||
return candidate
|
||
index += 1
|
||
|
||
|
||
def command_schema(args):
|
||
output_dir = Path(OUTPUT_DIR_NAME)
|
||
output_dir.mkdir(exist_ok=True)
|
||
text = json.dumps(SCHEMA_EXAMPLE, ensure_ascii=False, indent=2)
|
||
if args.output:
|
||
schema_path = output_dir / "schema_example.json"
|
||
schema_path.write_text(text + "\n", encoding="utf-8")
|
||
print(f"已写入 {schema_path}")
|
||
else:
|
||
print(text)
|
||
|
||
|
||
def command_fill(args):
|
||
result = load_json(args.json_path)
|
||
validate_result(result)
|
||
ensure_fill_dependencies()
|
||
|
||
json_path = Path(args.json_path)
|
||
output_dir = json_path.parent
|
||
archive_dir = output_dir.parent if output_dir.name == OUTPUT_DIR_NAME else Path(".")
|
||
|
||
output_file = fill_template(result, archive_dir)
|
||
verify_output(output_file, result)
|
||
final_archive_dir = finalize_archive_dir(archive_dir, result)
|
||
final_output_file = final_archive_dir / output_file.name
|
||
cleanup_output_dir(final_archive_dir / OUTPUT_DIR_NAME)
|
||
cleanup_pycache()
|
||
|
||
print(f"完成：{final_output_file}")
|
||
print(f"本次归档目录：{final_archive_dir}")
|
||
|
||
|
||
def load_json(json_path):
|
||
path = Path(json_path)
|
||
if not path.exists():
|
||
raise SystemExit(f"找不到 JSON 文件：{json_path}")
|
||
|
||
with open(path, encoding="utf-8") as f:
|
||
try:
|
||
return json.load(f)
|
||
except json.JSONDecodeError as exc:
|
||
raise SystemExit(f"JSON 格式错误：{exc}")
|
||
|
||
|
||
def validate_result(result):
|
||
require_keys(result, ["meta", "expenses", "summary"], "根对象")
|
||
|
||
meta = result["meta"]
|
||
expenses = result["expenses"]
|
||
summary = result["summary"]
|
||
|
||
require_keys(meta, [
|
||
"出差人", "出差事由", "出发日期", "结束日期", "出发地", "目的地",
|
||
"出行方式", "出差天数", "事由简写"
|
||
], "meta")
|
||
|
||
require_keys(expenses, ["住宿", "交通", "饮食", "其他"], "expenses")
|
||
for category in ["住宿", "交通", "饮食", "其他"]:
|
||
require_keys(expenses[category], ["金额", "单据张数", "说明"], f"expenses.{category}")
|
||
require_number(expenses[category]["金额"], f"expenses.{category}.金额")
|
||
require_integer(expenses[category]["单据张数"], f"expenses.{category}.单据张数")
|
||
|
||
require_keys(summary, ["附件总张数", "报销总额", "预借差旅费", "补领金额", "退还金额"], "summary")
|
||
|
||
expense_total = sum(float(expenses[c]["金额"]) for c in ["住宿", "交通", "饮食", "其他"])
|
||
if abs(float(summary["报销总额"]) - expense_total) > 0.01:
|
||
raise SystemExit(
|
||
f"报销总额不一致：summary.报销总额={summary['报销总额']}，"
|
||
f"费用合计={expense_total:.2f}"
|
||
)
|
||
|
||
receipt_total = sum(int(expenses[c]["单据张数"]) for c in ["住宿", "交通", "饮食", "其他"])
|
||
if int(summary["附件总张数"]) != receipt_total:
|
||
raise SystemExit(
|
||
f"附件总张数不一致：summary.附件总张数={summary['附件总张数']}，"
|
||
f"分类单据合计={receipt_total}"
|
||
)
|
||
|
||
|
||
def require_keys(obj, keys, label):
|
||
missing = [key for key in keys if key not in obj]
|
||
if missing:
|
||
raise SystemExit(f"{label} 缺少字段：{', '.join(missing)}")
|
||
|
||
|
||
def require_number(value, label):
|
||
if not isinstance(value, (int, float)):
|
||
raise SystemExit(f"{label} 必须是数字，当前为：{value!r}")
|
||
|
||
|
||
def require_integer(value, label):
|
||
if not isinstance(value, int):
|
||
raise SystemExit(f"{label} 必须是整数，当前为：{value!r}")
|
||
|
||
|
||
def ensure_fill_dependencies():
|
||
if not TEMPLATE_FILE.exists():
|
||
raise SystemExit(
|
||
"找不到模板文件：差旅费明细表.xlsx\n"
|
||
"请将 Excel 模板另存为 .xlsx 后放到当前目录。"
|
||
)
|
||
|
||
try:
|
||
import openpyxl # noqa: F401
|
||
except ImportError:
|
||
raise SystemExit("缺少 openpyxl，请运行：python -m pip install openpyxl")
|
||
|
||
|
||
def fill_template(result, output_dir):
|
||
output_dir.mkdir(exist_ok=True)
|
||
|
||
meta = result["meta"]
|
||
output_file = output_dir / build_output_filename(meta)
|
||
|
||
shutil.copy(TEMPLATE_FILE, output_file)
|
||
update_xlsx_cells(output_file, build_cell_values(result), build_formula_cached_values(result))
|
||
return output_file
|
||
|
||
|
||
def build_cell_values(result):
|
||
meta = result["meta"]
|
||
expenses = result["expenses"]
|
||
return {
|
||
"E3": meta["出差事由"],
|
||
"A6": f'{meta["出发日期"]}-{meta["结束日期"]}',
|
||
"B6": f'{meta["出发地"]}-{meta["目的地"]}',
|
||
"C6": f'{meta["出差天数"]}天',
|
||
"D6": meta["出行方式"],
|
||
"F6": meta["出差天数"],
|
||
"H6": expenses["住宿"]["单据张数"],
|
||
"I6": expenses["住宿"]["金额"],
|
||
"H7": expenses["饮食"]["单据张数"],
|
||
"I7": expenses["饮食"]["金额"],
|
||
"H8": expenses["交通"]["单据张数"],
|
||
"I8": expenses["交通"]["金额"],
|
||
"H9": expenses["其他"]["单据张数"],
|
||
"I9": expenses["其他"]["金额"],
|
||
}
|
||
|
||
|
||
def build_formula_cached_values(result):
|
||
summary = result["summary"]
|
||
advance = summary["预借差旅费"]
|
||
total = summary["报销总额"]
|
||
return {
|
||
"H2": summary["附件总张数"],
|
||
"F9": result["meta"]["出差天数"],
|
||
"C10": total,
|
||
"H10": max(total - advance, 0),
|
||
"H11": max(advance - total, 0),
|
||
}
|
||
|
||
|
||
def update_xlsx_cells(xlsx_path, cell_values, formula_cached_values):
|
||
temp_file = NamedTemporaryFile(delete=False, suffix=".xlsx", dir=xlsx_path.parent)
|
||
temp_path = Path(temp_file.name)
|
||
temp_file.close()
|
||
|
||
try:
|
||
with zipfile.ZipFile(xlsx_path, "r") as source:
|
||
with zipfile.ZipFile(temp_path, "w", compression=zipfile.ZIP_DEFLATED) as target:
|
||
for item in source.infolist():
|
||
data = source.read(item.filename)
|
||
if item.filename == "xl/worksheets/sheet1.xml":
|
||
data = update_sheet_xml(data, cell_values, formula_cached_values)
|
||
target.writestr(item, data)
|
||
temp_path.replace(xlsx_path)
|
||
except Exception:
|
||
temp_path.unlink(missing_ok=True)
|
||
raise
|
||
|
||
|
||
def update_sheet_xml(sheet_xml, cell_values, formula_cached_values):
|
||
namespace = "http://schemas.openxmlformats.org/spreadsheetml/2006/main"
|
||
ET.register_namespace("", namespace)
|
||
root = ET.fromstring(sheet_xml)
|
||
for cell_ref, value in cell_values.items():
|
||
cell = root.find(f".//{{{namespace}}}c[@r='{cell_ref}']")
|
||
if cell is None:
|
||
raise SystemExit(f"模板缺少单元格：{cell_ref}")
|
||
set_cell_xml_value(cell, value, namespace)
|
||
for cell_ref, value in formula_cached_values.items():
|
||
cell = root.find(f".//{{{namespace}}}c[@r='{cell_ref}']")
|
||
if cell is None:
|
||
raise SystemExit(f"模板缺少单元格：{cell_ref}")
|
||
set_formula_cached_value(cell, value, namespace)
|
||
return ET.tostring(root, encoding="utf-8", xml_declaration=True)
|
||
|
||
|
||
def set_cell_xml_value(cell, value, namespace):
|
||
for child in list(cell):
|
||
if child.tag in {f"{{{namespace}}}v", f"{{{namespace}}}f", f"{{{namespace}}}is"}:
|
||
cell.remove(child)
|
||
|
||
if isinstance(value, (int, float)):
|
||
cell.attrib.pop("t", None)
|
||
value_node = ET.SubElement(cell, f"{{{namespace}}}v")
|
||
value_node.text = format_number_for_excel(value)
|
||
else:
|
||
cell.set("t", "inlineStr")
|
||
inline_string = ET.SubElement(cell, f"{{{namespace}}}is")
|
||
text_node = ET.SubElement(inline_string, f"{{{namespace}}}t")
|
||
text_node.text = str(value)
|
||
|
||
|
||
def set_formula_cached_value(cell, value, namespace):
|
||
cell.attrib.pop("t", None)
|
||
value_node = None
|
||
for child in cell:
|
||
if child.tag == f"{{{namespace}}}v":
|
||
value_node = child
|
||
break
|
||
if value_node is None:
|
||
value_node = ET.SubElement(cell, f"{{{namespace}}}v")
|
||
value_node.text = format_number_for_excel(value)
|
||
|
||
|
||
def format_number_for_excel(value):
|
||
if isinstance(value, int):
|
||
return str(value)
|
||
return f"{float(value):.10f}".rstrip("0").rstrip(".")
|
||
|
||
|
||
def finalize_archive_dir(archive_dir, result):
|
||
if archive_dir == Path("."):
|
||
return archive_dir
|
||
|
||
target_dir = unique_archive_dir(archive_dir.parent / build_archive_dir_name(result["meta"]))
|
||
if archive_dir.resolve() == target_dir.resolve():
|
||
return archive_dir
|
||
|
||
archive_dir.rename(target_dir)
|
||
return target_dir
|
||
|
||
|
||
def unique_archive_dir(path):
|
||
if not path.exists():
|
||
return path
|
||
|
||
index = 1
|
||
while True:
|
||
candidate = path.with_name(f"{path.name}-{index}")
|
||
if not candidate.exists():
|
||
return candidate
|
||
index += 1
|
||
|
||
|
||
def cleanup_output_dir(output_dir):
|
||
if output_dir.exists():
|
||
shutil.rmtree(output_dir)
|
||
|
||
|
||
def cleanup_pycache():
|
||
pycache_dir = Path("__pycache__")
|
||
if pycache_dir.exists():
|
||
shutil.rmtree(pycache_dir)
|
||
|
||
|
||
def build_archive_dir_name(meta):
|
||
start_date = normalize_date_for_filename(meta["出发日期"])
|
||
end_date = normalize_date_for_filename(meta["结束日期"])
|
||
route = sanitize_filename(f'{meta["出发地"]}-{meta["目的地"]}')
|
||
reason = sanitize_filename(meta["事由简写"])
|
||
return f"{start_date}-{end_date}_{route}_{reason}"
|
||
|
||
|
||
def build_output_filename(meta):
|
||
date_str = normalize_date_for_filename(meta["出发日期"])
|
||
safe_parts = [
|
||
"差旅费明细表",
|
||
sanitize_filename(meta["出差人"]),
|
||
date_str,
|
||
sanitize_filename(meta["目的地"]),
|
||
sanitize_filename(meta["事由简写"]),
|
||
]
|
||
return "_".join(safe_parts) + ".xlsx"
|
||
|
||
|
||
def normalize_date_for_filename(date_text):
|
||
parts = re.split(r"[/-]", str(date_text))
|
||
if len(parts) != 3:
|
||
return sanitize_filename(str(date_text))
|
||
return f"{int(parts[0])}{int(parts[1]):02d}{int(parts[2]):02d}"
|
||
|
||
|
||
def sanitize_filename(text):
|
||
text = str(text).strip()
|
||
text = re.sub(r"[\\/:*?\"<>|\s]+", "", text)
|
||
return text or "未填写"
|
||
|
||
|
||
def verify_output(output_file, result):
|
||
import openpyxl
|
||
|
||
wb = openpyxl.load_workbook(output_file, data_only=False)
|
||
ws = wb.active
|
||
meta = result["meta"]
|
||
|
||
expected_route = f'{meta["出发地"]}-{meta["目的地"]}'
|
||
expected_date = f'{meta["出发日期"]}-{meta["结束日期"]}'
|
||
|
||
checks = [
|
||
(ws["E3"].value, meta["出差事由"], "出差事由"),
|
||
(ws["A6"].value, expected_date, "日期"),
|
||
(ws["B6"].value, expected_route, "起止地点"),
|
||
]
|
||
for actual, expected, label in checks:
|
||
if actual != expected:
|
||
raise SystemExit(f"输出校验失败：{label} 实际={actual!r} 预期={expected!r}")
|
||
|
||
formula_checks = {
|
||
"H2": "附件张数合计",
|
||
"F9": "天数合计",
|
||
"C10": "报销总额",
|
||
"H10": "补领金额",
|
||
}
|
||
for cell, label in formula_checks.items():
|
||
value = ws[cell].value
|
||
if not isinstance(value, str) or not value.startswith("="):
|
||
raise SystemExit(f"输出校验失败：{label} 单元格 {cell} 公式丢失，当前值={value!r}")
|
||
|
||
verify_xlsx_styles_preserved(output_file)
|
||
|
||
|
||
def verify_xlsx_styles_preserved(output_file):
|
||
with zipfile.ZipFile(TEMPLATE_FILE, "r") as template, zipfile.ZipFile(output_file, "r") as output:
|
||
if template.read("xl/styles.xml") != output.read("xl/styles.xml"):
|
||
raise SystemExit("输出校验失败：样式表 xl/styles.xml 与模板不一致")
|
||
|
||
template_styles = read_sheet_cell_style_ids(template.read("xl/worksheets/sheet1.xml"))
|
||
output_styles = read_sheet_cell_style_ids(output.read("xl/worksheets/sheet1.xml"))
|
||
for cell_ref in filled_cell_refs():
|
||
if template_styles.get(cell_ref) != output_styles.get(cell_ref):
|
||
raise SystemExit(
|
||
f"输出校验失败：单元格 {cell_ref} 样式编号变化，"
|
||
f"模板={template_styles.get(cell_ref)!r} 输出={output_styles.get(cell_ref)!r}"
|
||
)
|
||
|
||
|
||
def read_sheet_cell_style_ids(sheet_xml):
|
||
namespace = "http://schemas.openxmlformats.org/spreadsheetml/2006/main"
|
||
root = ET.fromstring(sheet_xml)
|
||
return {
|
||
cell.attrib["r"]: cell.attrib.get("s")
|
||
for cell in root.findall(f".//{{{namespace}}}c")
|
||
if "r" in cell.attrib
|
||
}
|
||
|
||
|
||
def filled_cell_refs():
|
||
return ["E3", "A6", "B6", "C6", "D6", "H6", "I6", "H7", "I7", "H8", "I8", "H9", "I9"]
|
||
|
||
|
||
def build_parser():
|
||
parser = argparse.ArgumentParser(description="差旅报销自动化脚本")
|
||
subparsers = parser.add_subparsers(dest="command", required=True)
|
||
|
||
extract_parser = subparsers.add_parser("extract", help="整理根目录附件并提取 PDF 原文到本次归档目录")
|
||
extract_parser.set_defaults(func=command_extract)
|
||
|
||
schema_parser = subparsers.add_parser("schema", help="输出大模型需要生成的 JSON schema 示例")
|
||
schema_parser.add_argument("--output", action="store_true", help="写入 报销输出/schema_example.json，而不是打印到终端")
|
||
schema_parser.set_defaults(func=command_schema)
|
||
|
||
fill_parser = subparsers.add_parser("fill", help="读取 JSON 并填入差旅费明细表模板，再归档本次目录")
|
||
fill_parser.add_argument("json_path", help="大模型生成的 analyzed_result.json 路径")
|
||
fill_parser.set_defaults(func=command_fill)
|
||
|
||
return parser
|
||
|
||
|
||
def main():
|
||
parser = build_parser()
|
||
args = parser.parse_args()
|
||
args.func(args)
|
||
|
||
|
||
if __name__ == "__main__":
|
||
main()
|