gavindiaz/mt5_report_parser
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

添加批量解析

Browse Source
This commit is contained in:
gavindiaz
2026-07-11 03:22:50 +08:00
parent 9cdf93cc3d
commit 491ea17137
985 changed files with 4285 additions and 103 deletions
Download Patch File Download Diff File Expand all files Collapse all files
mt5_report_parser.py
+372 -7
View File
@@ -14,19 +14,108 @@ MT5 策略测试报告解析器（中文版 xlsx）
from __future__ import annotations
import re
import os as _os_std
from dataclasses import dataclass, field
from typing import Any, Dict, List, Optional, Tuple
from html.parser import HTMLParser
import numpy as np
import openpyxl
import pandas as pd
# 报告中区段标题（A 列出现的精确字符串）
# 报告中区段标题
SEC_SETTINGS = "设置"
SEC_RESULTS = "结果"
SEC_ORDERS = "订单"
SEC_DEALS = "成交"
# 支持的报告格式
_Supported_Extensions = {".xlsx", ".htm", ".html"}
class _MT5HTMLParser(HTMLParser):
"""MT5 HTML 报告解析器。"""
def __init__(self):
super().__init__()
self.in_td = False
self.in_th = False
self.in_b = False
self.current_row: List[Optional[str]] = []
self.current_cell: str = ""
self.b_content: str = "" # Content inside <b> tags
self.rows: List[List[Optional[str]]] = []
self.section_rows: Dict[str, int] = {} # Track section headers
def handle_starttag(self, tag, attrs):
if tag == "tr":
if self.current_row:
self.rows.append(self.current_row)
self.current_row = []
elif tag == "td":
self.current_cell = ""
self.b_content = ""
self.in_td = True
self.in_b = False
elif tag == "th":
self.current_cell = ""
self.b_content = ""
self.in_th = True
self.in_b = False
elif tag == "b":
self.in_b = True
self.b_content = ""
def handle_endtag(self, tag):
if tag == "tr":
if self.current_row:
self.rows.append(self.current_row)
self.current_row = []
self.current_cell = ""
self.in_td = False
self.in_th = False
elif tag == "td":
# Build full cell: non-b text + <b> content
full_cell = self.current_cell + self.b_content
self.current_row.append(full_cell.strip() if full_cell.strip() else "")
self.current_cell = ""
self.b_content = ""
self.in_td = False
elif tag == "th":
full_cell = self.current_cell + self.b_content
# Check if this is a section header
b_val = self.b_content.strip()
if b_val in (SEC_SETTINGS, SEC_RESULTS, SEC_ORDERS, SEC_DEALS):
row_idx = len(self.rows)
self.section_rows[b_val] = row_idx
self.current_row.append(full_cell.strip() if full_cell.strip() else "")
self.current_cell = ""
self.b_content = ""
self.in_th = False
elif tag == "b":
self.in_b = False
def handle_data(self, data):
if self.in_td or self.in_th:
if not self.in_b:
self.current_cell += data
else:
self.b_content += data
def finish(self):
if self.current_row:
self.rows.append(self.current_row)
# Also search all cells for section markers (they may be in <td> not just <th>)
for i, row in enumerate(self.rows):
if row:
for j, cell in enumerate(row):
if cell:
b_val = _get_b_text(cell)
if b_val and b_val.strip() in (SEC_SETTINGS, SEC_RESULTS, SEC_ORDERS, SEC_DEALS):
sec = b_val.strip()
if sec not in self.section_rows:
self.section_rows[sec] = i
return self.rows, self.section_rows
@dataclass
class MT5Report:
@@ -121,10 +210,253 @@ def _pct_to_float(x: Any) -> Optional[float]:
# --------------------------------------------------------------------------- #
# 主入口
# HTML 报告解析
# --------------------------------------------------------------------------- #
def parse_report(path: str) -> MT5Report:
"""解析一份 MT5 xlsx 报告。"""
def _parse_html_report(text: str) -> MT5Report:
"""解析 MT5 HTML 报告。"""
parser = _MT5HTMLParser()
parser.feed(text)
rows, section_rows = parser.finish()
r_settings = section_rows.get(SEC_SETTINGS, 0)
r_results = section_rows.get(SEC_RESULTS, r_settings)
r_orders = section_rows.get(SEC_ORDERS, r_results)
r_deals = section_rows.get(SEC_DEALS, r_orders)
rep = MT5Report()
rep.meta, rep.summary = _parse_html_meta_and_summary(rows, r_settings, r_results, r_orders)
rep.orders, rep.deals = _parse_html_tables(rows, r_orders, r_deals, len(rows))
rep.trades = _reconstruct_trades(rep.deals)
_normalize_summary_numbers(rep)
return rep
def _read_html_file(path: str) -> str:
"""读取 HTML 文件，自动检测编码。"""
with open(path, "rb") as f:
raw = f.read()
# 检测 BOM
if raw[:2] == b"\xff\xfe":
return raw.decode("utf-16-le")
elif raw[:2] == b"\xfe\xff":
return raw.decode("utf-16-be")
elif raw[:3] == b"\xef\xbb\xbf":
return raw[3:].decode("utf-8")
elif raw[:4] == b"\xff\xfe\x00\x00":
return raw[4:].decode("utf-32-le")
# 尝试 UTF-8
try:
return raw.decode("utf-8")
except UnicodeDecodeError:
pass
# 回退到 latin-1
return raw.decode("latin-1")
def _get_b_text(cell_text: str) -> Optional[str]:
"""从 HTML 单元格文本中提取 <b>...</b> 内的文本。"""
if cell_text is None or not cell_text.strip():
return None
# 直接提取 <b>...</b> 内容
m = re.search(r'<b>(.*?)</b>', cell_text, re.DOTALL)
if m:
text = m.group(1).strip()
# 去除内嵌的 <br> 标签
text = re.sub(r'<br\s*/?>', '', text, flags=re.IGNORECASE)
return text if text else None
# 如果没有 <b> 标签，返回去除 HTML 标签的纯文本
clean = re.sub(r'<[^>]+>', '', cell_text).strip()
return clean if clean else None
def _parse_html_meta_and_summary(
rows: List[List[Optional[str]]],
r_settings: int, r_results: int, r_orders: int,
) -> Tuple[Dict[str, Any], Dict[str, Any]]:
"""解析 HTML 报告中的设置和结果部分。"""
meta: Dict[str, Any] = {}
summary: Dict[str, Any] = {}
# 解析设置段
for i in range(r_settings, r_results):
if i >= len(rows):
break
row = rows[i]
if not row or len(row) < 2:
continue
label = _get_b_text(row[0] or "")
if label and label.endswith(":"):
label = label[:-1].strip()
# 值在第二个非空单元格
for j in range(1, len(row)):
val = _get_b_text(row[j] or "")
if val is not None:
if label == "输入":
# 输入参数
continue
meta[label] = val
break
# 解析输入参数（在设置段中查找 Key=Value 格式）
inputs: Dict[str, str] = {}
for i in range(r_settings, r_results):
if i >= len(rows):
break
for j in range(len(rows[i])):
cell = rows[i][j]
if cell:
b_val = _get_b_text(cell)
if b_val and "=" in b_val and not b_val.strip().startswith("==="):
k, _, val = b_val.partition("=")
inputs[k.strip()] = val.strip()
meta["inputs"] = inputs
# 解析结果段
for i in range(r_results, r_orders):
if i >= len(rows):
break
row = rows[i]
if not row or len(row) < 2:
continue
label = _get_b_text(row[0] or "")
if label and label.endswith(":"):
label = label[:-1].strip()
for j in range(1, len(row)):
val = _get_b_text(row[j] or "")
if val is not None:
summary[label] = val
break
return meta, summary
def _parse_html_tables(
rows: List[List[Optional[str]]],
r_orders: int, r_deals: int, max_row: int,
) -> Tuple[Optional[pd.DataFrame], Optional[pd.DataFrame]]:
"""解析 HTML 报告中的订单表和成交表。"""
orders_df: Optional[pd.DataFrame] = None
deals_df: Optional[pd.DataFrame] = None
# HTML 报告列名到标准英文列名的映射
DEALS_COL_MAP = {
"时间": "time",
"成交": "deal_id",
"交易品种": "symbol",
"类型": "type",
"趋势": "entry",
"交易量": "volume",
"价位": "price",
"订单": "order",
"手续费": "commission",
"库存费": "swap",
"盈利": "profit",
"结余": "balance",
"注释": "comment",
}
ORDERS_COL_MAP = {
"时间": "time",
"交易": "ticket",
"交易品种": "symbol",
"类型": "type",
"状态": "state",
"交易量": "volume",
"价位": "price",
"止损": "sl",
"止盈": "tp",
"注释": "comment",
"订单": "order",
}
def _parse_table_section(start_row: int, end_row: int, col_map: Dict[str, str]) -> Optional[pd.DataFrame]:
"""Parse a table section from rows."""
if start_row >= max_row:
return None
# Find header row
header_row = None
for i in range(start_row + 1, min(start_row + 5, end_row)):
if i >= len(rows):
break
row = rows[i]
if row:
b_count = sum(1 for cell in row if cell and _get_b_text(cell))
if b_count >= 3:
header_row = i
break
if header_row is None:
return None
# Extract headers
headers = [_get_b_text(cell) for cell in rows[header_row] if cell and _get_b_text(cell)]
if not headers:
return None
# Rename columns using col_map
renamed_headers = [col_map.get(h, h) for h in headers]
# Extract data rows
data_rows = []
for i in range(header_row + 1, end_row):
if i >= len(rows):
break
row = rows[i]
if row:
values = [_get_b_text(cell) or "" for cell in row]
if any(v for v in values):
while len(values) < len(renamed_headers):
values.append("")
data_rows.append(values[:len(renamed_headers)])
if not data_rows:
return None
df = pd.DataFrame(data_rows, columns=renamed_headers)
return df
if r_orders < max_row:
orders_df = _parse_table_section(r_orders, r_deals, ORDERS_COL_MAP)
if r_deals < max_row:
deals_df = _parse_table_section(r_deals, max_row, DEALS_COL_MAP)
return orders_df, deals_df
# --------------------------------------------------------------------------- #
# 解析缓存（文件级 LRU 缓存）
# --------------------------------------------------------------------------- #
_parse_cache: Dict[str, MT5Report] = {}
def parse_report(path: str, use_cache: bool = True) -> MT5Report:
"""解析一份 MT5 报告（支持 xlsx / htm / html 格式）。"""
real_path = _os_std.path.realpath(str(path))
if use_cache and real_path in _parse_cache:
return _parse_cache[real_path]
ext = _os_std.path.splitext(real_path)[1].lower()
if ext == ".xlsx":
rep = _parse_xlsx_report(real_path)
elif ext in (".htm", ".html"):
text = _read_html_file(real_path)
rep = _parse_html_report(text)
rep.source_file = real_path
else:
raise ValueError(f"不支持的报告格式: {ext}（仅支持 .xlsx / .htm / .html）")
_parse_cache[real_path] = rep
return rep
def _parse_xlsx_report(path: str) -> MT5Report:
"""解析 xlsx 格式报告（原 parse_report 逻辑）。"""
wb = openpyxl.load_workbook(path, data_only=True)
ws = wb[wb.sheetnames[0]]
rep = MT5Report(source_file=path)
@@ -136,16 +468,15 @@ def parse_report(path: str) -> MT5Report:
r_deals = sec.get(SEC_DEALS, r_orders)
# ---- 元信息 / 设置 ----
# 整个设置块（设置标题行之后到结果段之前）逐行解析 “标签: 值” 对
meta: Dict[str, Any] = {}
for r in range(r_settings + 1, r_results):
kv = _kv_in_row(ws, r)
for k, v in kv.items():
if k == "输入": # “输入:” 是子段标题，其值为分节字符串，跳过
if k == "输入": # "输入:" 是子段标题，其值为分节字符串，跳过
continue
meta[k] = v
# 输入参数：在 “输入:” 段落里，列 D 形如 Key=Value
# 输入参数：在 "输入:" 段落里，列 D 形如 Key=Value
inputs: Dict[str, str] = {}
for r in range(r_settings + 1, r_results):
for c in (4,): # 经验上输入参数在 D 列
@@ -173,9 +504,15 @@ def parse_report(path: str) -> MT5Report:
rep.summary = summary
rep.trades = _reconstruct_trades(rep.deals)
_normalize_summary_numbers(rep)
return rep
def clear_parse_cache() -> None:
"""清除解析缓存（调试 / 释放内存用）。"""
_parse_cache.clear()
# --------------------------------------------------------------------------- #
# 逐笔交易重建
# --------------------------------------------------------------------------- #
@@ -256,6 +593,34 @@ def _reconstruct_trades(deals: pd.DataFrame) -> pd.DataFrame:
return pd.DataFrame(trades)
# --------------------------------------------------------------------------- #
# 单段通用指标（供 walk_forward / run_analysis 复用）
# --------------------------------------------------------------------------- #
def compute_segment_metrics(net: pd.Series) -> Dict[str, float]:
"""
由净盈亏序列计算核心指标。
无状态、纯函数，可直接用于单段或单份报告。
"""
n = len(net)
if n == 0:
return {"n": 0, "net": 0.0, "pf": 0.0, "win": 0.0, "dd": 0.0, "exp": 0.0}
wins = net[net > 0]
losses = net[net <= 0]
gp = wins.sum()
gl = -losses.sum()
pf = gp / gl if gl > 0 else np.inf
equity = net.cumsum()
dd = float((equity - equity.cummax()).min())
return {
"n": int(n),
"net": float(net.sum()),
"pf": float(pf),
"win": float(len(wins) / n * 100),
"dd": dd,
"exp": float(net.mean()),
}
# --------------------------------------------------------------------------- #
# 汇总指标数值化（便于程序对比）
# --------------------------------------------------------------------------- #