xhong
f608e3afab
- 将单一城市名称参数改为省、市、县三级参数，提升解析准确性 - 实现层级查找策略：全国->省->市->县，并添加参数校验 - 更新输出文件命名规则，使用"省_市"或"省_市_县"格式 - 完善函数文档字符串和错误提示信息
297 lines
9.5 KiB
Python
297 lines
9.5 KiB
Python
"""
|
|
访问阿里云 DataV 下载的行政区边界数据保存为原始 JSON, 有部分字段与GDAL不兼容, 导致无法
|
|
直接读取, 需要先清洗再保存为 GeoJSON.
|
|
|
|
- 官方地址: https://datav.aliyun.com/portal/school/atlas/area_selector
|
|
- Step1: 按"省-市-县名称"解析为行政区划代码 (adcode);
|
|
- Step2: 将 DataV 原始数据保存为 `省-市-县名称.json` 文件;
|
|
- Step3: 移除不兼容 GDAL/GeoPandas 的属性字段 (parent, center, centroid, acroutes);
|
|
- Step4: 将清洗后的结果写出为 `省-市-县名称.geojson` 文件.
|
|
|
|
-------------------------------------------------------------------------------
|
|
Authors: CVEO Team
|
|
Last Updated: 2026-04-13
|
|
===============================================================================
|
|
"""
|
|
|
|
import json
|
|
import re
|
|
from pathlib import Path
|
|
from typing import Optional
|
|
|
|
import requests
|
|
|
|
|
|
def _validate_region_params(province: str, city: str, county: Optional[str]) -> None:
|
|
"""
|
|
Validate that province and city are non-empty.
|
|
|
|
Raises
|
|
------
|
|
ValueError
|
|
If province or city is None or empty string.
|
|
"""
|
|
if not province or not province.strip():
|
|
raise ValueError("province (省) cannot be empty")
|
|
if not city or not city.strip():
|
|
raise ValueError("city (市) cannot be empty")
|
|
|
|
|
|
def get_datav_json(accode: str) -> dict:
|
|
"""
|
|
从 DataV 接口获取行政区边界的原始 JSON 数据并返回字典.
|
|
"""
|
|
# 使用路径式接口, 支持如 "420100" 或 "420100_full"
|
|
url = f"https://geo.datav.aliyun.com/areas_v3/bound/{accode}.json"
|
|
response = requests.get(url, timeout=15)
|
|
response.raise_for_status()
|
|
return response.json()
|
|
|
|
|
|
def fetch_and_save_geojson(accode: str, region_name: str, out_dir: Path) -> Path:
|
|
"""
|
|
获取 DataV 原始数据, 先保存为 .json; 随后清洗属性并另存为 .geojson.
|
|
|
|
Parameters
|
|
----------
|
|
accode : str
|
|
行政区划代码, 如 "420100" 或 "420100_full".
|
|
region_name : str
|
|
区域名称, 用于输出文件名. 当使用省-市-县三级参数时, 文件名为
|
|
"{province}_{city}.geojson" 或 "{province}_{city}_{county}.geojson".
|
|
out_dir : Path
|
|
输出目录.
|
|
"""
|
|
raw_data = get_datav_json(accode)
|
|
|
|
# 处理 features: 移除不兼容 GeoPandas 的属性
|
|
def sanitize_properties(props: dict) -> dict:
|
|
out = {}
|
|
for k, v in props.items():
|
|
# 移除嵌套对象
|
|
if k in ("parent", "center", "centroid", "acroutes"):
|
|
continue
|
|
# 仅保留标量类型; 丢弃其他嵌套结构
|
|
if isinstance(v, (str, int, float, bool)) or v is None:
|
|
out[k] = v
|
|
return out
|
|
|
|
# 输出路径(确保目录存在)
|
|
out_dir_path = Path(out_dir)
|
|
out_dir_path.mkdir(parents=True, exist_ok=True)
|
|
# 先保存原始 JSON(未清洗)
|
|
raw_json_path = out_dir_path / f"{region_name}.json"
|
|
with raw_json_path.open("w", encoding="utf-8") as f:
|
|
json.dump(raw_data, f, ensure_ascii=False)
|
|
# 再保存清洗后的 GeoJSON
|
|
out_path = out_dir_path / f"{region_name}.geojson"
|
|
# 深拷贝后进行清洗, 避免影响原始数据
|
|
data = json.loads(json.dumps(raw_data))
|
|
features = data.get("features", [])
|
|
for feature in features:
|
|
props = feature.get("properties", {})
|
|
feature["properties"] = sanitize_properties(props)
|
|
|
|
# 写出为 .geojson, 确保 UTF-8 且保留中文字符
|
|
with out_path.open("w", encoding="utf-8") as f:
|
|
json.dump(data, f, ensure_ascii=False)
|
|
|
|
return out_path
|
|
|
|
|
|
def _normalize_name(name: str) -> str:
|
|
name = name.strip()
|
|
# 简单去除常见后缀提高匹配鲁棒性
|
|
for suffix in ("市", "省", "地区", "盟", "自治州", "自治县", "特别行政区"):
|
|
if name.endswith(suffix):
|
|
name = name[: -len(suffix)]
|
|
return name
|
|
|
|
|
|
def _name_matches_exact(target: str, candidate: str) -> bool:
|
|
return _normalize_name(target) == _normalize_name(candidate)
|
|
|
|
|
|
def resolve_adcode_by_name(
|
|
province: str,
|
|
city: str,
|
|
county: Optional[str] = None,
|
|
prefer_full: bool = False,
|
|
) -> Optional[str]:
|
|
"""
|
|
通过省-市-县名称解析 DataV 行政区划代码.
|
|
|
|
采用层级查找策略: 全国数据 -> 省级数据 -> 市级数据 -> 区县数据.
|
|
前两级 (省/市) 必须提供且不能为空; 第三级 (县) 可为空.
|
|
|
|
Parameters
|
|
----------
|
|
province : str
|
|
省份名称, 如 "湖北省". 不能为空.
|
|
city : str
|
|
城市名称, 如 "武汉市". 不能为空.
|
|
county : str, optional
|
|
区县名称, 如 "江岸区". 可为空 (表示只解析到市级).
|
|
prefer_full : bool, optional
|
|
是否返回下辖完整边界代码 (如 "420100_full"), 默认 False.
|
|
|
|
Returns
|
|
-------
|
|
str or None
|
|
行政区划代码, 如 "420100" 或 "420100_full", 找不到则返回 None.
|
|
|
|
Raises
|
|
------
|
|
ValueError
|
|
如果 province 或 city 为空.
|
|
"""
|
|
_validate_region_params(province, city, county)
|
|
|
|
# Step 1: 从全国数据中查找省份
|
|
try:
|
|
cn = requests.get(
|
|
"https://geo.datav.aliyun.com/areas_v3/bound/100000_full.json",
|
|
timeout=20,
|
|
).json()
|
|
except Exception:
|
|
return None
|
|
|
|
pcode = None
|
|
for feat in cn.get("features", []):
|
|
props = feat.get("properties", {})
|
|
if props.get("level") == "province":
|
|
name = props.get("name", "")
|
|
code = str(props.get("adcode", ""))
|
|
if re.fullmatch(r"\d{6}", code):
|
|
if _name_matches_exact(province, name):
|
|
pcode = code
|
|
break
|
|
|
|
if not pcode:
|
|
return None
|
|
|
|
# Step 2: 从省份数据中查找城市
|
|
# 特殊处理直辖市: 当 province == city 时 (如 "北京市" == "北京市")
|
|
if _name_matches_exact(province, city):
|
|
ccode = pcode
|
|
else:
|
|
try:
|
|
prov_data = requests.get(
|
|
f"https://geo.datav.aliyun.com/areas_v3/bound/{pcode}_full.json",
|
|
timeout=20,
|
|
).json()
|
|
except Exception:
|
|
return None
|
|
|
|
ccode = None
|
|
for feat in prov_data.get("features", []):
|
|
props = feat.get("properties", {})
|
|
level = props.get("level")
|
|
name = props.get("name", "")
|
|
code = str(props.get("adcode", ""))
|
|
|
|
# 匹配城市或区县级别
|
|
if level in ("city", "district") and re.fullmatch(r"\d{6}", code):
|
|
if _name_matches_exact(city, name):
|
|
ccode = code
|
|
break
|
|
|
|
if not ccode:
|
|
return None
|
|
|
|
# Step 3: 如果提供了区县名称, 从城市数据中查找区县
|
|
if county:
|
|
try:
|
|
city_data = requests.get(
|
|
f"https://geo.datav.aliyun.com/areas_v3/bound/{ccode}_full.json",
|
|
timeout=20,
|
|
).json()
|
|
except Exception:
|
|
return None
|
|
|
|
dcode = None
|
|
for feat in city_data.get("features", []):
|
|
props = feat.get("properties", {})
|
|
level = props.get("level")
|
|
name = props.get("name", "")
|
|
code = str(props.get("adcode", ""))
|
|
|
|
# 匹配区县 (包括县级市, 其 level 可能为 "city")
|
|
if level in ("district", "city") and re.fullmatch(r"\d{6}", code):
|
|
if _name_matches_exact(county, name):
|
|
dcode = code
|
|
break
|
|
|
|
if not dcode:
|
|
return None
|
|
# 区县级别已经是最终边界，不需要 _full 后缀
|
|
return dcode
|
|
|
|
# 只解析到市级
|
|
return f"{ccode}_full" if prefer_full else ccode
|
|
|
|
|
|
def fetch_and_save_geojson_by_name(
|
|
province: str,
|
|
city: str,
|
|
county: Optional[str],
|
|
out_dir: Path,
|
|
prefer_full: bool = False,
|
|
) -> Path:
|
|
"""
|
|
通过省-市-县名称解析 adcode, 并直接拉取与保存 GeoJSON.
|
|
|
|
Parameters
|
|
----------
|
|
province : str
|
|
省份名称, 如 "湖北省". 不能为空.
|
|
city : str
|
|
城市名称, 如 "武汉市". 不能为空.
|
|
county : str, optional
|
|
区县名称, 如 "江岸区". 可为空 (表示只解析到市级).
|
|
out_dir : Path
|
|
输出目录, 用于保存 GeoJSON 文件.
|
|
prefer_full : bool, optional
|
|
是否下载下辖区划的 GeoJSON, 默认 False.
|
|
|
|
Returns
|
|
-------
|
|
Path
|
|
保存的 GeoJSON 文件路径.
|
|
|
|
Raises
|
|
------
|
|
ValueError
|
|
如果省或市为空, 或无法解析到行政区划代码.
|
|
"""
|
|
code = resolve_adcode_by_name(province, city, county, prefer_full=prefer_full)
|
|
if not code:
|
|
raise ValueError(
|
|
f"无法通过名称解析到行政区划代码: province={province}, city={city}, county={county}"
|
|
)
|
|
|
|
# 构建输出文件名
|
|
if county:
|
|
region_name = f"{province}_{city}_{county}"
|
|
else:
|
|
region_name = f"{province}_{city}"
|
|
|
|
return fetch_and_save_geojson(code, region_name, out_dir)
|
|
|
|
|
|
if __name__ == "__main__":
|
|
# 示例 1: 只获取市级边界 (province + city)
|
|
# out = fetch_and_save_geojson_by_name("湖北省", "武汉市", None, out_dir=Path("./data/vectors/"))
|
|
|
|
# 示例 2: 获取十堰市县区级边界
|
|
# out = fetch_and_save_geojson_by_name(
|
|
# "湖北省", "十堰市", None, out_dir=Path("./data/vectors/"), prefer_full=True
|
|
# )
|
|
|
|
# 示例 3: 获取区县级边界 (province + city + county)
|
|
out = fetch_and_save_geojson_by_name(
|
|
"湖北省", "十堰市", "郧西县", out_dir=Path("./data/vectors/"), prefer_full=True
|
|
)
|
|
|
|
print(f"Saved raw JSON and GeoJSON: {out}.")
|