refactor: 重构项目结构,将geo_tools重命名为app并更新相关引用

- 将主包名从geo_tools改为app
- 更新所有模块中的引用路径
- 迁移并更新测试用例
- 添加项目规则文档
- 保持原有功能不变,仅进行结构调整
This commit is contained in:
2026-04-12 19:49:56 +08:00
parent fcb8e1f255
commit db51d41aef
41 changed files with 4132 additions and 808 deletions

1
app/analysis/__init__.py Normal file
View File

@@ -0,0 +1 @@
"""geo_tools.analysis 包 —— 空间分析层。"""

149
app/analysis/spatial_ops.py Normal file
View File

@@ -0,0 +1,149 @@
"""
geo_tools.analysis.spatial_ops
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
空间叠加与邻域分析操作。
"""
from __future__ import annotations
from typing import Any
import geopandas as gpd
import pandas as pd
from app.utils.logger import get_logger
logger = get_logger(__name__)
def buffer_and_overlay(
source: gpd.GeoDataFrame,
distance: float,
target: gpd.GeoDataFrame,
how: str = "intersection",
projected_crs: str | None = None,
) -> gpd.GeoDataFrame:
"""对 source 执行缓冲区后与 target 执行叠置分析。
Parameters
----------
source:
源图层(生成缓冲区)。
distance:
缓冲距离(与 ``projected_crs`` 单位一致)。
target:
叠置目标图层。
how:
叠置类型:``"intersection"``、``"union"``、``"difference"``、``"symmetric_difference"``、``"identity"``。
projected_crs:
执行缓冲区前先投影到此 CRS建议使用平面坐标系以保证距离精度
``None`` 则使用 source 的当前 CRS地理 CRS 下 distance 单位为度)。
Returns
-------
gpd.GeoDataFrame
"""
original_crs = source.crs
if projected_crs:
source = source.to_crs(projected_crs)
target = target.to_crs(projected_crs)
buffered = source.copy()
buffered["geometry"] = buffered.geometry.buffer(distance)
logger.debug("缓冲区完成distance=%.2f执行叠置分析how=%s", distance, how)
result = gpd.overlay(buffered, target, how=how, keep_geom_type=False)
if projected_crs:
result = result.to_crs(original_crs) # type: ignore
logger.info("叠置分析完成:%d 条结果", len(result))
return result
def overlay(
df1: gpd.GeoDataFrame,
df2: gpd.GeoDataFrame,
how: str = "intersection",
keep_geom_type: bool = True,
) -> gpd.GeoDataFrame:
"""封装 geopandas overlay自动对齐 CRS。
Parameters
----------
how:
叠置类型:``"intersection"``、``"union"``、``"difference"``、
``"symmetric_difference"``、``"identity"``。
"""
if df1.crs != df2.crs:
df2 = df2.to_crs(df1.crs) # type: ignore
result = gpd.overlay(df1, df2, how=how, keep_geom_type=keep_geom_type)
logger.debug("overlay(%s)%d 条结果", how, len(result))
return result
def nearest_features(
source: gpd.GeoDataFrame,
target: gpd.GeoDataFrame,
k: int = 1,
max_distance: float | None = None,
) -> gpd.GeoDataFrame:
"""为 source 中每条要素找到 target 中最近的 k 个要素。
Parameters
----------
source:
查询图层。
target:
被查询图层。
k:
最近邻数量。
max_distance:
最大搜索距离(与 CRS 单位一致),``None`` 表示无限制。
Returns
-------
gpd.GeoDataFrame
连接了最近 target 属性的 source GDF可能包含重复行每行对应一个近邻
"""
if source.crs != target.crs:
target = target.to_crs(source.crs) # type: ignore
result = gpd.sjoin_nearest(
source,
target,
how="left",
max_distance=max_distance,
distance_col="nearest_distance",
lsuffix="left",
rsuffix="right",
)
logger.debug("最近邻分析完成k=%d%d 条结果", k, len(result))
return result
def select_by_location(
source: gpd.GeoDataFrame,
selector: gpd.GeoDataFrame,
predicate: str = "intersects",
) -> gpd.GeoDataFrame:
"""按位置关系从 source 中选取要素(等同于 ArcGIS「按位置选择」
Parameters
----------
predicate:
空间谓词:``"intersects"``、``"within"``、``"contains"``、``"touches"``。
Returns
-------
gpd.GeoDataFrame
满足条件的 source 子集。
"""
if source.crs != selector.crs:
selector = selector.to_crs(source.crs) # type: ignore
joined = gpd.sjoin(source, selector, how="inner", predicate=predicate)
result = source.loc[source.index.isin(joined.index)].copy()
logger.debug("按位置选择(%s%d / %d", predicate, len(result), len(source))
return result

136
app/analysis/stats.py Normal file
View File

@@ -0,0 +1,136 @@
"""
geo_tools.analysis.stats
~~~~~~~~~~~~~~~~~~~~~~~~~
空间统计工具:属性汇总、面积加权均值、空间自相关指数等。
"""
from __future__ import annotations
import geopandas as gpd
import numpy as np
import pandas as pd
from app.utils.logger import get_logger
logger = get_logger(__name__)
def area_weighted_mean(
gdf: gpd.GeoDataFrame,
value_col: str,
group_col: str | None = None,
projected_crs: str = "EPSG:3857",
) -> pd.Series | pd.DataFrame:
"""计算面积加权均值。
Parameters
----------
gdf:
输入 GeoDataFrame面要素
value_col:
需要加权平均的属性列名。
group_col:
分组字段名;若为 ``None`` 则对整个 GDF 计算单一结果。
projected_crs:
用于计算面积的平面投影 CRS。
Returns
-------
pd.Series无分组或 pd.DataFrame有分组
"""
gdf = gdf.copy()
# 计算面积
if not gdf.crs or not gdf.crs.is_projected:
projected = gdf.to_crs(projected_crs)
else:
projected = gdf
gdf["_area"] = projected.geometry.area
if group_col is None:
total_area = gdf["_area"].sum()
result = (gdf[value_col] * gdf["_area"]).sum() / total_area
return pd.Series({"area_weighted_mean": result, "total_area": total_area})
def _weighted(group: pd.DataFrame) -> float:
return float((group[value_col] * group["_area"]).sum() / group["_area"].sum())
result = gdf.groupby(group_col).apply(_weighted, include_groups=False).rename("area_weighted_mean") # type: ignore[no-untyped-call]
area_sum = gdf.groupby(group_col)["_area"].sum().rename("total_area")
return pd.concat([result, area_sum], axis=1).reset_index()
def summarize_attributes(
gdf: gpd.GeoDataFrame,
columns: list[str] | None = None,
group_col: str | None = None,
agg_funcs: list[str] | None = None,
) -> pd.DataFrame:
"""对属性列进行统计汇总(最大、最小、均值、总和等)。
Parameters
----------
gdf:
输入 GeoDataFrame。
columns:
统计的列名列表;``None`` 则自动选取所有数值列。
group_col:
分组字段名;``None`` 则对全局统计。
agg_funcs:
聚合函数列表,默认 ``["count", "mean", "min", "max", "sum", "std"]``。
Returns
-------
pd.DataFrame
"""
if agg_funcs is None:
agg_funcs = ["count", "mean", "min", "max", "sum", "std"]
df = gdf.drop(columns=["geometry"], errors="ignore")
if columns is None:
columns = df.select_dtypes(include="number").columns.tolist()
if not columns:
raise ValueError("未找到数值列,请显式指定 columns 参数。")
subset = df[columns]
if group_col is None:
return subset.agg(agg_funcs).T.rename_axis("column").reset_index() # type: ignore[no-untyped-call]
df_with_group = df[[group_col] + columns]
return df_with_group.groupby(group_col)[columns].agg(agg_funcs).reset_index()
def count_by_polygon(
points: gpd.GeoDataFrame,
polygons: gpd.GeoDataFrame,
count_col: str = "point_count",
) -> gpd.GeoDataFrame:
"""统计每个面要素内的点要素数量(类似 ArcGIS「面要素统计点」
Parameters
----------
points:
点图层。
polygons:
面图层。
count_col:
新增计数列名。
Returns
-------
gpd.GeoDataFrame
含 ``count_col`` 列的 polygons 副本。
"""
if points.crs != polygons.crs:
points = points.to_crs(polygons.crs) # type: ignore
joined = gpd.sjoin(points, polygons, how="inner", predicate="within")
point_counts = joined.groupby("index_right").size().rename(count_col)
result = polygons.copy()
result = result.join(point_counts)
result[count_col] = result[count_col].fillna(0).astype(int)
return result