feat: add geopandas notebook

CHANGELOG.md

@@ -6,6 +6,10 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 
 ## [Unreleased]
 
+### Added
+
+- Create a item collection from an arrow table ([#57](https://github.com/stac-utils/stacrs/pull/57))
+
 ## [0.5.6-beta.0] - 2025-02-22
 
 ### Added

Cargo.toml

@@ -9,9 +9,10 @@ name = "stacrs"
 crate-type = ["cdylib"]
 
 [dependencies]
-clap = "4.5.30"
+clap = "4.5.31"
 geojson = "0.24.1"
-pyo3 = { version = "0.23.4", features = ["extension-module"] }
+geoarrow = "0.4.0-beta.3"
+pyo3 = { version = "0.23.5", features = ["extension-module"] }
 pyo3-async-runtimes = { version = "0.23.0", features = [
     "tokio",
     "tokio-runtime",

pyproject.toml

@@ -25,6 +25,9 @@ dependencies = []
 
 [project.optional-dependencies]
 arrow = ["arro3-core>=0.4.5"]
+docs = [
+    "jinja2>=3.1.4",
+]
 
 [project.scripts]
 stacrs = "stacrs:main"
@@ -58,6 +61,9 @@ dev = [
     "stac-geoparquet>=0.6.0",
 ]
 docs = [
+    "contextily>=1.6.2",
+    "humanize>=4.12.1",
+    "jinja2>=3.1.4",
     "mike>=2.1.3",
     "mkdocs-jupyter>=0.25.1",
     "mkdocs-material[imaging]>=9.5.45",

src/arrow.rs

@@ -0,0 +1,55 @@
+use crate::{Error, Json, Result};
+use geoarrow::table::Table;
+use pyo3::{prelude::*, IntoPyObjectExt};
+use pyo3_arrow::PyTable;
+use serde_json::Value;
+use stac::{Item, ItemCollection};
+
+#[pyfunction]
+pub fn from_arrow(py: Python<'_>, table: PyTable) -> PyResult<Bound<PyAny>> {
+    let (record_batches, mut schema) = table.into_inner();
+    let record_batches = record_batches
+        .into_iter()
+        .map(|record_batch| {
+            let record_batch = stac::geoarrow::with_native_geometry(record_batch, "geometry")?;
+            Ok(record_batch)
+        })
+        .collect::<Result<Vec<_>>>()?;
+    if !record_batches.is_empty() {
+        schema = record_batches[0].schema();
+    }
+    let table = Table::try_new(record_batches, schema).map_err(Error::from)?;
+    let item_collection = stac::geoarrow::from_table(table).map_err(Error::from)?;
+    let item_collection = Json(item_collection).into_pyobject(py)?;
+    Ok(item_collection)
+}
+
+#[pyfunction]
+pub fn to_arrow(py: Python<'_>, items: Bound<PyAny>) -> PyResult<PyObject> {
+    let value: Value = pythonize::depythonize(&items)?;
+    let item_collection = if let Value::Array(array) = value {
+        let items = array
+            .into_iter()
+            .map(|value| serde_json::from_value::<Item>(value).map_err(Error::from))
+            .collect::<Result<Vec<_>>>()?;
+        ItemCollection::from(items)
+    } else {
+        serde_json::from_value(value).map_err(Error::from)?
+    };
+    // TODO we might want to just allow use to go WKB right when we got to table?
+    let (record_batches, mut schema) = stac::geoarrow::to_table(item_collection)
+        .map_err(Error::from)?
+        .into_inner();
+    let record_batches = record_batches
+        .into_iter()
+        .map(|record_batch| {
+            stac::geoarrow::with_wkb_geometry(record_batch, "geometry").map_err(Error::from)
+        })
+        .collect::<Result<Vec<_>>>()?;
+    if !record_batches.is_empty() {
+        schema = record_batches[0].schema();
+    }
+    let table = PyTable::try_new(record_batches, schema)?;
+    let table = table.to_arro3(py)?;
+    Ok(table.into_py_any(py)?)
+}

src/error.rs

@@ -12,6 +12,9 @@ pub enum Error {
     #[error(transparent)]
     Geojson(#[from] geojson::Error),
 
+    #[error(transparent)]
+    Geoarrow(#[from] geoarrow::error::GeoArrowError),
+
     #[error(transparent)]
     Io(#[from] std::io::Error),
 

src/lib.rs

@@ -1,5 +1,6 @@
 #![deny(unused_crate_dependencies)]
 
+mod arrow;
 mod cli;
 mod duckdb;
 mod error;
@@ -20,6 +21,8 @@ fn stacrs(py: Python<'_>, m: &Bound<'_, PyModule>) -> PyResult<()> {
 
     m.add_class::<duckdb::DuckdbClient>()?;
 
+    m.add_function(wrap_pyfunction!(arrow::from_arrow, m)?)?;
+    m.add_function(wrap_pyfunction!(arrow::to_arrow, m)?)?;
     m.add_function(wrap_pyfunction!(cli::main, m)?)?;
     m.add_function(wrap_pyfunction!(migrate::migrate, m)?)?;
     m.add_function(wrap_pyfunction!(migrate::migrate_href, m)?)?;

stacrs.pyi

@@ -212,6 +212,36 @@ async def read(
         >>> item = await stacrs.read("item.json")
     """
 
+def from_arrow(
+    table: arro3.core.Table,
+) -> dict[str, Any]:
+    """
+    Converts an [arro3.core.table][] to a STAC item collection.
+
+    Requires **stacrs** to be installed with the `arrow` extra.
+
+    Args:
+        table: The table
+
+    Returns:
+        dict[str, Any]: The STAC item collection
+    """
+
+def to_arrow(
+    items: list[dict[str, Any]] | dict[str, Any],
+) -> arro3.core.Table:
+    """
+    Converts items to an [arro3.core.table][].
+
+    Requires **stacrs** to be installed with the `arrow` extra.
+
+    Args:
+        items: Either an iterable of items or a item collection
+
+    Returns:
+        arro3.core.Table: The table
+    """
+
 async def search(
     href: str,
     *,