Merge pull request #6 from HiDiHlabs/dev

Finalize first release

.pre-commit-config.yaml

@@ -1,5 +1,6 @@
 ci:
   autoupdate_schedule: quarterly
+  skip: [fmt, cargo-check]
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
     rev: v4.6.0

Cargo.toml

@@ -16,6 +16,7 @@ ndarray-stats = { version = "0.5.1" }
 num           = { version = "0.4.1" }
 numpy         = { version = "0.21.0" }
 polars        = { version = "0.41", features = ["partition_by", "dtype-categorical"] }
+polars-arrow  = { version = "0.41" }
 pyo3          = { version = "0.21.0", features = ["extension-module"] }
 pyo3-polars   = { version = "0.15.0" }
 rayon         = { version = "1.8.0" }

README.md

@@ -36,6 +36,12 @@ This project follows the [SemVer](https://semver.org/) guidelines for versioning
 
 ## Citations
 
+If you are using `sainsc` for your research please cite
+
+Müller-Bötticher, N., Tiesmeyer, S., Eils, R., and Ishaque, N.
+"Sainsc: a computational tool for segmentation-free analysis of in-situ capture"
+bioRxiv (2024) https://doi.org/10.1101/2024.08.02.603879
+
 ## License
 
 This project is licensed under the MIT License - for details please refer to the

docs/source/index.rst

@@ -10,10 +10,14 @@ It is easily integratable with the `scverse <https://github.com/scverse>`_
 `AnnData <https://anndata.readthedocs.io/>`_ or
 `SpatialData <https://spatialdata.scverse.org/>`_ format.
 
-.. Citations
-.. ---------
+Citations
+---------
 
-.. tbd
+If you are using `sainsc` for your research please cite
+
+Müller-Bötticher, N., Tiesmeyer, S., Eils, R., and Ishaque, N.
+"Sainsc: a computational tool for segmentation-free analysis of in-situ capture"
+bioRxiv (2024) https://doi.org/10.1101/2024.08.02.603879
 
 .. toctree::
    :maxdepth: 1

docs/source/installation.rst

@@ -5,19 +5,17 @@ Installation
 PyPI and ``pip``
 ----------------
 
-``sainsc`` will soon be available to install from `PyPI <https://pypi.org/>`_.
+To install ``sainsc`` from `PyPI <https://pypi.org/>`_ using ``pip`` just run
 
-.. To install ``sainsc`` from `PyPI <https://pypi.org/>`_ using ``pip`` just run
-
-.. .. code-block:: bash
+.. code-block:: bash
 
-..     pip install sainsc
+    pip install sainsc
 
-.. If you want to have support for :py:mod:`spatialdata` use
+If you want to have support for :py:mod:`spatialdata` use
 
-.. .. code-block:: bash
+.. code-block:: bash
 
-..     pip install 'sainsc[spatialdata]'
+    pip install sainsc[spatialdata]
 
 
 Bioconda and ``conda``

sainsc/_utils_rust.pyi

@@ -135,6 +135,15 @@ class GridCounts:
         sainsc.GridCounts
         """
 
+    def as_dataframe(self) -> DataFrame:
+        """
+        Convert to a dataframe with 'gene', 'x', 'y', and 'count' column.
+
+        Returns
+        -------
+        polars.DataFrame
+        """
+
     def __getitem__(self, key: str) -> _CsxArray: ...
     def __setitem__(self, key: str, value: _Csx): ...
     def __delitem__(self, key: str): ...

src/gridcounts.rs

@@ -1,7 +1,7 @@
 use crate::sparsearray_conversion::WrappedCsx;
 use crate::utils::create_pool;
 use bincode::{deserialize, serialize};
-use itertools::MultiUnzip;
+use itertools::Itertools;
 use ndarray::Array2;
 use num::Zero;
 use numpy::{IntoPyArray, PyArray2, PyReadonlyArray2};
@@ -12,12 +12,13 @@ use polars::{
     },
     prelude::*,
 };
+use polars_arrow::array::{DictionaryArray, UInt32Array, Utf8Array};
 use pyo3::{
     exceptions::{PyKeyError, PyRuntimeError, PyValueError},
     prelude::*,
     types::{PyBytes, PyType},
 };
-use pyo3_polars::PyDataFrame;
+use pyo3_polars::{error::PyPolarsErr, PyDataFrame};
 use rayon::{
     iter::{
         IntoParallelIterator, IntoParallelRefIterator, IntoParallelRefMutIterator, ParallelIterator,
@@ -31,6 +32,7 @@ use sprs::{
 use std::{
     cmp::min,
     collections::{HashMap, HashSet},
+    iter::repeat,
     ops::AddAssign,
 };
 
@@ -203,7 +205,7 @@ impl GridCounts {
         let resolution = resolution.map(|r| r * binsize.unwrap_or(1.));
 
         match threadpool.install(|| _from_dataframe(df.into(), binsize)) {
-            Err(e) => Err(PyValueError::new_err(e.to_string())),
+            Err(e) => Err(PyPolarsErr::from(e).into()),
             Ok((counts, shape)) => Ok(Self {
                 counts,
                 shape,
@@ -426,4 +428,41 @@ impl GridCounts {
             });
         });
     }
+
+    fn as_dataframe(&mut self) -> PyResult<PyDataFrame> {
+        self.to_format(CSR);
+
+        let genes: Vec<_> = self.counts.keys().sorted().collect();
+
+        let ((counts, (x, y)), gene_idx): ((Vec<&Count>, (Vec<_>, Vec<_>)), Vec<_>) = genes
+            .iter()
+            .zip(0u32..)
+            .flat_map(|(&gene, i)| {
+                self.get_view(gene)
+                    .expect("gene exists because we collected the keys above")
+                    .iter_rbr()
+                    .zip(repeat(i))
+            })
+            .multiunzip();
+
+        let counts = Series::from_iter(counts).with_name("count");
+        let x = Series::from_vec("x", x);
+        let y = Series::from_vec("y", y);
+        // construct categorical gene array from codes and categories
+        let genes = Series::from_arrow(
+            "gene",
+            Box::new(
+                DictionaryArray::try_from_keys(
+                    UInt32Array::from_vec(gene_idx),
+                    Box::new(Utf8Array::<i32>::from_iter(genes.into_iter().map(Some))),
+                )
+                .map_err(PyPolarsErr::from)?,
+            ),
+        )
+        .map_err(PyPolarsErr::from)?;
+
+        let df = DataFrame::new(vec![genes, x, y, counts]).map_err(PyPolarsErr::from)?;
+
+        Ok(PyDataFrame(df))
+    }
 }