Developer (#164)

- State before merging the new cluster vector API

---------

Co-authored-by: Patrick <patrick.sieberer@psi.ch>
Co-authored-by: JulianHeymes <julian.heymes@psi.ch>
Co-authored-by: Dhanya Thattil <dhanya.thattil@psi.ch>
Co-authored-by: Xiangyu Xie <45243914+xiangyuxie@users.noreply.github.com>
Co-authored-by: xiangyu.xie <xiangyu.xie@psi.ch>
Co-authored-by: siebsi <sieb.patr@gmail.com>

python/CMakeLists.txt

@@ -1,12 +1,13 @@
 
-find_package (Python 3.10 COMPONENTS Interpreter Development REQUIRED)
+find_package (Python 3.10 COMPONENTS Interpreter Development.Module REQUIRED)
+set(PYBIND11_FINDPYTHON ON) # Needed for RH8
 
 # Download or find pybind11 depending on configuration
 if(AARE_FETCH_PYBIND11)
     FetchContent_Declare(
         pybind11
         GIT_REPOSITORY https://github.com/pybind/pybind11
-        GIT_TAG        v2.13.0
+        GIT_TAG        v2.13.6
     )
     FetchContent_MakeAvailable(pybind11)
 else()
@@ -58,10 +59,16 @@ endforeach(FILE ${PYTHON_EXAMPLES})
 
 
 if(AARE_INSTALL_PYTHONEXT)
-    install(TARGETS _aare
+    install(
+        TARGETS _aare
         EXPORT "${TARGETS_EXPORT_NAME}"
         LIBRARY DESTINATION aare
+        COMPONENT python
     )
 
-    install(FILES ${PYTHON_FILES} DESTINATION aare)
+    install(
+        FILES ${PYTHON_FILES} 
+        DESTINATION aare
+        COMPONENT python
+        )
 endif()

python/aare/__init__.py

@@ -2,7 +2,7 @@
 from . import _aare
 
 
-from ._aare import File, RawMasterFile, RawSubFile
+from ._aare import File, RawMasterFile, RawSubFile, JungfrauDataFile
 from ._aare import Pedestal_d, Pedestal_f, ClusterFinder, VarClusterFinder
 from ._aare import DetectorType
 from ._aare import ClusterFile
@@ -13,11 +13,15 @@ from ._aare import ClusterFinderMT, ClusterCollector, ClusterFileSink, ClusterVe
 
 from ._aare import fit_gaus, fit_pol1
 from ._aare import Interpolator
+
+
+from ._aare import apply_custom_weights
+
 from .CtbRawFile import CtbRawFile
 from .RawFile import RawFile
 from .ScanParameters import ScanParameters
 
-from .utils import random_pixels, random_pixel, flat_list
+from .utils import random_pixels, random_pixel, flat_list, add_colorbar
 
 
 #make functions available in the top level API

python/aare/utils.py

@@ -1,4 +1,6 @@
 import numpy as np
+import matplotlib.pyplot as plt
+from mpl_toolkits.axes_grid1 import make_axes_locatable
 
 def random_pixels(n_pixels, xmin=0, xmax=512, ymin=0, ymax=1024):
     """Return a list of random pixels.
@@ -24,4 +26,11 @@ def random_pixel(xmin=0, xmax=512, ymin=0, ymax=1024):
 
 def flat_list(xss):
     """Flatten a list of lists."""
-    return [x for xs in xss for x in xs]
+    return [x for xs in xss for x in xs]
+
+def add_colorbar(ax, im, size="5%", pad=0.05):
+    """Add a colorbar with the same height as the image."""
+    divider = make_axes_locatable(ax)
+    cax = divider.append_axes("right", size=size, pad=pad)
+    plt.colorbar(im, cax=cax)
+    return ax, im, cax

python/src/ctb_raw_file.hpp

@@ -10,6 +10,8 @@
 #include "aare/decode.hpp"
 // #include "aare/fClusterFileV2.hpp"
 
+#include "np_helper.hpp"
+
 #include <cstdint>
 #include <filesystem>
 #include <pybind11/iostream.h>
@@ -65,35 +67,54 @@ m.def("adc_sar_04_decode64to16", [](py::array_t<uint8_t> input) {
     return output;
 });
 
- py::class_<CtbRawFile>(m, "CtbRawFile")
-        .def(py::init<const std::filesystem::path &>())
-        .def("read_frame",
-             [](CtbRawFile &self) {
-                 size_t image_size = self.image_size_in_bytes();
-                 py::array image;
-                 std::vector<ssize_t> shape;
-                 shape.reserve(2);
-                 shape.push_back(1);
-                 shape.push_back(image_size);
+m.def(
+    "apply_custom_weights",
+    [](py::array_t<uint16_t, py::array::c_style | py::array::forcecast> &input,
+       py::array_t<double, py::array::c_style | py::array::forcecast>
+           &weights) {
+        
 
-                 py::array_t<DetectorHeader> header(1);
+        // Create new array with same shape as the input array (uninitialized values)
+        py::buffer_info buf = input.request();
+        py::array_t<double> output(buf.shape);
 
-                 // always read bytes
-                 image = py::array_t<uint8_t>(shape);
+        // Use NDViews to call into the C++ library
+        auto weights_view = make_view_1d(weights);
+        NDView<uint16_t, 1> input_view(input.mutable_data(), {input.size()});
+        NDView<double, 1> output_view(output.mutable_data(), {output.size()});
 
-                 self.read_into(
-                     reinterpret_cast<std::byte *>(image.mutable_data()),
-                     header.mutable_data());
+        apply_custom_weights(input_view, output_view, weights_view);
+        return output;
+    });
 
-                 return py::make_tuple(header, image);
-             })
-        .def("seek", &CtbRawFile::seek)
-        .def("tell", &CtbRawFile::tell)
-        .def("master", &CtbRawFile::master)
+py::class_<CtbRawFile>(m, "CtbRawFile")
+    .def(py::init<const std::filesystem::path &>())
+    .def("read_frame",
+         [](CtbRawFile &self) {
+             size_t image_size = self.image_size_in_bytes();
+             py::array image;
+             std::vector<ssize_t> shape;
+             shape.reserve(2);
+             shape.push_back(1);
+             shape.push_back(image_size);
 
-        .def_property_readonly("image_size_in_bytes",
-                               &CtbRawFile::image_size_in_bytes)
+             py::array_t<DetectorHeader> header(1);
 
-        .def_property_readonly("frames_in_file", &CtbRawFile::frames_in_file);
+             // always read bytes
+             image = py::array_t<uint8_t>(shape);
 
-}
+             self.read_into(reinterpret_cast<std::byte *>(image.mutable_data()),
+                            header.mutable_data());
+
+             return py::make_tuple(header, image);
+         })
+    .def("seek", &CtbRawFile::seek)
+    .def("tell", &CtbRawFile::tell)
+    .def("master", &CtbRawFile::master)
+
+    .def_property_readonly("image_size_in_bytes",
+                           &CtbRawFile::image_size_in_bytes)
+
+    .def_property_readonly("frames_in_file", &CtbRawFile::frames_in_file);
+
+}

python/src/file.hpp

@@ -20,6 +20,9 @@
 namespace py = pybind11;
 using namespace ::aare;
 
+
+
+
 //Disable warnings for unused parameters, as we ignore some
 //in the __exit__ method
 #pragma GCC diagnostic push
@@ -214,36 +217,9 @@ void define_file_io_bindings(py::module &m) {
     
 
 
-    py::class_<RawSubFile>(m, "RawSubFile")
-        .def(py::init<const std::filesystem::path &, DetectorType, size_t,
-                      size_t, size_t>())
-        .def_property_readonly("bytes_per_frame", &RawSubFile::bytes_per_frame)
-        .def_property_readonly("pixels_per_frame",
-                               &RawSubFile::pixels_per_frame)
-        .def("seek", &RawSubFile::seek)
-        .def("tell", &RawSubFile::tell)
-        .def_property_readonly("rows", &RawSubFile::rows)
-        .def_property_readonly("cols", &RawSubFile::cols)
-        .def("read_frame",
-             [](RawSubFile &self) {
-                 const uint8_t item_size = self.bytes_per_pixel();
-                 py::array image;
-                 std::vector<ssize_t> shape;
-                 shape.reserve(2);
-                 shape.push_back(self.rows());
-                 shape.push_back(self.cols());
-                 if (item_size == 1) {
-                     image = py::array_t<uint8_t>(shape);
-                 } else if (item_size == 2) {
-                     image = py::array_t<uint16_t>(shape);
-                 } else if (item_size == 4) {
-                     image = py::array_t<uint32_t>(shape);
-                 }
-                 fmt::print("item_size: {} rows: {} cols: {}\n", item_size, self.rows(), self.cols());
-                 self.read_into(
-                     reinterpret_cast<std::byte *>(image.mutable_data()));
-                 return image;
-             });
+
+
+
 
 #pragma GCC diagnostic pop
     // py::class_<ClusterHeader>(m, "ClusterHeader")

python/src/jungfrau_data_file.hpp

@@ -0,0 +1,116 @@
+
+#include "aare/JungfrauDataFile.hpp"
+#include "aare/defs.hpp"
+
+#include <cstdint>
+#include <filesystem>
+#include <pybind11/iostream.h>
+#include <pybind11/numpy.h>
+#include <pybind11/pybind11.h>
+#include <pybind11/stl.h>
+#include <pybind11/stl/filesystem.h>
+#include <string>
+
+namespace py = pybind11;
+using namespace ::aare;
+
+// Disable warnings for unused parameters, as we ignore some
+// in the __exit__ method
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wunused-parameter"
+
+auto read_dat_frame(JungfrauDataFile &self) {
+    py::array_t<JungfrauDataHeader> header(1);
+    py::array_t<uint16_t> image({
+        self.rows(), 
+        self.cols() 
+    });
+
+    self.read_into(reinterpret_cast<std::byte *>(image.mutable_data()),
+                   header.mutable_data());
+
+    return py::make_tuple(header, image);
+}
+
+auto read_n_dat_frames(JungfrauDataFile &self, size_t n_frames) {
+    // adjust for actual frames left in the file
+    n_frames = std::min(n_frames, self.total_frames() - self.tell());
+    if (n_frames == 0) {
+        throw std::runtime_error("No frames left in file");
+    }
+
+    py::array_t<JungfrauDataHeader> header(n_frames);
+    py::array_t<uint16_t> image({
+        n_frames, self.rows(), 
+        self.cols()});
+
+    self.read_into(reinterpret_cast<std::byte *>(image.mutable_data()),
+                   n_frames, header.mutable_data());
+
+    return py::make_tuple(header, image);
+}
+
+void define_jungfrau_data_file_io_bindings(py::module &m) {
+    // Make the JungfrauDataHeader usable from numpy
+    PYBIND11_NUMPY_DTYPE(JungfrauDataHeader, framenum, bunchid);
+
+    py::class_<JungfrauDataFile>(m, "JungfrauDataFile")
+        .def(py::init<const std::filesystem::path &>())
+        .def("seek", &JungfrauDataFile::seek,
+             R"(
+               Seek to the given frame index.
+               )")
+        .def("tell", &JungfrauDataFile::tell,
+             R"(
+               Get the current frame index.
+               )")
+        .def_property_readonly("rows", &JungfrauDataFile::rows)
+        .def_property_readonly("cols", &JungfrauDataFile::cols)
+        .def_property_readonly("base_name", &JungfrauDataFile::base_name)
+        .def_property_readonly("bytes_per_frame",
+                               &JungfrauDataFile::bytes_per_frame)
+        .def_property_readonly("pixels_per_frame",
+                               &JungfrauDataFile::pixels_per_frame)
+        .def_property_readonly("bytes_per_pixel",
+                               &JungfrauDataFile::bytes_per_pixel)
+        .def_property_readonly("bitdepth", &JungfrauDataFile::bitdepth)
+        .def_property_readonly("current_file", &JungfrauDataFile::current_file)
+        .def_property_readonly("total_frames", &JungfrauDataFile::total_frames)
+        .def_property_readonly("n_files", &JungfrauDataFile::n_files)
+        .def("read_frame", &read_dat_frame,
+             R"(
+               Read a single frame from the file.
+               )")
+        .def("read_n", &read_n_dat_frames,
+             R"(
+               Read maximum n_frames frames from the file.
+               )")
+        .def(
+            "read",
+            [](JungfrauDataFile &self) {
+                self.seek(0);
+                auto n_frames = self.total_frames();
+                return read_n_dat_frames(self, n_frames);
+            },
+            R"(
+              Read all frames from the file. Seeks to the beginning before reading.
+              )")
+        .def("__enter__", [](JungfrauDataFile &self) { return &self; })
+        .def("__exit__",
+             [](JungfrauDataFile &self,
+                const std::optional<pybind11::type> &exc_type,
+                const std::optional<pybind11::object> &exc_value,
+                const std::optional<pybind11::object> &traceback) {
+                 //  self.close();
+             })
+        .def("__iter__", [](JungfrauDataFile &self) { return &self; })
+        .def("__next__", [](JungfrauDataFile &self) {
+            try {
+                return read_dat_frame(self);
+            } catch (std::runtime_error &e) {
+                throw py::stop_iteration();
+            }
+        });
+}
+
+#pragma GCC diagnostic pop

python/src/module.cpp

@@ -10,6 +10,9 @@
 #include "cluster_file.hpp"
 #include "fit.hpp"
 #include "interpolation.hpp"
+#include "raw_sub_file.hpp"
+
+#include "jungfrau_data_file.hpp"
 
 //Pybind stuff
 #include <pybind11/pybind11.h>
@@ -20,6 +23,7 @@ namespace py = pybind11;
 PYBIND11_MODULE(_aare, m) {
     define_file_io_bindings(m);
     define_raw_file_io_bindings(m);
+    define_raw_sub_file_io_bindings(m);
     define_ctb_raw_file_io_bindings(m);
     define_raw_master_file_bindings(m);
     define_var_cluster_finder_bindings(m);
@@ -33,5 +37,6 @@ PYBIND11_MODULE(_aare, m) {
     define_cluster_file_sink_bindings(m);
     define_fit_bindings(m);
     define_interpolation_bindings(m);
+    define_jungfrau_data_file_io_bindings(m);
 
 }

python/src/raw_sub_file.hpp

@@ -0,0 +1,110 @@
+#include "aare/CtbRawFile.hpp"
+#include "aare/File.hpp"
+#include "aare/Frame.hpp"
+#include "aare/RawFile.hpp"
+#include "aare/RawMasterFile.hpp"
+#include "aare/RawSubFile.hpp"
+
+#include "aare/defs.hpp"
+// #include "aare/fClusterFileV2.hpp"
+
+#include <cstdint>
+#include <filesystem>
+#include <pybind11/iostream.h>
+#include <pybind11/numpy.h>
+#include <pybind11/pybind11.h>
+#include <pybind11/stl.h>
+#include <pybind11/stl/filesystem.h>
+#include <string>
+
+namespace py = pybind11;
+using namespace ::aare;
+
+auto read_frame_from_RawSubFile(RawSubFile &self) {
+    py::array_t<DetectorHeader> header(1);
+    const uint8_t item_size = self.bytes_per_pixel();
+    std::vector<ssize_t> shape{static_cast<ssize_t>(self.rows()),
+        static_cast<ssize_t>(self.cols())};
+    
+    py::array image;
+    if (item_size == 1) {
+        image = py::array_t<uint8_t>(shape);
+    } else if (item_size == 2) {
+        image = py::array_t<uint16_t>(shape);
+    } else if (item_size == 4) {
+        image = py::array_t<uint32_t>(shape);
+    }
+    self.read_into(reinterpret_cast<std::byte *>(image.mutable_data()),
+                   header.mutable_data());
+
+    return py::make_tuple(header, image);
+}
+
+auto read_n_frames_from_RawSubFile(RawSubFile &self, size_t n_frames) {
+    py::array_t<DetectorHeader> header(n_frames);
+    const uint8_t item_size = self.bytes_per_pixel();
+    std::vector<ssize_t> shape{
+        static_cast<ssize_t>(n_frames), 
+        static_cast<ssize_t>(self.rows()),
+        static_cast<ssize_t>(self.cols())
+    };
+    
+    py::array image;
+    if (item_size == 1) {
+        image = py::array_t<uint8_t>(shape);
+    } else if (item_size == 2) {
+        image = py::array_t<uint16_t>(shape);
+    } else if (item_size == 4) {
+        image = py::array_t<uint32_t>(shape);
+    }
+    self.read_into(reinterpret_cast<std::byte *>(image.mutable_data()), n_frames,
+                   header.mutable_data());
+
+    return py::make_tuple(header, image);
+}
+
+
+//Disable warnings for unused parameters, as we ignore some
+//in the __exit__ method
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Wunused-parameter"
+
+void define_raw_sub_file_io_bindings(py::module &m) {
+    py::class_<RawSubFile>(m, "RawSubFile")
+        .def(py::init<const std::filesystem::path &, DetectorType, size_t,
+                      size_t, size_t>())
+        .def_property_readonly("bytes_per_frame", &RawSubFile::bytes_per_frame)
+        .def_property_readonly("pixels_per_frame",
+                               &RawSubFile::pixels_per_frame)
+        .def_property_readonly("bytes_per_pixel", &RawSubFile::bytes_per_pixel)                    
+        .def("seek", &RawSubFile::seek)
+        .def("tell", &RawSubFile::tell)
+        .def_property_readonly("rows", &RawSubFile::rows)
+        .def_property_readonly("cols", &RawSubFile::cols)
+        .def_property_readonly("frames_in_file", &RawSubFile::frames_in_file)
+        .def("read_frame", &read_frame_from_RawSubFile)
+        .def("read_n", &read_n_frames_from_RawSubFile)
+        .def("read", [](RawSubFile &self){
+            self.seek(0);
+            auto n_frames = self.frames_in_file();
+            return read_n_frames_from_RawSubFile(self, n_frames);
+        })
+        .def("__enter__", [](RawSubFile &self) { return &self; })
+        .def("__exit__",
+             [](RawSubFile &self,
+                const std::optional<pybind11::type> &exc_type,
+                const std::optional<pybind11::object> &exc_value,
+                const std::optional<pybind11::object> &traceback) {
+             })
+        .def("__iter__", [](RawSubFile &self) { return &self; })
+        .def("__next__", [](RawSubFile &self) {
+            try {
+                return read_frame_from_RawSubFile(self);
+            } catch (std::runtime_error &e) {
+                throw py::stop_iteration();
+            }
+        });
+
+}
+
+#pragma GCC diagnostic pop

python/tests/conftest.py

@@ -0,0 +1,29 @@
+import os
+from pathlib import Path
+import pytest
+
+
+
+def pytest_addoption(parser):
+    parser.addoption(
+        "--files", action="store_true", default=False, help="run slow tests"
+    )
+
+
+def pytest_configure(config):
+    config.addinivalue_line("markers", "files: mark test as needing image files to run")
+
+
+def pytest_collection_modifyitems(config, items):
+    if config.getoption("--files"):
+        return
+    skip = pytest.mark.skip(reason="need --files option to run")
+    for item in items:
+        if "files" in item.keywords:
+            item.add_marker(skip)
+
+
+@pytest.fixture
+def test_data_path():
+    return Path(os.environ["AARE_TEST_DATA"])
+

python/tests/test_RawSubFile.py

@@ -0,0 +1,36 @@
+import pytest
+import numpy as np
+from aare import RawSubFile, DetectorType
+
+
+@pytest.mark.files
+def test_read_a_jungfrau_RawSubFile(test_data_path):
+    with RawSubFile(test_data_path / "raw/jungfrau/jungfrau_single_d0_f1_0.raw", DetectorType.Jungfrau, 512, 1024, 16) as f:
+        assert f.frames_in_file == 3
+
+        headers, frames = f.read()
+
+    assert headers.size == 3
+    assert frames.shape == (3, 512, 1024)
+    
+    # Frame numbers in this file should be 4, 5, 6
+    for i,h in zip(range(4,7,1), headers):
+        assert h["frameNumber"] == i
+
+    # Compare to canned data using numpy
+    data = np.load(test_data_path / "raw/jungfrau/jungfrau_single_0.npy")
+    assert np.all(data[3:6] == frames)
+
+@pytest.mark.files
+def test_iterate_over_a_jungfrau_RawSubFile(test_data_path):
+
+    data = np.load(test_data_path / "raw/jungfrau/jungfrau_single_0.npy")
+
+    with RawSubFile(test_data_path / "raw/jungfrau/jungfrau_single_d0_f0_0.raw", DetectorType.Jungfrau, 512, 1024, 16) as f:
+        i = 0
+        for header, frame in f:
+            assert header["frameNumber"] == i+1
+            assert np.all(frame == data[i])
+            i += 1
+        assert i == 3
+        assert header["frameNumber"] == 3

python/tests/test_jungfrau_dat_files.py

@@ -0,0 +1,92 @@
+import pytest
+import numpy as np
+from aare import JungfrauDataFile
+
+@pytest.mark.files
+def test_jfungfrau_dat_read_number_of_frames(test_data_path):
+    with JungfrauDataFile(test_data_path / "dat/AldoJF500k_000000.dat") as dat_file:
+        assert dat_file.total_frames == 24
+
+    with JungfrauDataFile(test_data_path / "dat/AldoJF250k_000000.dat") as dat_file:
+        assert dat_file.total_frames == 53
+    
+    with JungfrauDataFile(test_data_path / "dat/AldoJF65k_000000.dat") as dat_file:
+        assert dat_file.total_frames == 113
+
+
+@pytest.mark.files
+def test_jfungfrau_dat_read_number_of_file(test_data_path):
+    with JungfrauDataFile(test_data_path / "dat/AldoJF500k_000000.dat") as dat_file:
+        assert dat_file.n_files == 4
+
+    with JungfrauDataFile(test_data_path / "dat/AldoJF250k_000000.dat") as dat_file:
+        assert dat_file.n_files == 7
+    
+    with JungfrauDataFile(test_data_path / "dat/AldoJF65k_000000.dat") as dat_file:
+        assert dat_file.n_files == 7
+
+
+@pytest.mark.files
+def test_read_module(test_data_path):
+    """
+    Read all frames from the series of .dat files. Compare to canned data in npz format. 
+    """
+
+    # Read all frames from the .dat file
+    with JungfrauDataFile(test_data_path / "dat/AldoJF500k_000000.dat") as f:
+        header, data = f.read()
+
+    #Sanity check
+    n_frames = 24
+    assert header.size == n_frames
+    assert data.shape == (n_frames, 512, 1024)
+
+    # Read reference data using numpy
+    with np.load(test_data_path / "dat/AldoJF500k.npz") as f:
+        ref_header = f["headers"]
+        ref_data = f["frames"]
+
+    # Check that the data is the same
+    assert np.all(ref_header == header)
+    assert np.all(ref_data == data)
+
+@pytest.mark.files
+def test_read_half_module(test_data_path):
+
+    # Read all frames from the .dat file
+    with JungfrauDataFile(test_data_path / "dat/AldoJF250k_000000.dat") as f:
+        header, data = f.read()
+
+    n_frames = 53
+    assert header.size == n_frames
+    assert data.shape == (n_frames, 256, 1024)
+
+    # Read reference data using numpy
+    with np.load(test_data_path / "dat/AldoJF250k.npz") as f:
+        ref_header = f["headers"]
+        ref_data = f["frames"]
+
+    # Check that the data is the same
+    assert np.all(ref_header == header)
+    assert np.all(ref_data == data)
+
+
+@pytest.mark.files
+def test_read_single_chip(test_data_path):
+
+    # Read all frames from the .dat file
+    with JungfrauDataFile(test_data_path / "dat/AldoJF65k_000000.dat") as f:
+        header, data = f.read()
+
+    n_frames = 113
+    assert header.size == n_frames
+    assert data.shape == (n_frames, 256, 256)
+
+    # Read reference data using numpy
+    with np.load(test_data_path / "dat/AldoJF65k.npz") as f:
+        ref_header = f["headers"]
+        ref_data = f["frames"]
+
+    # Check that the data is the same
+    assert np.all(ref_header == header)
+    assert np.all(ref_data == data)