Jungfraujoch/reader/JFJochHDF5Reader.cpp

// SPDX-FileCopyrightText: 2024 Filip Leonarski, Paul Scherrer Institute <filip.leonarski@psi.ch>
// SPDX-License-Identifier: GPL-3.0-only

#include "JFJochHDF5Reader.h"
#include "../common/JFJochException.h"

void JFJochHDF5Reader::ReadFile(const std::string &filename) {
    std::unique_lock ul(hdf5_mutex);
    image_source_.Clear();
    snapshots_.clear();
    active_metadata_.reset();
    active_snapshot_.clear();
    number_of_images = 0;
    try {
        auto metadata = std::make_shared<HDF5MetadataSource>();
        auto open_result = metadata->Open(filename, default_experiment);

        image_source_.Configure(std::move(open_result.image_layout));
        // Original file: per-image metadata is co-located with the pixels.
        metadata->UseImageSourceForMetadata(&image_source_);
        number_of_images = open_result.number_of_images;

        snapshots_["Original"] = metadata;
        active_metadata_ = metadata;
        active_snapshot_ = "Original";
        SetStartMessage(metadata->Dataset());
    } catch (const std::exception &e) {
        image_source_.Clear();
        snapshots_.clear();
        active_metadata_.reset();
        active_snapshot_.clear();
        number_of_images = 0;
        SetStartMessage({});
        throw;
    }
}

uint64_t JFJochHDF5Reader::GetNumberOfImages() const {
    std::unique_lock ul(hdf5_mutex);
    return number_of_images;
}

void JFJochHDF5Reader::Close() {
    std::unique_lock ul(hdf5_mutex);
    image_source_.Clear();
    snapshots_.clear();
    active_metadata_.reset();
    active_snapshot_.clear();
    number_of_images = 0;
    SetStartMessage({});
}

HDF5ImageLocator::Location JFJochHDF5Reader::GetImageLocation(int64_t image_number) const {
    if (image_number >= static_cast<int64_t>(number_of_images) || image_number < 0)
        throw JFJochException(JFJochExceptionCategory::HDF5, "Image out of bounds");
    return image_source_.Resolve(image_number);
}

std::shared_ptr<JFJochReaderRawImage> JFJochHDF5Reader::GetRawImage(int64_t image_number) {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Cannot load image if file not loaded");

    auto loc = GetImageLocation(image_number);
    auto ret = std::make_shared<JFJochReaderRawImage>();
    ret->image = image_source_.ReadImageAt(ret->image_buffer, loc);
    return ret;
}

bool JFJochHDF5Reader::LoadImage_i(std::shared_ptr<JFJochReaderDataset> &dataset,
                                   DataMessage &message,
                                   std::vector<uint8_t> &buffer,
                                   int64_t image_number,
                                   bool update_dataset) {
    std::unique_lock ul(hdf5_mutex);
    (void) update_dataset;

    if (!dataset)
        return false;
    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Cannot load image if file not loaded");

    // Pixels from the shared image source, per-image metadata from the active snapshot.
    auto loc = GetImageLocation(image_number);
    message.image = image_source_.ReadImageAt(buffer, loc);
    message.number = image_number;
    active_metadata_->FillPerImage(message, image_number, dataset);
    return true;
}

std::vector<HDF5DataSourceMessage> JFJochHDF5Reader::GetHDF5DataSource(uint64_t first_image,
                                                                       std::optional<uint64_t> image_count) const {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Cannot generate HDF5 source mapping if file not loaded");

    return image_source_.GetSourceMapping(first_image, image_count, number_of_images);
}

std::vector<IntegrationOutcome> JFJochHDF5Reader::ReadReflections(size_t start_image,
                                                                  std::optional<size_t> end_image) const {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Cannot read reflections if file not loaded");

    return active_metadata_->ReadReflections(start_image, end_image);
}

std::vector<SpotToSave> JFJochHDF5Reader::ReadSpots(int64_t image) const {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Cannot read spots if file not loaded");

    return active_metadata_->ReadSpots(image);
}

bool JFJochHDF5Reader::HasStoredSpots() const {
    std::unique_lock ul(hdf5_mutex);
    return active_metadata_ && active_metadata_->HasSpots();
}

CompressedImage JFJochHDF5Reader::ReadCalibration(std::vector<uint8_t> &tmp, const std::string &name) const {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid, "Master file not loaded");

    return active_metadata_->ReadCalibration(tmp, name);
}

void JFJochHDF5Reader::RegisterSnapshot(const std::string &name, const std::string &master_path) {
    std::unique_lock ul(hdf5_mutex);

    if (!active_metadata_)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Open a dataset before registering a snapshot");

    auto metadata = std::make_shared<HDF5MetadataSource>();
    auto open_result = metadata->Open(master_path, default_experiment);

    // A snapshot may cover a subset of the images (a sub-range or filtered reprocessing); its
    // /entry/detector/number map (read in Open) ties each snapshot image back to an original one.
    // It just must not claim more images than the dataset has.
    if (open_result.number_of_images > number_of_images)
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid,
                              "Snapshot has more images than the open dataset");

    // Snapshot pixels come from the existing image source; its own (integrated) master holds the
    // per-image metadata at the global index.
    metadata->UseImageSourceForMetadata(nullptr);
    snapshots_[name] = metadata;
}

void JFJochHDF5Reader::RemoveSnapshot(const std::string &name) {
    std::unique_lock ul(hdf5_mutex);
    if (name == "Original")
        return;   // the original file metadata is always kept
    snapshots_.erase(name);
    if (active_snapshot_ == name) {
        active_metadata_ = snapshots_.at("Original");
        active_snapshot_ = "Original";
        SetStartMessage(active_metadata_->Dataset());
    }
}

void JFJochHDF5Reader::SetActiveSnapshot(const std::string &name) {
    std::unique_lock ul(hdf5_mutex);

    auto it = snapshots_.find(name);
    if (it == snapshots_.end())
        throw JFJochException(JFJochExceptionCategory::InputParameterInvalid, "Unknown snapshot: " + name);

    active_metadata_ = it->second;
    active_snapshot_ = name;
    SetStartMessage(active_metadata_->Dataset());
}

std::vector<std::string> JFJochHDF5Reader::SnapshotNames() const {
    std::unique_lock ul(hdf5_mutex);

    std::vector<std::string> names;
    names.reserve(snapshots_.size());
    for (const auto &[name, _]: snapshots_)
        names.push_back(name);
    return names;
}

std::string JFJochHDF5Reader::ActiveSnapshot() const {
    std::unique_lock ul(hdf5_mutex);
    return active_snapshot_;
}

std::vector<std::pair<std::string, std::shared_ptr<const JFJochReaderDataset>>>
JFJochHDF5Reader::AllSnapshotDatasets() const {
    std::unique_lock ul(hdf5_mutex);
    std::vector<std::pair<std::string, std::shared_ptr<const JFJochReaderDataset>>> out;
    out.reserve(snapshots_.size());
    // "Original" first, then the rest, so overlay colours stay stable across updates.
    if (auto it = snapshots_.find("Original"); it != snapshots_.end())
        out.emplace_back(it->first, it->second->Dataset());
    for (const auto &[name, source]: snapshots_)
        if (name != "Original")
            out.emplace_back(name, source->Dataset());
    return out;
}