* Add print function to collections

* Make the IReader interface capable of jumping to events

* Split python parts into its own shared lib

* Make it possible to read SIO files via python

* Add podio-dump tool to inspect podio files

Make sure print flushes the stream to ensure that python and c++ stdout
are "synchronized"

* Make podio-dump also dump event data

* PodioConfig: set ENABLE_SIO variable to avoid warning about cmake policy CMP0012 in downstream packages

* CMake: Explicitly require version 3 (or newer) for Catch2

* Use latest versions of actions and introduce log groups

* Update to LCG_102 in test lcg based environments

* Add workflow that builds EDM4hep after podio

* Make python code a proper module

* Reorganize CMake config for shared libraries

* Refactor library and dict generation in cmake config

* Add basic version of python bindings for working with Frames

* Lift an existing root and sio legacy frame reader from git

* Add python bindings for legacy readers

* Add dumping functionality for Frames

* Add a few basic tests for podio-dump

…IDASoft#351)

* Fix erroneously considering empty collections as not present

* Handle missing SIO backend more gracefully in podio-dump

…#354)

* Add /include to ROOT_INCLUDE_PATH

Necessary for podio-dump

* Fix README with the new path

* Use the choices from argparse for the handlers

* Initialize unique_ptr in initializer list for nvcc

* Skip tests for podio build in edm4hep workflows

Co-authored-by: Valentin Volkl <[email protected]>

* Add basic I/O tests for datamodel extensions

* Generate and build extension model earlier

* Add extension datamodel to gitignore

Co-authored-by: Valentin Volkl <[email protected]>

…Soft#363)

* Use defaultdict instead of essentially hand-rolling one

* Fix cmake configure dependencies again after AIDASoft#343

* Move SIO utilities to existing private utils header

* Use sio_utils also for legacy writer

* Do not trigger twice on PRs, add workflow_dispatch to run manually

* Add comment to run the workflows manually

Frame::Frame(), get(), put(), and putParameter() methods had (non-templated) definitions in the header file. These have now been specified as inline.

* Adding ID to the short output

* Raise errors when the corresponding library is not found

Co-authored-by: jmcarcell <[email protected]>

* Initial commit that works

* Add a legend and a few fixes

* Exchange blue and red

* Fix filename

* Fix pylint issues

* Fix linter issues and add some text when saving a file

* Address comments; add grouping and change default format to svg

* Add information in the README and delete an useless argument

* Fix linter issues

* Simplify code a bit since I wasn't able to align the legend

* Fix change in the README too

* Remove renderer

* Add comment in the README and check for graphviz being installed

* Exit if it doesn't exist

* Change exit() -> sys.exit()

* Remove trailing whitespace

…#380)

* Add more parameter functionality to Frame interface

* Mark getGenericParamtersForWrite as deprecated

Now that getParamters exists

* Mark inline defined methods as inline

* Add deprecation warning for EventStore based I/O

* Add deprecation warning for EventStore python bindings

* Make CI pass while deprecation warnings are in

)

* Add double map to GenericParamters

Add tests for constructors

* Make python bindings work for double parameters again

* Make sure to also print double values

* Add test case for float value retrieval in python

* Make sure to write double params in SIO, add legacy test case

* Make SIO reading backward compatible

* Fix typos and remove redundant test

…IDASoft#384)

* Split model reading into file reading and parsing

* Add possibility to dump parsed EDMs to JSON

* Add registry for datamodel JSON defintions

- Generate model definitions in JSON format as constexpr string literals
- Register constexpr string literals in registry

* Populate definition registry via static variable initialization

- Also make the collections query-able

* Write definition and read it back afterwards (ROOT)

* Read and write EDM definitions in SIO

* Refactor EDM defintion I/O functionality for less code duplication

* Add first version of model dumping

* Make dumped models look more like inputs

- Change order of main keys (options, components, datatypes)
- Slightly tweak formatting (as far as possible with PyYAML)

* Add roundtrip tests for stored EDM definitions

* Add documentation for EDM definition embedding

* Fix documentation

* Add warnings output when trying to retrieve non existant EDMs

* Rename EDMDefinitionRegistry and clarify documentation

* Fix python bindings and rename tests

* Make utility classes members instead of using them as mixin

* Rename header file to match new terminology

* Update documentation to match implementation again

* Update test names also in ignored tests list

Co-authored-by: jmcarcell <[email protected]>

AIDASoft#390)

* Modify initial clang check

* Test for the existence of .clang-format

* Fix linter

---------

Co-authored-by: jmcarcell <[email protected]>

Co-authored-by: jmcarcell <[email protected]>

* add stripped down schema evolution; relying on ROOT for the moment

---------

Co-authored-by: Thomas Madlener <[email protected]>

Keep them in place even if the clean target is specified. See AIDASoft#396

Co-authored-by: Paul Gessinger-Befurt <[email protected]>

…#399)

* Actually check extension model dumping

* Rework script and CMakeLists to pass in more information in
  environment

* Also check when clang format is used
---------

Co-authored-by: jmcarcell <[email protected]>
Co-authored-by: Thomas Madlener <[email protected]>

…rs and `size()` (AIDASoft#408)

* fix: use type `size_t` for index in `Collection` element accessors

* fix: `unsigned int` -> `size_t` in `Collection.cc.jinja2`

* fix: use `size_t` in `macro multi_relation_handling`

* fix: `size_t` -> `std::size_t`

* fix: add `#include <cstddef>` to `{,Mutable}Object.h.jinja2`

…#394)

* Add schema version to datamodel meta information

* Use meta namespace to get schema version info

* Make sure to have non-zero positive schema versions

* Remove unnecessary attributes

* Add basic buffer factory and switch over UserDataCollections

* Add documentation

* Fix dumping of UserDataCollection

* Add unittest to cover fix

…AIDASoft#415)

* Remove the deprecated getters and setters from the generic parameters

* Add some friends not to make public the getMap functions

…DASoft#416)

Co-authored-by: jmcarcell <[email protected]>

…t#418)

* Make string/float/int maps available again from GenericParameters

* Remove unnecessary typedefs

* Add documentation for opening of files API

* Fix seg fault when reading from multiple files

* Add tests and properly fix file switching

* Ignore new test in sanitizer runs

…ft#420)

…DASoft#422)

* Introduce version subdirectory for legacy tests

* Rename executable and add more legacy versions for testing

* Introduce macro to cleanup CMakeLists slightly

* Rename frame reading test

* Add frame reading tests for legacy files

* Don't check build versions in legacy files

* Fix reading of v00-16 files and adapt tests to expected contents

* Add legacy file reading test to ignore list for sanitizer builds

* Make sure to properly configure CTestCustom.cmake

…Soft#405)

* Add RelationNames and code gen to populate it

* Write branches with proper names in ROOTFrameWriter

* Make the Frame reader use the registry to get branch names

* Make sure that the EDM always is in the registry

* Make sure that the names are populated when they are used

* Make relation branch names start with an underscore

Less prone to accidental collisions. Plus sort of conveying that they
are "hidden"

* Make it possible to read index based files as well

- Keep branch names around for resetting them (since invalidated
branches can no longer provide that information)

* Fix legacy reading

* Manually increase patch to make if clause work properly

…#425)

* Make unittest environment slightly more robust

* Make catch test discovery skippable

* Specifically use python3 for python unittests

…t#412)

* add hashing feature to CollectionID table

* move collectionID from mix of int/unsigned to consistent uint32_t

* fix collectionID in SIOBlock

* use 32 bit hash; use murmurhash3

* Ignore tests in UB sanitizer runs

* Add standalone executable for collision detection

* protect frame against double insert

---------

Co-authored-by: tmadlener <[email protected]>

…t#402)

* Add static constexpr char* names and hand out views for them

* Work around Catch2 problem in test discovery

* Add AUTO as possible option to USE_EXTERNAL_CATCH2

* Switch key4hep workflows to build Catch2 if necessary

* Delay library loading as long as possible

* Add a --version flag for dumping the podio version

* Dump collections in alphabetical order

* Use tabulate for easier printing of overview table

* Upate README with new requirements

* [format] Fix pylint and flake8 issues

* Fix repeated download of test inputs if they are alrady present

* Split tests into subdirectories to declutter things a bit

* Move test utility functionality into dedicated file

- Introduce PODIO_SET_TEST_ENV function to set consistent test environment

* Move dumpmodel roundtrip tests into separate subfolder

* Move commonly used declaration to separate header (Fixes clang-tidy not being able to find an included header otherwise)

…ASoft#434)

* Download more legacy file versions

Rework download logic to have one source of legacy versions

* Add tests that read legacy files through the legacy Frame reader

* Make ROOTLegacyReader handle older versions correctly

* Remove unused fields

* Add SchemaEvolution singleton to hold evolution functions

* Inject type information into collection buffers

* Inject current schema version into buffers from buffer factory

* [wip] Start populating SchemaEvolution registry

* [wip] Split registration into two steps

Easier this way to get the current version information into the whole
system

* [wip] Require registration of each evolution function

Remove two step registration again

* [clang-tidy] Mark inputs as const& for now

* Remove registry from docstrings

* Add a RNTuple writer

* Cleanup and add a reader

* Add compilation instructions for RNTuple

* Add tests

* Fix the reader and writer so that they pass most of the tests

* Commit missing changes in the header

* Add support for Generic Parameters

* Add an ugly workaround to the unique_ptr issue

* Read also vector members and remove some comments

* Do a bit of cleanup

* Do more cleanup, also compiler warnings

* Add names in rootUtils.h, fix a few compiler warnings

* Add a few minor changes

* Add missing changes in the headers

* Change map -> unordered_map and use append in CMakeLists.txt

* Simplify writing and reading of generic parameters

* Only create the ID table once

* Add CollectionInfo structs

* Add a ROOT version check

* Add missing endif()

* Add Name at the end of some names

* Add missing Name at the end

* Cast to rvalue

* Cache entries and reserve

* Add comment and remove old comments

* Remove a few couts

* Remove intermediate variables and use std::move

* Run clang-format

* Use clang-format on tests too

* Enable RNTuple I/O in Key4hep CI

* Check if dev3 workflows come with recent enough ROOT

* Change MakeField to the new signature

* Update the RNTuple reader and writer to use the buffer factory

* Run clang-format

* Update the RNTuple writer to use a bare model

* Add friends for Generic Parameters

* Update changes after the changes in the collectionID and string_view

* Run clang-format

* Update the reader and writer to conform to AIDASoft#405

* Reorganize and clean up code in the reader

* Run clang-format

* Simplify how the references are filled

---------

Co-authored-by: jmcarcell <[email protected]>
Co-authored-by: tmadlener <[email protected]>

Co-authored-by: jmcarcell <[email protected]>

* Switch to a newer LCG stack with root 6.28/00

* Try with key4hep nightlies

* Allow git to run necessary commands for pre-commit

* Fix clang-format complaints

* Fix clang-tidy complaints

* Reorder some imports to make pylint happy again

* Enbale building RNTuple support in more workflows

Co-authored-by: jmcarcell <[email protected]>

Co-authored-by: jmcarcell <[email protected]>