- DLT coefficient to/from intrinsic and extrinsic camera parameters (pinhole model)
- Improve modeling of 3d object, camera, mirror on a gimbal, and virtual camera
- test rotation matrices and pinhole camera model

+ Changed convention for normal vector of non-rotated object (from [1, 0, 0] to [0, 0, 1])
+ Use extrinsic rotation 'xyz' instead of intrinsic 'ZYX' for scipy.transform.Rotation
+ Implemented quaternions for calibration, Object3D, Camer, etc. But still need to implement them in skeleton_fitter
+ Improve documentation (doctstrings)

+ Other minor changes

Moved VirtualCamera(s) classes in their own .py + Various changes to VirtualCamera to prevent problems (e.g. it's not possible to set X,Y,Z and alpha,beta,gamma)

+Small changes to README.md

Worked on camera classes to handle 3d object(s), (virtual and real) camera(s) and mirror(s). (Work in progress)
+ Moved all virtual camera classes to cameras.py

Added functions to compute dlt coefficients from and to intrinsinc/extrincic parameters. (Work in progress)
Added camera classes to handle 3d object(s), (virtual and real) camera(s) and mirror(s). (Work in progress)
+ Moved example calibration files
+ Minor changes in README.md

All settings can now be read from a unique yaml file that contains settings for the batch processing and the protocol (e.g. all process steps)

- Updated import calib in reconstructor3d.py
- Updated dlc_cfg_path to '/home/acribellier/Desktop/Antoine/_maDLC/config.yaml'
- Corrected docstrings

- Updated config.yaml
- Updated dlc_cfg_path to '../data/examples/mosquito_escapes/config.yaml'
- Corrected function name (convert_kwargs_to_be_writen_in_yaml > convert_kwargs_to_yaml)

# Conflicts:
#	point_tracker/reconstructor3d.py
#	process/batch_processing.py
#	process/dlc_postprocessing.py
#	process/utils.py
#	skeleton_fitter/animals/fly.py
#	skeleton_fitter/animals/fly_geo.py
#	skeleton_fitter/animals/fly_hybrid.py
#	skeleton_fitter/modules/bodies/insect_body_slim.py
#	skeleton_fitter/modules/limbs/insect_wings_flat.py
#	skeleton_fitter/skeleton_modules/insect_wings_flat_geo.py

Added a new utils package. It contains scripts to print header with the versions and settings. And a new framework for writing settings dataclasses. This framework is now used everywhere. For example: process_dict has been replaced by a settings class

Now check if frame that have been recorded by some camera(s) are missing on the other(s), if so raise an error
Does not save the missing frames in the process_dict anymore

- Added .yaml to run test scripts of multi_process_batch
- Added the possibility to crop using 2d coordinates (not from 3d tracks)
- Changed default name of sequence in mutli_image_sequence to be seq1, seq2, seq3, ...
- Corrected rec_name_to_process in test functions to be 20200216_075901-sample_binned instead of 20200216_075901-selection
- Changed names of args for deeplabcut.analyse()

- Made a BodyModuleSettings for insect_hybrid_body_slim.py
- Modified insect_wings_flat_geo.py to be used as a normal limb (use new numbered labels around each original label)
- Removed and added TODOs
- Small bug fixes

- Corrected dynamic imports
- Corrected file paths and names
- Added bounds for default_pitch_a
- Added a param_names_to_optimize (list with only the names of the parameters to optimize)
- Changed variable names
- Made minor docstring changes

Made two classes inherite from BodyModuleSettings, LimbsModuleSettings (to avoid having twice the same code)

Made csv_path dynamic, if updated, then param_names, labels, and skeleton3d_init will be updated automatically

Added a new utils package. It contains scripts to print header with the versions and settings. And a new framework for writing settings dataclasses. This framework is now used everywhere. For example: process_dict has been replaced by a settings class

Now check if frame that have been recorded by some camera(s) are missing on the other(s), if so raise an error
Does not save the missing frames in the process_dict anymore