* bob/bob.buildout
  * minor
     - bob/bob.buildout!32 New CI using bdt
     - bob/bob.buildout!33 Try to use cp instead of install for doc-dir installation:
     - bob/bob.buildout!35 Removing MACOSX_DEPLOYMENT_TARGET variable
* bob/bob.extension
  * patch
     - bob/bob.extension!89 Improve the error message of click tests by printing the traceback
     - bob/bob.extension!91 Fix for CONDA_BUILD_SYSROOT variable missing on mac at test time
     - bob/bob.extension!75 Fix a bug when commands where invoked multiple times
     - bob/bob.extension!95 Improve support for Extension building by acknowledging CXXFLAGS
     - bob/bob.extension!90 Disable sphinx nitpick and Python 2
     - bob/bob.extension!96 [click] Account for cases when exception is None in click results
     - bob/bob.extension!97 [ci] Update gitlab-ci file (depend on bdt): No description for this MR
     - bob/bob.extension!98 Update for click 7 upgrade
     - bob/bob.extension!99 Augments the extension for building C/C++ bindings with tests for SDKROOT and MACOSX_DEPLOYMENT_TARGET environment variables.
     - bob/bob.extension!100 Document setup of CONDA_BUILD_SYSROOT: This MR properly documents how to setup one's environment to compile on macOS machines.
* bob/bob.blitz
  * patch
* bob/bob.core
  * patch
* bob/bob.io.base
  * patch
     - bob/bob.io.base!27 Add a function to load features in a memory efficient way
* bob/bob.math
  * patch
* bob/bob.measure
  * minor
     - bob/bob.measure!89 Add a base class for subplot-based plots
     - bob/bob.measure!90 Fixing histogram test
     - bob/bob.measure!91 Fix decimal option in metrics command: Fixes bob/bob.measure#52
     - bob/bob.measure!43 Resolve "FAR and FRR thresholds are computed even when there is no data support"
     - bob/bob.measure!92 Fix minimizing threshold: Closes bob/bob.measure#59
     - bob/bob.measure!93 Click scripts fixes: Improvements and bug fixes to click scripts
     - bob/bob.measure!94 Fix --figsize help message
* bob/bob.io.image
  * patch
     - bob/bob.io.image!24 Resolve "GIF loading seems to be completely broken"
     - bob/bob.io.image!42 Fix c++ exception throw
     - bob/bob.io.image!43 Fix readout of netpbm images containing empty lines (such as those produced by GIMP)
* bob/bob.db.base
  * patch
* bob/bob.io.video
  * major
     - bob/bob.io.video!10 Fixes to multiple issues: This MR closes bob/bob.io.video#6, closes bob/bob.io.video#8 and closes bob/bob.io.video#9.
     - bob/bob.io.video!13 Frame free fix: The API to free the frames have changed in ffmpeg >= 3.0    Closes bob/bob.io.video#12.
     - bob/bob.io.video!15 Adds a delete method to reader.iter types
     - bob/bob.io.video!16 Increasing the threshold in the test_memory_leak
     - bob/bob.io.video!17 Port to ffmpeg 4, disable (libvpx|vp8 and mov) tests: libvpx and vp8 codecs are not working with .mov containers
* bob/bob.io.matlab
  * patch
* bob/bob.io.audio
  * patch
* bob/bob.sp
  * patch
* bob/bob.ap
  * patch
* bob/bob.ip.base
  * patch
* bob/bob.ip.color
  * patch
* bob/bob.ip.draw
  * patch
* bob/bob.ip.gabor
  * patch
* bob/bob.learn.activation
  * patch
* bob/bob.learn.libsvm
  * patch
* bob/bob.learn.linear
  * patch
* bob/bob.learn.mlp
  * patch
* bob/bob.learn.boosting
  * patch
     - bob/bob.learn.boosting!10 Do not build a library to avoid C++ API incompatibilities (see bob/bob.learn.boosting#2)
* bob/bob.db.iris
  * patch
* bob/bob.learn.em
  * patch
     - bob/bob.learn.em!16 implemented EM with multiprocessing
* bob/bob.db.wine
  * patch
* bob/bob.db.mnist
  * patch
* bob/bob.db.atnt
  * patch
* bob/bob.ip.facedetect
  * patch
* bob/bob.ip.optflow.hornschunck
  * patch
* bob/bob.ip.optflow.liu
  * patch
* bob/bob.ip.flandmark
  * patch
* bob/gridtk
  * patch
     - bob/gridtk!24 Fix submission to gpu queues removing memfree and hvmem settings (causing issues with virtual memory allocation)
     - bob/gridtk!25 Do not delete the logs of the jobs that are not resubmitted: Fixes bob/gridtk#29
     - bob/gridtk!26 Do not automatically remove G from the GPU memory requirements: Fixes bob/gridtk#30
     - bob/gridtk!27 Automatic injection of bob.extension.rc to jinja variables: Fixes bob/gridtk#31
     - bob/gridtk!28 Add a helper function to easily slice into a list in array jobs
* bob/bob.ip.qualitymeasure
  * patch
     - bob/bob.ip.qualitymeasure!13 added 2 lines for pkg_config: Updated meta.yaml to include pkg_config.
* bob/bob.ip.skincolorfilter
  * patch
* bob/bob.ip.facelandmarks
  * patch
* bob/bob.ip.dlib
  * patch
     - bob/bob.ip.dlib!14 Fix left eye right eye landmarks
* bob/bob.ip.mtcnn
  * patch
     - bob/bob.ip.mtcnn!7 Fixes conda recipe by setting caffe dependence to work on "linux" only
* bob/bob.db.arface
  * patch
* bob/bob.db.asvspoof
  * patch
* bob/bob.db.asvspoof2017
  * patch
* bob/bob.db.atvskeystroke
  * patch
* bob/bob.db.avspoof
  * patch
* bob/bob.db.banca
  * patch
* bob/bob.db.biosecure
  * patch
* bob/bob.db.biosecurid.face
  * patch
* bob/bob.db.casia_fasd
  * patch
* bob/bob.db.casme2
  * patch
* bob/bob.db.caspeal
  * patch
* bob/bob.db.cohface
  * patch
* bob/bob.db.frgc
  * patch
* bob/bob.db.gbu
  * patch
* bob/bob.db.hci_tagging
  * patch
* bob/bob.db.ijbc
  * patch
* bob/bob.db.kboc16
  * patch
* bob/bob.db.lfw
  * patch
* bob/bob.db.livdet2013
  * patch
* bob/bob.db.mobio
  * patch
* bob/bob.db.msu_mfsd_mod
  * patch
* bob/bob.db.multipie
  * patch
* bob/bob.db.nist_sre12
  * patch
* bob/bob.db.putvein
  * patch
* bob/bob.db.replay
  * patch
     - bob/bob.db.replay!6 Use the SQL class from bob.db.base to benefit from improvements
* bob/bob.db.replaymobile
  * patch
     - bob/bob.db.replaymobile!11 Improve db interface
* bob/bob.db.scface
  * patch
* bob/bob.db.utfvp
  * patch
* bob/bob.db.verafinger
  * patch
* bob/bob.db.fv3d
  * patch
* bob/bob.db.voicepa
  * patch
* bob/bob.db.xm2vts
  * patch
* bob/bob.db.youtube
  * patch
* bob/bob.db.pericrosseye
  * patch
* bob/bob.db.maskattack
  * patch
* bob/bob.db.casiasurf
  * major
* bob/bob.bio.base
  * patch
     - bob/bob.bio.base!171 Implemented multiple subjects score generation, unknown probes and fixed issues:
     - bob/bob.bio.base!172 Fix the tests after legend loc option change in bob.measure!89
     - bob/bob.bio.base!175 small fix in help message
     - bob/bob.bio.base!174 Adds a sort command and small fixes
* bob/bob.bio.gmm
  * patch
     - bob/bob.bio.gmm!22 Adds a threading option to the GMM algorithm
* bob/bob.bio.face
  * patch
     - bob/bob.bio.face!53 Do not allow upside down faces in FaceCrop by default: Fixes bob/bob.bio.face#33
     - bob/bob.bio.face!54 Remove bob.ip.dlib annotator: bob.ip.dlib is no longer maintained.
     - bob/bob.bio.face!52 Add Fargo database: New face recognition database.
     - bob/bob.bio.face!55 added FARGO stuff wherever it was needed:
     - bob/bob.bio.face!44 Fix the header level of index.rst
     - bob/bob.bio.face!57 Improve the warning message in FaceCrop
     - bob/bob.bio.face!56 Improve Replay Mobile interface
* bob/bob.bio.spear
  * patch
     - bob/bob.bio.spear!45 Fix import
* bob/bob.bio.video
  * patch
     - bob/bob.bio.video!37 Keep the order of FrameContainers when saving:
     - bob/bob.bio.video!39 Improve logging in case of failures in preprocessing
     - bob/bob.bio.video!38 Add replay mobile interface     
* bob/bob.bio.vein
  * patch
* bob/bob.bio.face_ongoing
  * patch
* bob/bob.bio.htface
  * patch
* bob/bob.db.voxforge
  * patch
* bob/bob.db.ijba
  * patch
* bob/bob.rppg.base
  * patch
* bob/bob.pad.base
  * minor
     - bob/bob.pad.base!54 Use the GridSubplot base class for epsc: Depends on bob.measure!89
     - bob/bob.pad.base!55 Use the version of scikit-learn that we use for testing:
     - bob/bob.pad.base!56 Use the builtin method for converting frame containers to arrays: 
     - bob/bob.pad.base!57 Correcting bug when using split_training_data_by_client flag in extractor
     - bob/bob.pad.base!53 Cross database testing evaluation: Adds a new command `bob pad cross` to evaluate PAD systems in cross database settings
     - bob/bob.pad.base!58 Fixes bob/bob.pad.base#8 Algorithm.read_toscore_object should not exist
     - bob/bob.pad.base!59 fix for wrong Axis label with -nG option: fix for wrong Axis label with -nG option, fixes bob/bob.pad.base#28
     - bob/bob.pad.base!60 Correct apcer calculation
     - bob/bob.pad.base!61 Bpcer threshold criteria: Added criteria for threshold selection based on desired BPCER criteria
     - bob/bob.pad.base!62 Add a new one class GMM based on bob's GMMs
* bob/bob.pad.face
  * minor
     - bob/bob.pad.face!73 Fix the broken if logic in ImageQualityMeasure
     - bob/bob.pad.face!74 Updated BATL HLDI: added rc, skin/non-skin annotations, multi-channel laoding functionalities
     - bob/bob.pad.face!75 Alignment for LightCNN: Added an alignment method, suitable for LightCNN
     - bob/bob.pad.face!76 remove missing bootstrap-buildout.py from MANIFEST.in
     - bob/bob.pad.face!79 Batl hldi update: Updted batl HLDI, and skin, non-skin annotation files.
     - bob/bob.pad.face!81 MC preprocessor: The pre-processor for multi-channel data, allowing to detect and crop faces, and extract patches from cropped images.
     - bob/bob.pad.face!80 Patch extraction from images
     - bob/bob.pad.face!83 Preprocessing and quality check
     - bob/bob.pad.face!77 Added CASIA-SURF database
     - bob/bob.pad.face!86 Batl loo protocols: Adds Leave one out protocols to the BATL HLDI.
     - bob/bob.pad.face!87 Chalearn pad: 1. Fixed small mistake in CASIA-SURF High level PAD BD  2. Added some stuff to a preprocessor (sanity check and logging)
     - bob/bob.pad.face!88 Added the doc on MLP training using AE embeddings
     - bob/bob.pad.face!89 Fixed CASIA SURF
     - bob/bob.pad.face!90 [database] added modification to high-level interface for CASIA-SURF
     - bob/bob.pad.face!91 Resolve "fix CASIA-SURF tests"
     - bob/bob.pad.face!95 Move bob.ip.dlib and bob.ip.facelandmarks to test dependencies
     - bob/bob.pad.face!96 Put bob.ip packages back as run requirements
     - bob/bob.pad.face!97 unit tests: Reverts the unit tests
     - bob/bob.pad.face!98 Add casia fasd and an optical flow extractor
* bob/bob.pad.voice
  * patch
     - bob/bob.pad.voice!17 Stop the TensorflowEval algorithm to load on imports
* bob/bob.pad.vein
  * patch
* bob/bob.fusion.base
  * patch
     - bob/bob.fusion.base!10 Add Tanh and ZNorm preprocessors: and clean-up code.
* bob/bob.db.oulunpu
  * patch
      - bob/bob.db.oulunpu!5 Fix left and right annotations being swapped
* bob/bob.db.uvad
  * patch
* bob/bob.db.cuhk_cufs
  * patch
* bob/bob.db.cbsr_nir_vis_2
  * patch
* bob/bob.db.nivl
  * patch
* bob/bob.db.pola_thermal
  * patch
* bob/bob.db.cuhk_cufsf
  * patch
* bob/bob.db.ldhf
  * patch
* bob/bob.ip.tensorflow_extractor
  * patch
     - bob/bob.ip.tensorflow_extractor!9 Removed plot since this is not used
     - bob/bob.ip.tensorflow_extractor!10 [facenet] Allow layer name
     - bob/bob.ip.tensorflow_extractor!11 Created a VGGFace for tensorflow: Added download link
     - bob/bob.ip.tensorflow_extractor!12 Add MTCNN (tensorflow) model: code and model from https://github.com/blaueck/tf-mtcnn
     - bob/bob.ip.tensorflow_extractor!13 Improve graph and session handling in facenet class
* bob/bob.ip.caffe_extractor
  * patch
     - bob/bob.ip.caffe_extractor!11 Fix conda recipe by setting caffe dependence to work on "linux" only
* bob/bob.bio.caffe_face
  * patch
* bob/bob.learn.tensorflow
  * minor
     - bob/bob.learn.tensorflow!72 [predict_bio] Some fixes and new features:
     - [predict_bio] fix the broken checkpoint path handling 
     - [predict_bio] Add an option to save data in FrameContainers 
     - [simplecnn] fix a bug with slim architecture
     - [regressor] add a name scope for train parts
     - [test] organize estimator scripts tests
     - [eval] improvements to bob tf eval
     - bob/bob.learn.tensorflow!73 Improvements to eval script: 
     - bob/bob.learn.tensorflow!74 Click 7 compatibility
     - bob/bob.learn.tensorflow!75 Add scripts for training keras models using keras API
     - Add two functions to allow dynamic weighting of samples per batch
     - Add several extractors which are usefull at inference time 
     - Add a generic Generator class alternative to BioGenerator
     - Add virtual adversarial training loss
     - Fix the euclidean function so that its gradients don't become nan. Also moves the bytes_to_human function
     - Add a hook to add some tensors to the summary
     - Add bob tf predict command  
     - Fix the image augmentation function 
     - Add dataset_to_tfrecord and dataset_from_tfrecord 
     - Make bob tf cache command useful 
     - Add a new MLP architecture  
     - Fixes in inception architectures
     - bob/bob.learn.tensorflow!77 Fixed the reuse in some operations of the inception-v1: 
     - bob/bob.learn.tensorflow!78 Fix issue with VGG16 from slim. The slim models adds the hot-encoded in the architecture function
     - bob/bob.learn.tensorflow!76 Estimators optimize loss:
* bob/bob.db.fargo
  * major
     - bob/bob.db.fargo!2 Disable Sphinx nitpicky and disable Python 2
     - bob/bob.db.fargo!3 Resolve "Make it sqlite":
     - bob/bob.db.fargo!4 Cleaning the package
