Adding changelog for v0.2.0.

docker/package/Dockerfile

@@ -25,18 +25,6 @@ RUN cd leveldb && mkdir -p build && cd build && cmake -DCMAKE_BUILD_TYPE=Release
 RUN cd leveldb && cd build && make -j $(nproc)
 RUN cd leveldb && cd build && make install
 
-#RUN apt-get install -y protobuf-compiler libprotobuf-dev
-#RUN apt-get install -y libgrpc++-dev libgrpc++1 libgrpc-dev libgrpc3 
-#RUN apt-get install -y protobuf-compiler-grpc
-#RUN apt-get install -y libleveldb-dev libleveldb1v5 python3-leveldb vim
-
-#RUN apt-get install -y gcc-8 g++-8
-
-#RUN update-alternatives --install /usr/bin/gcc gcc /usr/bin/gcc-8 800
-#RUN update-alternatives --install /usr/bin/g++ g++ /usr/bin/g++-8 800
-#RUN update-alternatives --set gcc /usr/bin/gcc-8
-#RUN update-alternatives --set g++ /usr/bin/g++-8
-
 # PyTorch
 ARG PYTHON_VERSION=3.6
 RUN curl -o ~/miniconda.sh -O  https://repo.continuum.io/miniconda/Miniconda3-latest-Linux-x86_64.sh  && \

docs/source/changelog.rst

@@ -4,8 +4,29 @@ Changelog for the EuclidesDB releases.
 
 Release v.0.2.0
 -------------------------------------------------------------------------------
-    - Bug-fix and feature addition release;
-    - TODO
+This is a bug-fix and feature addition release with many good news ! The main new features are: integration with Faiss (see :ref:`search-config` for more information), new models, database compression, new exact linear search and internal codebase refactoring.
+
+Thanks for all the users that opened issues and contributors who helped with this release.
+
+Changes in this release:
+
+	- **[Enhancement]**: examples doesn't require ``torchvision`` anymore (`#8 <https://github.com/perone/euclidesdb/pull/8>`_);
+	- **[Bug]**: wrong model name in client call can cause the server to quit (`#1 <https://github.com/perone/euclidesdb/issues/1>`_);
+	- **[Enhancement]**: major refactoring of indexing types, they're now called **Search Engines** and have their own units and configuration;
+	- **[Bug]**: search engines were called with Variables instead of Tensors;
+	- **[Enhancement]**: added the new search engine called ``exact_dist`` that will do a on-disk search (as opposed to in-memory search) using linear exact search (see :ref:`search-config` for more information);
+	- **[Enhancement]**: each search engine has now their own requirement for refresh the index upon adding new items or not;
+	- **[Enhancement]**: added the new search engine called ``faiss`` that integrated Faiss/OpenMP/Blas together with EuclidesDB, any Faiss index type is now supported on EuclidesDB (see :ref:`search-config` for more information);
+	- **[Enhancement]**: to avoid memory allocations and improve performance, the reply vectors are now pre-allocated with top-k size;
+	- **[Enhancement]**: enabled database compression support (snappy);
+	- **[Enhancement]**: added Resnet101 model support;
+	- **[Enhancement]**: added internal database versioning machanism to support future underlying changes;
+	- **[Bug]**: fixed an issue with Python API (missing ``close()`` channel call);
+	- **[Enhancement]**: ``FindSimilar`` RPC call is now called ``FindSimilarImage``;
+	- **[Enhancement]**: added a new RPC call called ``FindSimilarImageById`` to search similar items based on items already indexed;
+	- **[Enhancement]**: added documentation for each Search Engine and their configurations (see :ref:`search-config` for more information);
+	- **[Enhancement]**: added documentation for each low-level gRPC call for advanced users (see :ref:`grpc-api` for more information);
+
 
 Release v.0.1.1
 -------------------------------------------------------------------------------

docs/source/grpcapi.rst

@@ -1,3 +1,5 @@
+.. _grpc-api:
+
 Low-level gRPC API
 ===============================================================================
 This section describes the low-level gRPC API that you can use from any other language or any other server/service proxy.