dota-replay-parser
A really bad C++ emscripten compiled port of Dotabuff's manta replay parser. Also borrows a lot from the Alice Source 1 replay parser.
Why? For Source 1, Alice could be compiled with Emscripten and I used that to make a browser based replay viewer on top of it. I wanted to do something similar for Source 2 but Alice has not been updated. I don't have the expertise or knowledge on Dota 2 replays or C++ to be able to write my own parser from scratch so I based everything off manta and Alice.
I also tried just compiling manta with GopherJS, but the performance was slow. I don't know how to write Go and combined with my poor knowledge of the inner workings of replays, I'm clueless as to how to maybe optimize it.
Building
Sorry, but I haven't used a build system yet. It's a miracle I got anything to compile to begin with since I'm a total noob. But I'll describe what I did at least and maybe that will be enough for someone to be able to produce a build. If I learn how to create a makefile I'll update this as well.
Dependencies
First get emscripten and boost installed. For me, emscripten ended up in /home/devilesk/emsdk_portable and boost in /home/devilesk/boost_1_59_0. I have /usr/local/include/boost symlinked to /home/devilesk/boost_1_59_0/boost. Follow the updating the sdk steps for emscripten.
Then download the snappy source and build using emscripten. Basically just ./emconfigure ./configure then ./emmake make in the snappy directory. I ended up with everything in /home/devilesk/snappy and the compiled code in /home/devilesk/snappy/.libs
Then download protobuf-emscripten and follow the steps in its README. I ended up with /home/devilesk/protobuf-emscripten and the compiled code in /home/devilesk/protobuf-emscripten/2.6.1/src/.libs.
Compiling
Download this project's source.
At this point it might be good to update the protos in the protobufs directory. To do so you'll need to build and install protobuf 2.6.1. Then update the protobufs/update_proto.sh script with the correct path to your protoc (for me it is /usr/local/bin/protoc) and path to the directory where the script is located which is where the protobufs will be downloaded to (for me it is /home/devilesk/aegis/protobufs).
Now to actually compile the replay parser with emscripten. Run the following commands. You'll need to change the library include paths based on where your dependencies are.
em++ -o parser.bc -std=c++11 parser.cpp packet_entity.cpp message_lookup.cpp huffman.cpp property_serializers.cpp string_table.cpp class_info.cpp flattened_serializers.cpp fieldpath.cpp quantizedfloat.cpp property.cpp property_decoder.cpp bitstream.cpp protobufs/demo.pb.cc protobufs/netmessages.pb.cc protobufs/networkbasetypes.pb.cc protobufs/network_connection.pb.cc protobufs/usermessages.pb.cc -I /home/devilesk/snappy -I /home/devilesk/protobuf-emscripten/2.6.1/src -I /usr/local/include -s TOTAL_MEMORY=512000000 -O3
em++ -o replayviewer.bc -std=c++11 binding.cpp replayviewer.cpp -I /home/devilesk/snappy -I /home/devilesk/protobuf-emscripten/2.6.1/src -I /usr/local/include -s TOTAL_MEMORY=512000000 -O3
em++ --bind -std=c++11 parser.bc replayviewer.bc -o replayviewer.js -I /home/devilesk/snappy -I /home/devilesk/protobuf-emscripten/2.6.1/src -I /usr/local/include /home/devilesk/protobuf-emscripten/2.6.1/src/.libs/libprotobuf.so.9.0.1 /home/devilesk/snappy/.libs/libsnappy.so -s TOTAL_MEMORY=512000000 -O3
The first two commands generate two bitcode files, parser.bc and replayviewer.bc. The last command links them together along with the other libraries and outputs the javascript library replayviewer.js and replayviewer.js.mem.
Usage
If you start up a local server to serve the directory containing the source and compiled files (python -m SimpleHTTPServer if you have Python 2.7) you can go to demo.html to try the parser.
I'm currently working on using this replay parser in my dota interactive map to create a replay viewer.