MgxParser

This version(0.5.2) was compiled on 20240326

Introduction

MgxParser is a C++ lib used to parse Age of Empires II game records.

It is firstly developed at 2020 to support running of aocrec.com as a replacement of a prior php parser( which was a modified version of recanalyst).

Use a docker image

A docker image is available at https://hub.docker.com/r/lichifeng/mgxparser Just pull and run it, send a record file to it and get a JSON response.

docker pull lichifeng/mgxparser
docker run -e MAX_CONNECTIONS=100 -p 4400:4400 lichifeng/mgxparser

Then, send a record file and a json command to it:

curl -X POST -F "file=@/path/to/record.mgx" -F "json={\"map\":\"HD\"}" http://localhost:4400/

To build the docker image manually, use the Dockerfile in root directory.

docker build -t mgxparser .
docker run -it -p 4400:4400 --rm mgxparser

Basic usage

MgxParser parse(Settings) function: One takes a record file path as input and returns a JSON string contains important information about this record.

If input is a zip archive, MgxParser will try to Extract2Files first file in the archive and parse it as a record. This function is intended to tackle compressed record files from old version of aocrec.com.

No error was expected even if something unexpected happend in parsing process. Users should check parsing result by the value of key status and message in JSON response.

Read src/include/MgxParse.h for more details.

A demo of MgxParser will be compiled using demo.cc. It can be used to parse a record file and print the result to stdout.

mgxparser --help

Output:

Usage: ./mgxparser [options] [file]
Options:
  -m    Generate a normal map
  -M    Generate a HD map
  -b/-B Generate base64 encoded normal/HD mapdata
  -u    Extract the original file in .zip archive
  -n    Indentation of the json string. i.e. -n2
  --help Display this help message
Example: ./mgxparser -m -n2 demo.mgx
Notice: if -b/-B is provided, -m/-M will be ignored.

For node addon usage, see src/node/README.md,
See Compile node addon with cmake-js for details.

JSON output

The JSON output of MgxParser is a string contains parsed information of a record.
Some keys are optional, they will be omitted if not available.
Those keys are ensured to be in the output:

• status: A string indicates the status of parsing.
• guid: A string of 32 characters, the unique identifier of this record.
• md5: A string of 32 characters, the md5 of the record file.
• fileext: File extension with leading dot of original input file. e.g. .mgx, .mgz, .aoe2record, .zip, etc. NOT THE EXTRACTED FILE EXTENSION!
• filename: File name of original input file with extension. e.g. "rec.mgz", "rec.zip", etc. NOT THE EXTRACTED FILE NAME!
• realfile: File name of the extracted file in .zip archive. e.g. "rec.mgx", "rec.aoe2record", etc.
• realext: File extension of the extracted file in .zip archive. e.g. .mgx, .aoe2record, etc.
• realsize: Size in bytes of the raw record file, not the zip archive.
• gameTime: Time when this game played. Later DE versions have this recorded. Older versions use last modified time of the record file. If the input stream is not a file, will be current time.
• message: A string contains some hints about the parsing process.
• matchup: Matchup of this game. e.g. "1v1", "4v4", "1v2v3", etc. "N/A" if not available.
• version.code: A string indicates the version of the record. e.g. "AOC10C", "DE", etc. "UNSUPPORTED" if not available.
• rawEncoding: The encoding of the record file. e.g. "UTF-8", "ISO-8859-1", etc. "UNKNOWN" if not detected.

About 'status'

The 'status' value can be checked to determine whether the parsing was successful:

• 'perfect': Means all data in the record was parsed.
• 'good': Means the main data in header part was scanned, and at least the map can be generated.
• 'valid': Means the header and body parts can be decompressed, but there are problems in parsing.
• 'invalid': Means the record is invalid or cannot be parsed, but MgxParser worked fine.

About 'guid'

The 'guid' value is a 32 characters string, it is the unique identifier of this record:

• In higher versions(HD, DE), it is contained in the record file.
• In earlier versions, it is generated from some stable data in the record, records from the same game are expected to have a same guid.
  In very rare cases, like some players quit at the very very beginning of the game and others continued, the guid may be different.
• If the record is invalid and no enough info for generating a guid, the md5 of the record file will be used.

Performance

I did't do elegent profiling for it, simply measured its performance
by \time -v MgxParser recordfile.mgx

Pure json output without map generation is very fast. Normally cost <50ms and <10m peak memory( in my synology DS1621+ NAS docker container).

Command being timed: "/Workspace/mgxParserCPP-master/build/MgxParser test/sampleRecords/AOC10_4v4_5_5e3b2a7e.mgx"
User time (seconds): 0.01
System time (seconds): 0.00
Percent of CPU this job got: 100%
Elapsed (wall clock) time (h:mm:ss or m:ss): 0:00.01
... ...
Maximum resident set size (kbytes): 9124
Average resident set size (kbytes): 0
... ...
Exit status: 0

Exceptions

Parsing an AoE record is fretful, exceptions are expected everywhere.
But most of the time, exceptions don't mean worthless, some part of the
record can still be read, so MgxParser is wrapped with an outermost
try-catch block to avoid easy crash.
When exceptions happen, they are normally caught and previously extracted
information are returned. Some status hints are attached, so users will
get notified.

Requirements

These libraries are required to build MgxParser:

• CImg(bundled): Used to generate mini maps. This library is bundled into libs/CImg and compiled with MgxParser. of record. Project page
• libpng: Required by CImg.

  sudo apt-get install libpng-dev

• libz: Required by libpng and header data decompressing.
• iconv: Used to convert encoding of old records.
• nlohmann/json: Required to generate JSON output. A single-header version of nlohmann/json is bundled into libs/nlohmann_json_3 and compiled with MgxParser.
• doxygen: Required to generate documents from source code. Not necessary.
• Graphviz: Required by doxygen.

  sudo apt-get install doxygen graphviz

  googletest: Required for testing.

All dependencies of MgxParser are bundled into the executive file after compilation.

How to compile

A complie option BUILD_STATIC is provided to control whether to build a static library. Because some denpendencies don't provide a static version on alpine.

To compile this project, latest version of CMake is recommended. See https://apt.kitware.com/ for details on this. I use CMake 3.24.1 and never tested other versions. On Ubuntu18 and newer versions, fresh version of cmake can be installed by:

sudo apt-get install cmake

A C++17 compiler is also needed. G++-9, Clang++-10 or newer compilers are recommended. Other compilers may work, but I never tested, too.

sudo apt-get install g++-8
#AND
sudo apt-get install clang
#Then, build the source file
cmake . -DCMAKE_CXX_COMPILER=/usr/bin/clang++-10 -DCMAKE_BUILD_TYPE=Release --fresh
#Last, compile it!
make

Compile node addon with cmake-js

To compile it to a node addon, cmake-js and node-addon-api is required.
Use npm to install cmake-js and node-addon-api:

npm install

Then, compile it with cmake-js:

npx cmake-js compile -p 6 # or other threads settings

The compiled node addon will be in build/Release directory.
A demo of node addon is in test/node_addon_test.js directory.

Version log

• 0.5.2: Modifications compiling on alpine. Add a static library option. Updated spdlog.
• 0.5.0: Refactored json output, make some keys mandatory. Some key names changed!
• 0.4.6: Replace md5 with openssl MD5 algorithm. Add ability to generate base64 encoded map data.
• 0.4.5: Reorganized source code. Refactored parse() function. Add node addon support. Add docker workflow.
• 0.4.2: A version used on aocrec.com before Feb. 2024.
• 0.4.0: Prepare to go online. Add english language pack. Fixed more bugs.
• 0.3.x: Bug fix and some changes to meet requirements of mgxhub.
• 0.2.0: Reoragnized source files, tested with 130000+ records in aocrec.com.
• 0.1.0: Now MgxParser can work with Node.js(Fastify) and parse a uploaded. record. All operations are done in memory without file storage.

Resources

Find records for test: https://www.ageofempires.com/stats/ageiide
https://github.com/topics/age-of-empires
DE Replays Manager: https://github.com/gregstein/DE-Replays-Manager
Awesome Age of Empires II resources: https://github.com/Arkanosis/awesome-aoe2
https://aok.heavengames.com/blacksmith/lister.php?category=utilities
https://aok.heavengames.com/blacksmith/lister.php?category=records
AoE II Development Collective https://siegeengineers.org/
https://www.twaoe2wiki.com/

Caution

This project is still under development. No warrenty for anything!