Open-source Python package for Systematic Evaluation of Climate and Earth System Models
Home Page: http://pcmdi.github.io/pcmdi_metrics/
License: BSD 3-Clause "New" or "Revised" License
should be
example: huss ?
Should the metrics dir be dropped into the master directory (so WGNE by default) rather than hidden under /tmp which it currently is..?
It seems to me like there is some disagreement in the units of the obs and model data:
an014:/home/Paul.Durack/140129_metrics/test> /home/Paul.Durack/140129_metrics/WGNE/bin/python /home/Paul.Durack/140129_metrics/WGNE/bin/wgne_metrics_driver.py -p /home/Paul.Durack/140129_metrics/test/wgne_input_parameters_test.py
ref is: all
refs: [u'default']
OBS SHAPE IS (12, 72, 144)
tos GFDL-ESM2G (12, 72, 144) (12, 72, 144) default
---- shapes (12, 72, 144) (12, 72, 144)
/home/Paul.Durack/140129_metrics/WGNE/lib/python2.7/site-packages/numpy/ma/core.py:936: RuntimeWarning: overflow encountered in multiply
result = self.f(da, db, _args, *_kwargs)
tos annual mean bias is 273.421894797
---- shapes (12, 24, 144) (12, 24, 144)
tos annual mean bias is 274.136384059
---- shapes (12, 24, 144) (12, 24, 144)
tos annual mean bias is 274.230836527
---- shapes (12, 24, 144) (12, 24, 144)
tos annual mean bias is 272.039584515
an014:/home/Paul.Durack/140129_metrics/test>
Either the code should be able to deal with a deg_C -> K conversion, or I need to rewrite the obs and correct the temperature units
For each variable/json file ...
Add to both models (under "SimulationDescription") and observations the following keys (both of which can be obtained from the netCDF files...
Looks like the installer is hanging on the final build step:
[ 96%] No patch step for 'windspharm'
[ 97%] No update step for 'windspharm'
[ 97%] No configure step for 'windspharm'
[ 98%] Performing build step for 'windspharm'
-- windspharm build command succeeded. See also /home/p1d/140828_metrics/PCMDI_METRICS/tmp/uvcdat/uvcdat_build/windspharm-prefix/src/windspharm-stamp/windspharm-build-*.log
[ 99%] Performing install step for 'windspharm'
-- windspharm install command succeeded. See also /home/p1d/140828_metrics/PCMDI_METRICS/tmp/uvcdat/uvcdat_build/windspharm-prefix/src/windspharm-stamp/windspharm-install-*.log
[ 99%] Completed 'windspharm'
[ 99%] Built target windspharm
-- SCIPY build command succeeded. See also /home/p1d/140828_metrics/PCMDI_METRICS/tmp/uvcdat/uvcdat_build/SCIPY-prefix/src/SCIPY-stamp/SCIPY-build-*.log
[ 99%] Performing install step for 'SCIPY'
-- SCIPY install command succeeded. See also /home/p1d/140828_metrics/PCMDI_METRICS/tmp/uvcdat/uvcdat_build/SCIPY-prefix/src/SCIPY-stamp/SCIPY-install-*.log
[ 99%] Completed 'SCIPY'
[ 99%] Built target SCIPY
*******************************
Setting up CDAT - (Python + CDMS)... master
*******************************
Fetching the cdat project from GIT Repo...
git://github.com/UV-CDAT/uvcdat.git
Initialized empty Git repository in /home/p1d/140805_metrics/WGNE/tmp/uvcdat/uvcdat/.git/
remote: Counting objects: 28673, done.
remote: Total 28673 (delta 0), reused 0 (delta 0)
Receiving objects: 100% (28673/28673), 188.30 MiB | 62.40 MiB/s, done.
Resolving deltas: 100% (17273/17273), done.
Already on 'master'
CMAKE ARGS: cmake -DCDAT_BUILD_PARALLEL=OFF -DCMAKE_INSTALL_PREFIX=/home/p1d/140805_metrics/WGNE -DZLIB_PATCH_SRC=3 -DCDAT_DOWNLOAD_SAMPLE_DATA=OFF -DCDAT_BUILD_MODE=LEAN -DCDAT_BUILD_ESMF_ESMP=ON -DGIT_PROTOCOL=git:// /home/p1d/140805_metrics/WGNE/tmp/uvcdat/uvcdat
PATH:/home/p1d/140805_metrics/WGNE/Externals/bin:/home/gfdl/bin:/usr/local/OSoverlay/bin:/bin:/usr/bin:/usr/local/x64/bin:/usr/local/bin:/usr/local/explorer/bin:/usr/local/pgi/linux86/bin:/usr/local/cleanscape:/usr/local/toolworks/totalview/bin:/usr/local/autodesk/maya/bin:.:/home/p1d/bin
PWD:/home/p1d/140805_metrics/WGNE/tmp/uvcdat/uvcdat
-- The C compiler identification is GNU 4.4.7
-- The CXX compiler identification is GNU 4.4.7
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
CMake Error at CMakeLists.txt:16 (message):
[ERROR] CDAT requires an out of source Build. Run 'git clean -dfx' to
restore source dir.
-- Configuring incomplete, errors occurred!
See also "/home/p1d/140805_metrics/WGNE/tmp/uvcdat/uvcdat/CMakeFiles/CMakeOutput.log".
ERROR: Could not configure (cmake) cdat code (1)
~/140805_metrics/WGNE/tmp/metrics
Sorry...
This action did not complete successfully
Please try to re-run this task until successful before continuing further
ERROR: Could not compile (make) cdat code
~/140805_metrics/WGNE/tmp/metrics
Sorry...
This action did not complete successfully
Please try to re-run this task until successful before continuing further
Unmatched ".
public1:/home/p1d>
Looks for json test file at wrong place:
doutriaux1@maryam:[~]:[14]> wgne_metrics_driver.py -p /Users/doutriaux1/WGNE/test/wgne/basic_test_parameters_file.py
Traceback (most recent call last):
File "/Users/doutriaux1/WGNE/bin/wgne_metrics_driver.py", line 21, in <module>
obs_dic = json.loads(open(os.path.join(sys.prefix,"share","wgne","obs_info_dictionary.json")).read())
IOError: [Errno 2] No such file or directory: '/Users/doutriaux1/WGNE/Library/Frameworks/Python.framework/Versions/2.7/share/wgne/obs_info_dictionary.json'
whereas it is really at:
doutriaux1@maryam:[~]:[14]> ll /Users/doutriaux1/WGNE/share/wgne/
total 24
-rw-r--r-- 1 doutriaux1 45773 11K Jul 31 17:22 obs_info_dictionary.json
So this testing is on the GFDL machine:
public1:/home/p1d/140129_metrics/test> /home/p1d/140129_metrics/WGNE/bin/python /home/p1d/140129_metrics/WGNE/bin/wgne_metrics_driver.py -p /home/p1d/140129_metrics/test/wgne_input_parameters_test.py
ref is: all
refs: [u'default', u'alternate']
Traceback (most recent call last):
File "/home/p1d/140129_metrics/WGNE/bin/wgne_metrics_driver.py", line 94, in
applyCustomKeys(OUT,parameters.custom_keys,var)
AttributeError: 'module' object has no attribute 'custom_keys'
public1:/home/p1d/140129_metrics/test>
It appears that there is a hard coded unit conversion in some of the python scripts that are running with the metrics driver. I've found the following line in '/home/Paul.Durack/140701_metrics/WGNE/lib/python2.7/site-packages/metrics/wgne/ mean_climate_metrics_calculations.py':
if var == 'tas': do_glb = do_glb + 273.13
We want to add a new diagnostic on the position of the subtropical jets
Considering the performance improvements we've discovered jumping from 1.7.1 to 1.9.0b1, I'd be keen to get a devel branch of the metrics using the latest numpy version.. I can then pull this across to GFDL for further testing..
I'd be reluctant to commit this to the master branch until we've verified that the numbers coming out of numpy 1.9.0b1 are reasonable, that masked arrays are being dealt with properly etc etc..
As I've now got some files generating using the mean/min/max/median/std from GFDL CM2.6, I can rerun this code with a new build and check to make sure numbers are reasonable..
As part of standard checks before runnnig metrics, check units are compatible.
tos_GFDL-ESM2G__historical__198501-200512.interpolated.linear.2.5x2.5-clim.nc
Also in this demo interpolated file:
It might be a good idea to consider renaming wgne-wgcm_metrics to pcmdi_metrics before things begin to gain too much more momentum..? Thoughts?
It appears the code is bombing if it can't find data, rather than skipping (the preferred result):
ref is: all
refs: [u'default', u'alternate', u'ref3']
OBS SHAPE IS (12, 72, 144)
psl GFDL-CM4 (12, 72, 144) (12, 72, 144) default
/home/Paul.Durack/140701_metrics/WGNE/bin/wgne_metrics_driver.py:241: UserWarning: Model and observation units differed, converted model (hPa) to observation unit (Pa)
warnings.warn("Model and observation units differed, converted model (%s) to observation unit (%s)" % (dm.units,do.units))
---- shapes (12, 72, 144) (12, 72, 144)
psl annual mean bias is 19.1959297534
---- shapes (12, 24, 144) (12, 24, 144)
psl annual mean bias is -199.580140365
---- shapes (12, 24, 144) (12, 24, 144)
psl annual mean bias is 100.163863037
---- shapes (12, 24, 144) (12, 24, 144)
psl annual mean bias is 157.004066588
Traceback (most recent call last):
File "/home/Paul.Durack/140701_metrics/WGNE/bin/wgne_metrics_driver.py", line 171, in <module>
metrics_dictionary["References"][ref] = obs_dic[var][obs_dic[var][ref]]
KeyError: u'rnl_ncep'
/git/WGNE/bin/python/git/wgne-wgcm_metrics/src/python/wgne/scripts/wgne_metrics_driver.py –p ~/git/wgne-wgcm_metrics/doc/wgne_input_parameters_test.py
This is related to CDAT/cdat#449 and CDAT/cdat#448
It seems like cleaning up those issues might solve the below:
public1:/home/p1d/140418_metrics/test> source ../WGNE/bin/setup_runtime.csh
public1:/home/p1d/140418_metrics/test> which python
../WGNE/bin/python
public1:/home/p1d/140418_metrics/test> which wgne_metrics_driver.py
../WGNE/bin/wgne_metrics_driver.py
public1:/home/p1d/140418_metrics/test> wgne_metrics_driver.py -p wgne_input_parameters_test.py
Traceback (most recent call last):
File "../WGNE/bin/wgne_metrics_driver.py", line 46, in
exec("import %s as parameters" % fnm)
File "", line 1, in
ImportError: No module named wgne_input_parameters_test
public1:/home/p1d/140418_metrics/test>
For clarity, we should the following keys, only for output (json and ascii files):
"default" => "defaultReference"
"alternate => "alternateReference"
no change to "ref3", ... "refn"
GFDL tos
need to explain this in error message or add current path to it at load time
@taylor13 please confimr this is right:
default should be: "cdat_lean" (cdms2+cdutil+genutil")
add option for graphics (vcs but NO GUI)
add option for cmor
Ok so this seems to be getting closer..?
[ 89%] Performing install step for 'Numexpr'
CMake Error at /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/Numexpr-prefix/src/Numexpr-stamp/Numexpr-install.cmake:56 (message):
CTEST_FULL_OUTPUT
Command failed: 1
'env' 'PYTHONPATH=/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages:/work/dur041/tmp/140426_metrics/WGNE/lib/python/site-packages' '/work/dur041/tmp/140426_metrics/WGNE/bin/python' 'setup.py' 'install' '--prefix=/work/dur041/tmp/140426_metrics/WGNE'
From /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/Numexpr-prefix/src/Numexpr-stamp/Numexpr-install-out.log
-------------------------------------------------
-------------------------------------------------
From /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/Numexpr-prefix/src/Numexpr-stamp/Numexpr-install-err.log
-------------------------------------------------
Traceback (most recent call last):
File "setup.py", line 16, in <module>
import numpy
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/__init__.py", line 137, in <module>
import add_newdocs
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/add_newdocs.py", line 9, in <module>
from numpy.lib import add_newdoc
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/lib/__init__.py", line 13, in <module>
from polynomial import *
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/lib/polynomial.py", line 17, in <module>
from numpy.linalg import eigvals, lstsq, inv
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/linalg/__init__.py", line 48, in <module>
from linalg import *
File "/work/dur041/tmp/140426_metrics/WGNE/lib/python2.7/site-packages/numpy/linalg/linalg.py", line 23, in <module>
from numpy.linalg import lapack_lite
ImportError: liblapack.so: cannot open shared object file: No such file or
directory
-------------------------------------------------
make[2]: *** [Numexpr-prefix/src/Numexpr-stamp/Numexpr-install] Error 1
make[1]: *** [CMakeFiles/Numexpr.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
-- PYLIBXML2 build command succeeded. See also /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/PYLIBXML2-prefix/src/PYLIBXML2-stamp/PYLIBXML2-build-*.log
[ 90%] Performing install step for 'PYLIBXML2'
-- PYLIBXML2 install command succeeded. See also /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/PYLIBXML2-prefix/src/PYLIBXML2-stamp/PYLIBXML2-install-*.log
[ 90%] Completed 'PYLIBXML2'
[ 90%] Built target PYLIBXML2
-- ESMF build command succeeded. See also /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/ESMF-prefix/src/ESMF-stamp/ESMF-build-*.log
[ 90%] Performing install step for 'ESMF'
-- ESMF install command succeeded. See also /work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat_build/ESMF-prefix/src/ESMF-stamp/ESMF-install-*.log
[ 90%] Completed 'ESMF'
[ 90%] Built target ESMF
make: *** [all] Error 2
ERROR: Could not compile (make) cdat code (4)
/work/dur041/tmp/140426_metrics/WGNE/tmp/uvcdat/uvcdat /work/dur041/tmp/140426_metrics/WGNE/tmp/metrics
Sorry...
This action did not complete successfully
Please re-run this task until successful before continuing further
Also please review the installation FAQ it may assist you
http://esgf.org/wiki/ESGFNode/FAQ
ERROR: Could not compile (make) cdat code
/work/dur041/tmp/140426_metrics/WGNE/tmp/metrics
Sorry...
This action did not complete successfully
Please re-run this task until successful before continuing further
Also please review the installation FAQ it may assist you
http://esgf.org/wiki/ESGFNode/FAQ
We should add UV-CDAT and date stamp to all output formats to make debugging and problem discovery easier
[ 99%] Performing install step for 'CDATLogger'
Ping succeeded.
[100%] Completed 'CDATLogger'
[100%] Built target CDATLogger
CMAKE ARGS
UVCDAT BDIR/Users/durack1/Desktop/metrics/wgne-wgcm_metrics/WGNE/tmp/uvcdat
Hangs indefinitely at this point
^C ERROR: Could not load CDMS (cdms2) module
~/Desktop/metrics/wgne-wgcm_metrics/WGNE/tmp/metrics
Sorry...
This action did not complete successfully
Please re-run this task until successful before continuing further
Also please review the installation FAQ it may assist you
http://esgf.org/wiki/ESGFNode/FAQ
As below
running install_egg_info
Writing /home/p1d/140701_metrics/WGNE/lib/python2.7/site-packages/metrics-0.1.0-py2.7.egg-info
rmdir: failed to remove `/home/p1d/140701_metrics/WGNE/sample_data': Directory not empty
*******************************
UVCDAT - master - Install Success
Metrics - devel - Install Success
*******************************
Please test as follow:
source /home/p1d/140701_metrics/WGNE/bin/setup_runtime.sh
wgne_metrics_driver.py -p /home/p1d/140701_metrics/WGNE/test/wgne/basic_test_parameters_file.py
compare: /home/p1d/140701_metrics/WGNE/test/wgne/tos_2.5x2.5_esmf_linear_metrics.json.good with wgne_install_test_results/metrics_results/installationTest/tos_2.5x2.5_esmf_linear_metrics.json
*******************************
Create your customized input_parameters.py (inspire yourself from examples in /home/p1d/140701_metrics/WGNE/doc/wgne_input_parameters_sample.py
Once you have a parameter file run:
source /home/p1d/140701_metrics/WGNE/bin/setup_runtime.sh
wgne_metrics_driver.py -p /path/to/your/edited/parameter_file.py
*******************************
Once everything is ok, you can safely remove the temporary directory: /home/p1d/140701_metrics/WGNE/tmp
*******************************
Is this expected?
public1:/home/p1d> ls | xargs du -sh
4.0G 140516_metrics/
4.0G 140701_metrics/
8.3G 140828_metrics/
Rather than 'default'
},
"source": "UKMETOFFICE-HadISST-v1-1"
}
},
"GridInfo": {
"GridName": "2.5x2.5",
"GridResolution": [
72,
144
],
"RegridMethod": "linear",
"RegridTool": "esmf"
},
"References": {
"default": {
"CMIP_CMOR_TABLE": "Omon",
"period": "198002-200501",
"shape": "(12, 180, 360)"
}
},
"RegionalMasking": {
"global": null,
"ocean": 0.0,
"terre": 100.0
},
"metrics_git_sha1": "b869df6dd490ce8c7abd001c874b2839f365d4d1",
"uvcdat_version": "esgf-v1.7.0-12-gfb2c39b"
}
would be nice to run a small test script on subset of "tos" for the sake of verifying everything probably went ok
need to add realm/fqcy
e.g:
ocn/mo
after the user defined root path to obs
The pytz module (I thought it came free as a standard python module) doesn't appear to be installed in our current metrics package..
It's very useful to deal with timezoning stuff, and I'm using it for global attribute creation for files.
I'd advocate we drop this package back into the metrics..
Ok so the most up-to-date build was tested on the GFDL system yesterday and it seems to be working fine with ocean data.
Additionally the axis_aliases also appear to pickup axes from their units attribute.
Is there a need to generate a specific "metrics" build option so that ongoing changes to the "ESGF" build flag will not affect the metrics package in the future?
Try to keep as much information from IPSL controlled vocabulary as possible, see:
https://github.com/PCMDI/wgne-wgcm_metrics/wiki/Reference-Syntax-indexation
The work done on the GFDL infrastructure to best use the metrics package to create climatologies and then generate metrics from these has been done. The scripts which have been written to do this should be available somewhere so that other users don't have to start from scratch.
public1:/home/p1d/140418_metrics/test> /home/p1d/140418_metrics/WGNE/bin/python /home/p1d/140418_metrics/WGNE/bin/wgne_metrics_driver.py -p /home/p1d/140418_metrics/test/wgne_input_parameters_test.py
ref is: all
refs: [u'default']
OBS SHAPE IS (12, 72, 144)
tos GFDL-ESM2G (12, 72, 144) (12, 72, 144) default
Error while processing observation tos for variable default:
Obs and Model -GFDL-ESM2G- have different units (K vs degC) cowardly refusing to proceed
possibly csv and/or database
Ok so first test trying to plot a portrait plot:
durack1@oceanonly:[test_portrait]:[test_branch]:[81]> python portrait_test.july23.py
Traceback (most recent call last):
File "portrait_test.july23.py", line 17, in <module>
import portraits
File "/export/durack1/git/wgne-wgcm_metrics/test_portrait/portraits.py", line 12, in <module>
import vcs
File "/usr/local/uvcdat/2014-08-13/lib/python2.7/site-packages/vcs/__init__.py", line 52, in <module>
install_vcs._files()
File "/usr/local/uvcdat/2014-08-13/lib/python2.7/site-packages/vcs/install_vcs.py", line 41, in _files
shutil.copyfile(icon_file_cp, icon_file)
File "/usr/local/uvcdat/2014-08-13/lib/python2.7/shutil.py", line 82, in copyfile
with open(src, 'rb') as fsrc:
IOError: [Errno 13] Permission denied: '/usr/local/uvcdat/2014-08-13/lib/python2.7/site-packages/vcs/../../../../bin/vcs_icon.xbm'
And yes, the only way I could get this far down the path was to use bash.. csh bombed admirably.. I'll await the updated build on oceanonly and try again either later this week or next..
if you have already a working cdat it should use this one. Or at least have an optino to tell the installer to use this one.
The problem with the rewritten GFDL data was a lack of CF compliance, so incorrect dimension names (zt,yh,xh) - it would be great if such problems were dealt with by software, as CMOR/CF lack of compliance with input data is likely to be a problem that appears time and time again..
Some sanity checks, along with axes checks (so ensure that 'tos' variables are 3D, rather than 4D for e.g.) would also be a good idea.
so that you can have "in-house" extra metrics
an102:/home/Paul.Durack/140129_metrics/test> /home/Paul.Durack/140129_metrics/WGNE/bin/python /home/Paul.Durack/140129_metrics/WGNE/bin/wgne_metrics_driver.py -p /home/Paul.Durack/140129_metrics/test/wgne_input_parameters_test.py
ref is: all
refs: [u'default']
OBS SHAPE IS (12, 72, 144)
Failed to get variable tos for version: GFDL-ESM2G, error:
'NoneType' object has no attribute 'getBounds'
from parameter file we should be able to have an option to point to an existing json file and extend it.
edit driver to look at parameter file and run custom local metrics.
Traceback (most recent call last):
File "/export/gleckler1/git/wgne-wgcm_metrics/src/python/wgne/scripts/wgne_metrics_driver.py", line 57, in
dup("ok table thing",obs_dic[var][obs_dic[var]["default"]]["CMIP_CMOR_TABLE"],"---")
KeyError: u'CNES_AVISO_L4'
All 4d data now fails (e.g., 'ta_850','ua_200') as a result of the new regions_dict in wgne_metrics_driver.py only reading root variable (without level).
As we've agreed that building udunits to attempt to most accurately deal with weird variable units is a good idea.. Maybe we should also add CMOR to the build - that way modeling groups have quite literally all the tools installed and built for them so they can rewrite their data if required before pushing through the metrics meat grinder..?
This needs to be done with the land sea mask of the model. A version of these for CMIP5 can be found at: /work/gleckler1/processed_data/cmip5_fixed_fields/sftlf
But I guess we also need land sea masks on a series of target grids... 5x5, 2.5x2x5, 1x1... we should build these in somehow.
This will also require new switches in the parameter file or a list[] of ocean-only and land-only variables.
need to come up with system to tell driver which set of metrics to use (-m/-e option?)
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
Django
The Web framework for perfectionists with deadlines.
Laravel
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft
Open source projects and samples from Microsoft.
Google
Google ❤️ Open Source for everyone.
Alibaba
Alibaba Open Source for everyone
D3
Data-Driven Documents codes.
Tencent
China tencent open source team.