Command-line utilities to check source code for metrics (via SciTools Understand) and fail the build and/or publish trends in SONAR and/or publish histograms, kiviats and scatter plots of these metrics.
It would be neat to see the before&after as a whole in kiviat format
STDEV, AVG etc do not auto-adapt in the json file
We have a very large system (6MLOC) which Understand takes many minutes to open. Running srccheck takes 40 minutes on the build machine. Running both srccheck and srcplot will double this time. Is it possible to allow srccheck to also generate these histograms?
Right now they differ.
The Readme needs to be updated
Can you show metric names and current values versus max values?
When circles overlap, it would be nice to click on them and send them to the back so that others can come to the surface.
This metric is not native to Understand. We should document we support it.
Traceback (most recent call last):
File "/Applications/PyCharm.app/Contents/helpers/pydev/pydevd.py", line 1580, in <module>
globals = debugger.run(setup['file'], None, None, is_module)
File "/Applications/PyCharm.app/Contents/helpers/pydev/pydevd.py", line 964, in run
pydev_imports.execfile(file, globals, locals) # execute the script
File "/Applications/PyCharm.app/Contents/helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfile
exec(compile(contents+"\n", file, 'exec'), glob, loc)
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 340, in <module>
main()
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 313, in main
[violation_count, class_tracked_metrics] = process_class_metrics(db, arguments)
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 222, in process_class_metrics
return process_generic_metrics(db,cmdline_arguments,"--maxClassMetrics", cmdline_arguments["--classQuery"], _print_class_violation, cmdline_arguments.get("--regexIgnoreClasses", None))
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 197, in process_generic_metrics
stats_value = lambda_stats(metric_values())
File "/Library/Frameworks/Python.framework/Versions/3.5/lib/python3.5/statistics.py", line 434, in mode
'no unique mode; found %d equally common values' % len(table)
statistics.StatisticsError: no unique mode; found 2 equally common values
Request from client: be able to define a way to limit the scope. Given a large UDB, traverse just some of the entities.
====== Routine Metrics that failed the filters ==========
CountLineCode 116 (unnamed_function_1)
CountLineCode 173 (unnamed_function_1)
CountLineCode 166 (unnamed_function_1).nxThemingGenerator
CountLineCode 106 (unnamed_function_1).Route
...........................................
INFO: HIGHEST CountLineCode violator found (violation threshold is 92):CountLineCode 173 (unnamed_function_1)
...........................................
CountParams 15 (unnamed_function_1).(unnamed_function_2)
...........................................
INFO: HIGHEST CountParams violator found (violation threshold is 12):CountParams 15 (unnamed_function_1).(unnamed_function_2)
...........................................
...........................................
INFO: HIGHEST CyclomaticModified non violator found (violation threshold is 5):CyclomaticModified 5 (unnamed_function_1).nxThemingGenerator.rgba
...........................................
...........................................
INFO: HIGHEST MaxNesting non violator found (violation threshold is 2):MaxNesting 2 (unnamed_function_1).nxThemingGenerator.applyContrast
...........................................
It would be nice to cap not only on absolute maximum values of metrics, but also on other properties that better characterize the whole: average, standard deviation, etc. For example, you may have a Class with 90 as lack of cohesion but it is an outlier. You may want to cap on the Average or the Mean instead.
*** Response error connecting to http://sonarunj/api/manual_measures: b'{"err_code":400,"err_msg":"Unknown metric: routine_mode:cyclomaticmodified"}'
*** Response error connecting to http://sonarunj/api/manual_measures: b'{"err_code":400,"err_msg":"Unknown metric: class_median:maxinheritancetree"}'
*** Response error connecting to http://sonarunj/api/manual_measures: b'{"err_code":400,"err_msg":"Unknown metric: file_mode:countdeclclass"}'
*** Response error connecting to http://sonarunj/api/manual_measures: b'{"err_code":400,"err_msg":"Unknown metric: file_mode:countdeclsubprogram"}'
Right now, if there is no stats violation, I don't see the highest value. For non-stats violations, these INFO record are shown (#3). Can we have INFO for stats as well?
Similar to https://github.com/sglebs/udbcmp but with plots
The Readme says nothing about srcplot
the color of circles is based on the file path containing the entity. we got a bfp.udb from s101 which hits this bug
Each time a stats function is used with histograms enabled, avg, mean, max are computed.
we should remember their values when the first histogram is computed and save that for later use.
Hi
It is possible to use the metrics files from und as parameters of your application?
the tool is producing metric datapoints with value zero when the metric is absent.
It is hard to figure out the number for large metric values with E notation as it is.
Please add support for a scatterplot of 3 vars (X, Y, ball sizes) like http://matplotlib.org/examples/shapes_and_collections/scatter_demo.html
For easier integration with shell scripts and variables, please allow ''as a regex to be the same as no regex.
On Wed, Dec 7, 2016 at 5:20 AM, Marcio Marchini <[email protected]> wrote:
Hi,
At Nexxera we would like to cap the maximum number of methods in C# classes with my srccheck program. But in the case of C# classes, we want to ignore the Stub methods (automatically generated for properties). Unfortunately CountDeclMethod is also counting the Stub methods generated for setters and getters.
So… is there a way to count the non-stub methods?
I see I can check CountDeclPropertyAuto (would count the Stub ones) but what I need is the “NOT Stub ones”. I believe it would be: CountDeclMethod - CountDeclPropertyAuto
Is there such a metric and I am missing it?
If not… is it possible to add it? Or do I need to add a new synthetic one to srccheck like I did with CountParams at https://github.com/sglebs/srccheck/blob/master/utilities/utils.py#L44 ?
Thanks!
Hi Marcio,
You would need to write your own for this. Sorry not to have a better answer.
Sometimes we see unexpectedly low MEANs etc. Seeing histograms would be handy.
This is a request from a user: Please print max value of each metric being tracked regardless of being a violation
====== Class Metrics that failed the filters ==========
MEDIAN:MaxInheritanceTree 5
MODE:MaxInheritanceTree 5
Traceback (most recent call last):
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 412, in <module>
main()
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 376, in main
[violation_count, class_tracked_metrics] = process_class_metrics(db, arguments)
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 301, in process_class_metrics
return process_generic_metrics(db,cmdline_arguments,"--maxClassMetrics", cmdline_arguments["--classQuery"], _print_class_violation, cmdline_arguments.get("--regexIgnoreClasses", None), "Class")
File "/Users/mqm/git/srccheck/utilities/srccheck.py", line 197, in process_generic_metrics
lambda_name, adjusted_metric = metric.split(":")
ValueError: too many values to unpack (expected 2)
json:
{"CountDeclMethod":374,"MODE:CountDeclMethod":10,"MEDIAN:CountDeclMethod":10,"MaxInheritanceTree":11,"MODE:MaxInheritanceTree":4,"MEDIAN:MaxInheritanceTree":4,"AVG:PercentLackOfCohesion":50,"STDDEV:PercentLackOfCohesion:PercentLackOfCohesion":50}
Traceback (most recent call last):
File "/usr/bin/srccheck", line 9, in <module>
load_entry_point('srccheck==0.0.7', 'console_scripts', 'srccheck')()
File "/usr/lib/python3.4/site-packages/utilities/srccheck.py", line 358, in main
[violation_count, file_tracked_metrics] = process_file_metrics(db, arguments)
File "/usr/lib/python3.4/site-packages/utilities/srccheck.py", line 287, in process_file_metrics
return process_generic_metrics(db,cmdline_arguments,"--maxFileMetrics", cmdline_arguments["--fileQuery"], _print_file_violation, cmdline_arguments.get("--regexIgnoreFiles", None), "File")
File "/usr/lib/python3.4/site-packages/utilities/srccheck.py", line 261, in process_generic_metrics
scope_name)
File "/usr/lib/python3.4/site-packages/utilities/utils.py", line 83, in save_histogram
plt.figure() # new one, or they will be mixed
File "/usr/lib64/python3.4/site-packages/matplotlib/pyplot.py", line 535, in figure
**kwargs)
File "/usr/lib64/python3.4/site-packages/matplotlib/backends/backend_tkagg.py", line 84, in new_figure_manager
return new_figure_manager_given_figure(num, figure)
File "/usr/lib64/python3.4/site-packages/matplotlib/backends/backend_tkagg.py", line 92, in new_figure_manager_given_figure
window = Tk.Tk()
File "/usr/lib64/python3.4/tkinter/__init__.py", line 1856, in __init__
self.tk = _tkinter.create(screenName, baseName, className, interactive, wantobjects, useTk, sync, use)
_tkinter.TclError: no display name and no $DISPLAY environment variable
Currently we traverse the json as a dict. This means unordered (hash map). It would be nice to sort by metric name (process in alphabetical order).
Now that you can load limits from files (#14) please add support to overwrite these files with the currently found max values. This allows srccheck to be come adaptive, always automatically lowering these limits.
The code has 2 kiviat implementations. One can't plot negative numbers. How about a single kiviat that can plot negative numbers?
As in the histogram, it would be nice to be able to see if these values increased or decreased.
Even if I do not use --verbose, a lot of stuff is printed (HIGHEST valued elements). Can this be disabled?
Due to the escaping need din the shell, some people would like to pass the params as files instead of literal values.
Now with #8 we can see that many datapoints are zero. We would like to compute mean etc discarding these datapoints that are zero.
If you use metric FOO and also its AVG:FOO and STDDEV:FOO, the full dataset in understand will be traversed/computed again. Maybe this slows things down too much?
Having to run the tool multiple times is slow. Can multiple plots be generated for the same run?
The steps require getting full sources and using a virtualenv. Pls update.
I am trying to generate a scatter plot for pascal routines on a pascal UDB file. At one point, while traversing the routines, it is failing:
ERROR. Missing metric CountLineCode for X Axis (entity=Predeclared Routine (Assigned), file=None)
This is what I use for the query: --routineQuery: 'function ~Unknown ~Unresolved,method ~Unknown ~Unresolved,procedure ~Unknown ~Unresolved,routine ~Unknown ~Unresolved,classmethod ~Unknown ~Unresolved’
Interestingly my query to fetch the container file yields None, so I am in the dark:
entity.ref("definein, declarein")
I don’t have access to the sources. So, I’d like to know what exactly Predeclared means, so that I can decide if I can safely skip everything that is Predeclared in the query.
Thanks a lot in advance!!!
Now that srccheck has histogram support, pls explain it.
Pls change the y and x offsets
--version prints a string, not a version number
It takes 55 seconds to get project metrics for a UDB from the django sources. I emailed SciTools and got this as answer:
Hi Marcio,
I sent this to the engineers to look at, but I don't know how long it will be before they can look at it. I ran a few tests and there are a few specific metrics that are significantly slower than the others. If you limit it to the ones you need it might speed things up until they can get to this.
Here are the metric retrieval times from a 2.6MLOC Fuzzt C++ project:
AltAvgLineBlank : 0s
AltAvgLineCode : 0s
AltAvgLineComment : 0s
AltCountLineBlank : 0s
AltCountLineCode : 1s
AltCountLineComment : 0s
AvgCyclomatic : 1s
AvgCyclomaticModified : 1s
AvgCyclomaticStrict : 1s
AvgEssential : 1s
AvgLine : 0s
AvgLineBlank : 1s
AvgLineCode : 0s
AvgLineComment : 0s
CountDeclClass : 0s
CountDeclClassMethod : 1s
CountDeclClassVariable : 0s
CountDeclFile : 0s
CountDeclFileCode : 1s
CountDeclFileHeader : 0s
CountDeclFunction : 0s
CountDeclInstanceMethod : 1s
CountDeclInstanceVariable: 0s
CountDeclInstanceVariablePrivate: 1s
CountDeclInstanceVariableProtected: 1s
CountDeclInstanceVariablePublic: 0s
CountDeclMethod : 1s
CountDeclMethodAll : 24s
CountDeclMethodConst : 1s
CountDeclMethodFriend : 1s
CountDeclMethodPrivate : 0s
CountDeclMethodProtected : 1s
CountDeclMethodPublic : 0s
CountLine : 1s
CountLineBlank : 0s
CountLineCode : 0s
CountLineCodeDecl : 0s
CountLineCodeExe : 0s
CountLineComment : 0s
CountLineInactive : 1s
CountLinePreprocessor : 0s
CountPath : 1s
CountPathLog : 0s
CountSemicolon : 0s
CountStmt : 0s
CountStmtDecl : 1s
CountStmtEmpty : 0s
CountStmtExe : 0s
Cyclomatic : 1s
CyclomaticModified : 1s
CyclomaticStrict : 1s
Essential : 2s
Knots : 1s
MaxCyclomatic : 1s
MaxCyclomaticModified : 1s
MaxCyclomaticStrict : 1s
MaxEssential : 3s
MaxEssentialKnots : 2s
MaxInheritanceTree : 12s
MaxNesting : 6s
MinEssentialKnots : 1s
RatioCommentToCode : 0s
SumCyclomatic : 3s
SumCyclomaticModified : 4s
SumCyclomaticStrict : 3s
SumEssential : 3s
Total Run Time: 89 seconds
Maybe we need to support a flag to pass names of project metrics to skip when printing?
we should always show the max
It would be useful to have median and mean lines in the histogram
we would like to click on the circle and see what element it is
srccheck supports -a but this is not explained.
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.