Code2flow generates call graphs for dynamic programming language. Code2flow supports Python, Javascript, Ruby, and PHP.

C:\Users\sterg\Documents\GitHub\sparks-baird\CrabNet\crabnet>code2flow model.py
Code2Flow: Found 1 files from sources argument.
Code2Flow: Implicitly detected language as 'py'.
Code2Flow: Processing 1 source file(s).
Code2Flow:   model.py
Traceback (most recent call last):
  File "C:\Users\sterg\AppData\Local\Programs\Python\Python39\Scripts\code2flow-script.py", line 33, in <module>
    sys.exit(load_entry_point('code2flow==2.2.0', 'console_scripts', 'code2flow')())
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\site-packages\code2flow\engine.py", line 625, in main
    code2flow(
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\site-packages\code2flow\engine.py", line 531, in code2flow
    file_groups, all_nodes, edges = map_it(sources, language, no_trimming,
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\site-packages\code2flow\engine.py", line 317, in map_it
    raise ex
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\site-packages\code2flow\engine.py", line 312, in map_it
    file_ast_trees.append((source, language.get_tree(source, lang_params)))
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\site-packages\code2flow\python.py", line 155, in get_tree
    tree = ast.parse(f.read())
  File "c:\users\sterg\appdata\local\programs\python\python39\lib\encodings\cp1252.py", line 23, in decode
    return codecs.charmap_decode(input,self.errors,decoding_table)[0]
UnicodeDecodeError: 'charmap' codec can't decode byte 0x8f in position 4895: character maps to <undefined>

Scenario: I have a method (ruby) defined in a model somewhere. I'd like to identify all call sites which call this method and call sites that refer to those call sites up to some given limit. This sounds like a subset of the graph which is generated.

Following the instructions in the README, using python install setup.py, code2flow is added to PATH, but when trying to run, whether with a file or by itself, it instead opens a "How do you want to open this file?" menu.

If you select a text editor it will bring up this:

#!"C:\Users\[User]\AppData\Local\Programs\Python\Python39\python.exe"
# EASY-INSTALL-SCRIPT: 'code2flow==2.0.1','code2flow'
__requires__ = 'code2flow==2.0.1'
__import__('pkg_resources').run_script('code2flow==2.0.1', 'code2flow')

Uninstalling and reinstalling it does nothing.

Using Python 3.9.1 on Windows 10.

This might be a tough feature, yet it would be exceptionally useful if available. jQuery calls start with a dollar sign, or with (jQuery) with the parenthesis if used in non-conflict mode (iirc). That's where a lot of drawing is done.

I could give it a shot when I find some time.

From my experience I have seen the messiest code written in PHP. Hence I'd like to see support for PHP.

What do you think is required to implement a new language in code2flow?

It'd be nice if prototypes worked correctly. But there's the question of how they'd be represented on the graph.

Just to think out loud, this would possibly require code2flow to be able to differentiate between object methods (e.g. toString) and normal functions. This would then allow the graph to represent which object method is being called (e.g. if a class overrode the Object.toString method, that could be represented.

One thing I'm curious about is your decision to 'home-grow' the regular expressions etc - have you deliberately decided not to use existing tools (say uglifyjs) to generate an AST and work from there? On a related note, have you seen https://github.com/abort/javascript-call-graph?

Just pondering on the most appropriate tool for the job at the moment.

Input file is called "main.js" (so nothing spectacular).

Here's the traceback (I tried to correct the output from Powershell):

PS C:\Users\xxx> ..\..\Python\2.7\python.exe .\Apps\code2flow-master\code2fl
ow .\Apps\code2flow-master\2graph\main.js
Mapping .\Apps\code2flow-master\2graph\main
Removing comments and strings...
Generating function nodes...
Traceback (most recent call last):
  File ".\Apps\code2flow-master\code2flow", line 151, in <module>
    groups,nodes,edges = mapper.map()
  File "C:\Users\xxx\Apps\code2flow-master\code2flowlib\engine.py", line 814, in map
    fileGroup = self.generateFileGroup(name=filename,source=source)
  File "C:\Users\xxx\Apps\code2flow-master\code2flowlib\languages\javascript.py", line 329, in generateFileGroup
    return Group(name=name,source=source,fullSource=source,isFunction=True)
  File "C:\Users\xxx\Apps\code2flow-master\code2flowlib\languages\javascript.py", line 79, in __init__
    super(Group,self).__init__(**kwargs)
  File "C:\Users\xxx\Apps\code2flow-master\code2flowlib\engine.py", line 215, in __init__
    self.newObjectPattern = self.generateNewObjectPattern()
  File "C:\Users\xxx\Apps\code2flow-master\code2flowlib\languages\javascript.py", line 170, in generateNewObjectPattern
    return re.compile(r'new\s+%s\s*\('%self.name)
  File "C:\Python\2.7\lib\re.py", line 190, in compile
    return _compile(pattern, flags)
  File "C:\Python\2.7\lib\re.py", line 244, in _compile
    raise error, v # invalid expression
sre_constants.error: bogus escape: '\\2'

Currently, in my project, I just care about how a couple of classes interact with each other. Using code2flow has been an amazing way to see just how they are connected and show it to others. The issue is I've had to write a script that gets every other namespace to ignore with the exceptions of the ones I want. The ability to say, I want a graph with just these classes would be amazing.

Sample code

from typing import Callable

def trace(fn: Callable) -> Callable:
    def wrapper(*args, **kwargs):
        print('traced call')
        return fn(*args, **kwargs)
    return wrapper

def do_something(msg):
    return msg + ' world'

message = 'hello'
new_message = trace(do_something)(message)

when fed into code2flow would produce an AssertionError

File "%/bin/code2flow", line 33, in <module>
sys.exit(load_entry_point('code2flow', 'console_scripts', 'code2flow')())
File "%/code2flow/engine.py", line 630, in main
code2flow(
File "%/code2flow/engine.py", line 536, in code2flow
file_groups, all_nodes, edges = map_it(sources, language, no_trimming,
File "%/code2flow/engine.py", line 324, in map_it
file_group = make_file_group(file_ast_tree, source, extension)
File "%/code2flow/engine.py", line 208, in make_file_group
file_group.add_node(language.make_root_node(body_trees, parent=file_group), is_root=True)
File "%/code2flow/python.py", line 229, in make_root_node
variables = make_local_variables(lines, parent)
File "%/code2flow/python.py", line 120, in make_local_variables
variables += process_assign(element)
File "%/code2flow/python.py", line 79, in process_assign
ret.append(Variable(token, call, element.lineno))
File "%/code2flow/model.py", line 159, in __init__
assert points_to

As per Russian standard for block-schemes, there is a shape for functions/methods defined elsewhere. It looks like this in plain style:

I'll try to take a look at the code on my free time, but will promise nothing

I am running this on a ~1.2mb 38kloc browserified source file, and it's been running for > 10 mins at 100% cpu on a core i7. no disk output, and no output from strace so it appears internally cpu bound. a note of advisement as to the applicability to project size would probably be helpful.

Update: Traceback when I Ctrl-C it is:

Traceback (most recent call last):
  File "/home/user/.pyenv/versions/2.7.9/bin/code2flow", line 4, in <module>
    __import__('pkg_resources').run_script('code2flow==0.2', 'code2flow')
  File "build/bdist.linux-x86_64/egg/pkg_resources/__init__.py", line 723, in run_script
  File "build/bdist.linux-x86_64/egg/pkg_resources/__init__.py", line 1643, in run_script
  File "/home/user/.pyenv/versions/2.7.9/lib/python2.7/site-packages/code2flow-0.2-py2.7.egg/EGG-INFO/scripts/code2flow", line 151, in <module>

  File "build/bdist.linux-x86_64/egg/code2flowlib/engine.py", line 810, in map
  File "build/bdist.linux-x86_64/egg/code2flowlib/engine.py", line 369, in __init__
  File "build/bdist.linux-x86_64/egg/code2flowlib/engine.py", line 721, in _removeCommentsAndStrings

Update 2: the original source file had (believe it or not) >2mb of sourceMap (a single massive line) appended to it. Once I removed it, code2flow did complete after several minutes, so perhaps there is a pathological case in extremely long comment.

Hello! I tried to use code2flow on some python projects, And found it cannot distinguish files with the same name but in different directories, or functions with the same name but in different files. Also it becomes a mess if the two cases happen in the same time.

I also looked into the source code. The first problem can be fixed by changing line 348 in engine.py to include the full path into group name. But the second one seems much more difficult. I guess what you get from an ast Call node is only the function name, not enough to tell which function it is if there are more than one function with the same name in the project. But maybe we can use import statement to do this. But to do this, we need to relate the module name in import statement with the corresponding file full path.

Maybe you can check this problem and see if there's another solution? I think this is not a rare case so a solution is needed.

I'm trying to use code2flow on a very large codebase and the resulting out.png file keeps getting scaled to the point where it is unreadable.

I would like to be able to set a graph depth without having to pick a particular function.

code2flow my_package --downstream-depth 3

Or maybe just a generic --depth.

Scaled out.png file

~I realize there's probably a way to get dot/graphwhiz to reduce the scaling, but I haven't figured that out yet.~ Also I only really care about the first few nodes in the graph.

Edit: changing the output type to .svg solved the issue with low res .png files. But I still think being able to set a depth without specifying a target function would be a good feature.

Very cool project BTW.