I've read PEP 328 and similar questions like this, but they still didn't answer my questions about python 3 import. My question is:
why does a single dot . relative import ever produce an error like the following?
code: (__init__.py exists in the same directory)
from . import module_x
error:
ImportError: attempted relative import with no known parent package
Yes, there are workarounds to get around of this error for example by running python -m from the parent directory, but why?
Shouldn't . always refer to the same directory of the current module and the example always works as long as the module_x exists in the same directory?
Particularly PyCharm tries to run unittest test cases from the directory where the test file is, and would always fail the . import. Some may argue it's PyCharm's problem, but I think a simple import like this should really always work regardless where the interpreter runs.
There is a detailed answer here, also here. Sorry I cannot write this as a comment.
I'm trying out Freeling's API for python. The installation and test were ok, they provide a sample.py file that works perfectly (I've played around a little bit with it and it works).
So I was trying to use it on some other python code I have, in a different folder (I'm kind of guessing this is a path issue), but whenever I import freeling (like it shows on the sample.py):
import freeling
FREELINGDIR = "/usr/local";
DATA = FREELINGDIR+"/share/freeling/";
LANG="es";
freeling.util_init_locale("default");
I get this error:
ModuleNotFoundError: No module named 'freeling'.
The sample.py is located on the ~/Freeling-4.0/APIs/Python/ folder, while my other file is located in ~/project/, I dont know if that can be an issue.
Thank you!
A simple solution is to have a copy of freeling.py in the same directory as your code, since python will look there.
A better solution is to either paste it in one of the locations where it usually checks (like the lib folder in its install directory), or to tell it that the path where your file is should be scanned for a module.
You can check out this question to see how it can be done on Windows. You are basically just setting the PYTHONPATH environment variable, and there will only be minor differences in how to do so for other OSes. This page gives instructions that should work on Linux systems.
I like this answer since it adds the path at runtime in the script itself, doesn't make persistent changes, and is largely independent of the underlying OS (apart from the fact that you need to use the appropriate module path of course).
You need to set PYTHONPATH so python can find the modules if they are not in the same folder.
What is the best way to call java from python?
(jython and RPC are not an option for me).
I've heard of JCC: http://pypi.python.org/pypi/JCC/1.9
a C++ code generator for calling Java from C++/Python
But this requires compiling every possible call; I would prefer another solution.
I've hear about JPype: http://jpype.sourceforge.net/
tutorial: http://www.slideshare.net/onyame/mixing-python-and-java
import jpype
jpype.startJVM(path to jvm.dll, "-ea")
javaPackage = jpype.JPackage("JavaPackageName")
javaClass = javaPackage.JavaClassName
javaObject = javaClass()
javaObject.JavaMethodName()
jpype.shutdownJVM()
This looks like what I need.
However, the last release is from Jan 2009 and I see people failing to compile JPype.
Is JPype a dead project?
Are there any other alternatives?
You could also use Py4J. There is an example on the frontpage and lots of documentation, but essentially, you just call Java methods from your python code as if they were python methods:
from py4j.java_gateway import JavaGateway
gateway = JavaGateway() # connect to the JVM
java_object = gateway.jvm.mypackage.MyClass() # invoke constructor
other_object = java_object.doThat()
other_object.doThis(1,'abc')
gateway.jvm.java.lang.System.out.println('Hello World!') # call a static method
As opposed to Jython, one part of Py4J runs in the Python VM so it is always "up to date" with the latest version of Python and you can use libraries that do not run well on Jython (e.g., lxml). The other part runs in the Java VM you want to call.
The communication is done through sockets instead of JNI and Py4J has its own protocol (to optimize certain cases, to manage memory, etc.)
Disclaimer: I am the author of Py4J
Here is my summary of this problem: 5 Ways of Calling Java from Python
http://baojie.org/blog/2014/06/16/call-java-from-python/ (cached)
Short answer: Jpype works pretty well and is proven in many projects (such as python-boilerpipe), but Pyjnius is faster and simpler than JPype
I have tried Pyjnius/Jnius, JCC, javabridge, Jpype and Py4j.
Py4j is a bit hard to use, as you need to start a gateway, adding another layer of fragility.
Pyjnius docs and Github.
From the github page:
A Python module to access Java classes as Python classes using JNI.
PyJNIus is a "Work In Progress".
Quick overview
>>> from jnius import autoclass
>>> autoclass('java.lang.System').out.println('Hello world')
Hello world
>>> Stack = autoclass('java.util.Stack')
>>> stack = Stack()
>>> stack.push('hello')
>>> stack.push('world')
>>> print stack.pop()
world
>>> print stack.pop()
hello
If you're in Python 3, there's a fork of JPype called JPype1-py3
pip install JPype1-py3
This works for me on OSX / Python 3.4.3. (You may need to export JAVA_HOME=/Library/Java/JavaVirtualMachines/your-java-version)
from jpype import *
startJVM(getDefaultJVMPath(), "-ea")
java.lang.System.out.println("hello world")
shutdownJVM()
I'm on OSX 10.10.2, and succeeded in using JPype.
Ran into installation problems with Jnius (others have too), Javabridge installed but gave mysterious errors when I tried to use it, PyJ4 has this inconvenience of having to start a Gateway server in Java first, JCC wouldn't install. Finally, JPype ended up working. There's a maintained fork of JPype on Github. It has the major advantages that (a) it installs properly and (b) it can very efficiently convert java arrays to numpy array (np_arr = java_arr[:])
The installation process was:
git clone https://github.com/originell/jpype.git
cd jpype
python setup.py install
And you should be able to import jpype
The following demo worked:
import jpype as jp
jp.startJVM(jp.getDefaultJVMPath(), "-ea")
jp.java.lang.System.out.println("hello world")
jp.shutdownJVM()
When I tried calling my own java code, I had to first compile (javac ./blah/HelloWorldJPype.java), and I had to change the JVM path from the default (otherwise you'll get inexplicable "class not found" errors). For me, this meant changing the startJVM command to:
jp.startJVM('/Library/Java/JavaVirtualMachines/jdk1.7.0_79.jdk/Contents/MacOS/libjli.dylib', "-ea")
c = jp.JClass('blah.HelloWorldJPype')
# Where my java class file is in ./blah/HelloWorldJPype.class
...
I've been integrating a lot of stuff into Python lately, including Java. The most robust method I've found is to use IKVM and a C# wrapper.
IKVM has a neat little application that allows you to take any Java JAR, and convert it directly to .Net DLL. It simply translates the JVM bytecode to CLR bytecode. See http://sourceforge.net/p/ikvm/wiki/Ikvmc/ for details.
The converted library behaves just like a native C# library, and you can use it without needing the JVM. You can then create a C# DLL wrapper project, and add a reference to the converted DLL.
You can now create some wrapper stubs that call the methods that you want to expose, and mark those methods as DllEport. See https://stackoverflow.com/a/29854281/1977538 for details.
The wrapper DLL acts just like a native C library, with the exported methods looking just like exported C methods. You can connect to them using ctype as usual.
I've tried it with Python 2.7, but it should work with 3.0 as well. Works on Windows and the Linuxes
If you happen to use C#, then this is probably the best approach to try when integrating almost anything into python.
I'm just beginning to use JPype 0.5.4.2 (july 2011) and it looks like it's working nicely...
I'm on Xubuntu 10.04
I'm assuming that if you can get from C++ to Java then you are all set. I've seen a product of the kind you mention work well. As it happens the one we used was CodeMesh. I'm not specifically endorsing this vendor, or making any statement about their product's relative quality, but I have seen it work in quite a high volume scenario.
I would say generally that if at all possible I would recommend keeping away from direct integration via JNI if you can. Some simple REST service approach, or queue-based architecture will tend to be simpler to develop and diagnose. You can get quite decent perfomance if you use such decoupled technologies carefully.
Through my own experience trying to run some java code from within python in a manner similar to how python code runs within java code in python, I was unable to find a straightforward methodology.
My solution to my problem was by running this java code as BeanShell scripts by calling the BeanShell interpreter as a shell command from within my python code after editing the java code in a temporary file with the appropriate packages and variables.
If what I am talking about is helpful in any manner, I am glad to help you share more details of my solutions.
I just installed a python based library that claims to be compatible with python 2.7+. However, when I go to the run the first sample in their readme I get the SyntaxError: Missing parentheses in call to 'print'.
Groan!
I suppose I can dial code back to use Python 2.7. But before I go there, I wonder if there might exist some other widget that can be installed that perhaps the developers had access to that automatically allows the unparenthesized print statement to work without recoding the whole mess?
I'm confused with some specific behaviour and can't find some informations that help me understand the error.
The situation is as follows: I made a small pyqt4 app that at some point dumps an OrderedDict to a yaml string using pyyaml or ruamel.yaml (tried both) and writes this to a file, or reads from this file. This goes very well executing the code as normal. Now I want to distribute my app by bundling it into a single file windows exe using pyinstaller.
Now if I directly use yaml.dump() or ruamel.yaml.dump() in a method of my pyqt4 form class to generate the yaml-string and write to a file (the standard way using with open ...), I am able to bundle the app using pyinstaller and the exe runs fine.
However, if I write a small function in a sub-folder/module that uses the exact same call to pyyaml (yaml.dump(dict)) or ruamel.yaml (ruamel.yaml.dump(dict, Dumper=ruamel.yaml.RoundTripDumper)) to generate the yaml string and save to a file using with open ... and use this in my pyqt4 method (I just wanted to make things more readable), pyinstaller starts to load a bunch of modules and does a lot more stuff (according to console output), resulting in the exe file beeing almost 5 times larger plus unusable throwing a fatal error pyi_rth_pkgres returned -1 at start.
Unfortunately, I don't understand much from either console output or warnings log, viewable in this gist. Maybe I am searching for the wrong terms. I also tried renaming the module to prevent shadowing.
Now my question is, does anybody know whats going on and can explain this behaviour?
After doing a lot of trial and error, I finally got it working.
I created a new module and build the dumping functions inside it. pyinstaller and the bundled exe work flawless. However, if i do the exact same thing in the previous module, even refactoring the name, it does not work. I even copied the complete code to the old module and it doesn't work. I have no idea why and at this point I am too afraid to ask :|
I am just glad it works now.